404
Not Found
![\"asciicast\"](\"https://asciinema.org/a/247316.svg\")
Not Found
Not Found
You can reach us via the following channels:
This is a SIG-node subproject, hosted under the Kubernetes SIGs organization in Github. The project was established in 2016 and was migrated to Kubernetes SIGs in 2018.
This is open source software released under the Apache 2.0 License.
You can reach us via the following channels:
This is a SIG-node subproject, hosted under the Kubernetes SIGs organization in Github. The project was established in 2016 and was migrated to Kubernetes SIGs in 2018.
This is open source software released under the Apache 2.0 License.
NFD offers two variants of the container image. Released container images are\navailable for x86_64 and Arm64 architectures.
\n\nThe default is a minimal image based on\nscratch\nand only supports running statically linked binaries.
\n\nFor backwards compatibility a container image tag with suffix -minimal\n(e.g. gcr.io/k8s-staging-nfd/node-feature-discovery:master-minimal) is provided.
This image is based on debian:bookworm-slim\nand contains a full Linux system for doing live debugging and diagnosis\nof the NFD images.
\n\nThe container image tag has suffix -full\n(e.g. gcr.io/k8s-staging-nfd/node-feature-discovery:master-full).
Welcome to Node Feature Discovery – a Kubernetes add-on for detecting hardware\nfeatures and system configuration!
\n\nContinue to:
\n\nIntroduction for more details on the\nproject.
\nQuick start for quick step-by-step\ninstructions on how to get NFD running on your cluster.
\n$ kubectl apply -k https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default?ref=master\n namespace/node-feature-discovery created\n serviceaccount/nfd-master created\n clusterrole.rbac.authorization.k8s.io/nfd-master created\n clusterrolebinding.rbac.authorization.k8s.io/nfd-master created\n configmap/nfd-worker-conf created\n service/nfd-master created\n deployment.apps/nfd-master created\n daemonset.apps/nfd-worker created\n\n$ kubectl -n node-feature-discovery get all\n NAME READY STATUS RESTARTS AGE\n pod/nfd-master-555458dbbc-sxg6w 1/1 Running 0 56s\n pod/nfd-worker-mjg9f 1/1 Running 0 17s\n...\n\n$ kubectl get nodes -o json | jq '.items[].metadata.labels'\n {\n \"kubernetes.io/arch\": \"amd64\",\n \"kubernetes.io/os\": \"linux\",\n \"feature.node.kubernetes.io/cpu-cpuid.ADX\": \"true\",\n \"feature.node.kubernetes.io/cpu-cpuid.AESNI\": \"true\",\n...\n\nThis software enables node feature discovery for Kubernetes. It detects\nhardware features available on each node in a Kubernetes cluster, and\nadvertises those features using node labels and optionally node extended\nresources, annotations and node taints. Node Feature Discovery is compatible\nwith any recent version of Kubernetes (v1.24+).
\n\nNFD consists of four software components:
\n\nNFD-Master is the daemon responsible for communication towards the Kubernetes\nAPI. That is, it receives labeling requests from the worker and modifies node\nobjects accordingly.
\n\nNFD-Worker is a daemon responsible for feature detection. It then communicates\nthe information to nfd-master which does the actual node labeling. One\ninstance of nfd-worker is supposed to be running on each node of the cluster,
\n\nNFD-Topology-Updater is a daemon responsible for examining allocated\nresources on a worker node to account for resources available to be allocated\nto new pod on a per-zone basis (where a zone can be a NUMA node). It then\ncreates or updates a\nNodeResourceTopology custom\nresource object specific to this node. One instance of nfd-topology-updater is\nsupposed to be running on each node of the cluster.
\n\nNFD-GC is a daemon responsible for cleaning obsolete\nNodeFeature and\nNodeResourceTopology objects.
\n\nOne instance of nfd-gc is supposed to be running in the cluster.
\n\nFeature discovery is divided into domain-specific feature sources:
\n\nEach feature source is responsible for detecting a set of features which. in\nturn, are turned into node feature labels. Feature labels are prefixed with\nfeature.node.kubernetes.io/ and also contain the name of the feature source.\nNon-standard user-specific feature labels can be created with the local and\ncustom feature sources.
An overview of the default feature labels:
\n\n{\n \"feature.node.kubernetes.io/cpu-<feature-name>\": \"true\",\n \"feature.node.kubernetes.io/custom-<feature-name>\": \"true\",\n \"feature.node.kubernetes.io/kernel-<feature name>\": \"<feature value>\",\n \"feature.node.kubernetes.io/memory-<feature-name>\": \"true\",\n \"feature.node.kubernetes.io/network-<feature-name>\": \"true\",\n \"feature.node.kubernetes.io/pci-<device label>.present\": \"true\",\n \"feature.node.kubernetes.io/storage-<feature-name>\": \"true\",\n \"feature.node.kubernetes.io/system-<feature name>\": \"<feature value>\",\n \"feature.node.kubernetes.io/usb-<device label>.present\": \"<feature value>\",\n \"feature.node.kubernetes.io/<file name>-<feature name>\": \"<feature value>\"\n}\nNFD also annotates nodes it is running on:
\n\n| Annotation | \nDescription | \n
|---|---|
| [<instance>.]nfd.node.kubernetes.io/feature-labels | \nComma-separated list of node labels managed by NFD. NFD uses this internally so must not be edited by users. | \n
| [<instance>.]nfd.node.kubernetes.io/feature-annotations | \nComma-separated list of node annotations managed by NFD. NFD uses this internally so must not be edited by users. | \n
| [<instance>.]nfd.node.kubernetes.io/extended-resources | \nComma-separated list of node extended resources managed by NFD. NFD uses this internally so must not be edited by users. | \n
| [<instance>.]nfd.node.kubernetes.io/taints | \nComma-separated list of node taints managed by NFD. NFD uses this internally so must not be edited by users. | \n
\n\n\nNOTE: the
\n-instance\ncommand line flag affects the annotation names
Unapplicable annotations are not created, i.e. for example\nnfd.node.kubernetes.io/extended-resources is only placed if some extended\nresources were created by NFD.
NFD takes use of some Kubernetes Custom Resources.
\n\nNodeFeatures\nis be used for representing node features and requesting node labels to be\ngenerated.
\n\nNFD-Master uses NodeFeatureRules\nfor custom labeling of nodes.
\n\nNFD-Topology-Updater creates\nNodeResourceTopology objects\nthat describe the hardware topology of node resources.
\n","dir":"/get-started/","name":"introduction.md","path":"get-started/introduction.md","url":"/get-started/introduction.html"},{"title":"Master cmdline reference","layout":"default","sort":1,"content":"To quickly view available command line flags execute nfd-master -help.\nIn a docker container:
docker run gcr.io/k8s-staging-nfd/node-feature-discovery:master nfd-master -help\nPrint usage and exit.
\n\nPrint version and exit.
\n\nThe -feature-gates flag is used to enable or disable non GA features.\nThe list of available feature gates can be found in the feature gates documentation.
Example:
\n\nnfd-master -feature-gates NodeFeatureGroupAPI=true\nThe -prune flag is a sub-command like option for cleaning up the cluster. It\ncauses nfd-master to remove all NFD related labels, annotations and extended\nresources from all Node objects of the cluster and exit.
DEPRECATED: Will be removed in NFD v0.17 and replaced by -port.
The -metrics flag specifies the port on which to expose\nPrometheus metrics. Setting this to 0 disables the\nmetrics server on nfd-master.
Default: 8081
\n\nExample:
\n\nnfd-master -metrics=12345\nThe -instance flag makes it possible to run multiple NFD deployments in\nparallel. In practice, it separates the node annotations between deployments so\nthat each of them can store metadata independently. The instance name must\nstart and end with an alphanumeric character and may only contain alphanumeric\ncharacters, -, _ or ..
Default: empty
\n\nExample:
\n\nnfd-master -instance=network\nThe -enable-leader-election flag enables leader election for NFD-Master.\nIt is advised to turn on this flag when running more than one instance of\nNFD-Master.
Default: false
\n\nnfd-master -enable-leader-election\nThe -enable-taints flag enables/disables node tainting feature of NFD.
Default: false
\n\nExample:
\n\nnfd-master -enable-taints=true\nThe -no-publish flag disables updates to the Node objects in the Kubernetes\nAPI server, making a “dry-run” flag for nfd-master. No Labels, Annotations or\nExtendedResources of nodes are updated.
Default: false
\n\nExample:
\n\nnfd-master -no-publish\nThe -label-whitelist specifies a regular expression for filtering feature\nlabels based on their name. Each label must match against the given regular\nexpression or it will not be published.
\n\n\nNOTE: The regular expression is only matches against the “basename” part\nof the label, i.e. to the part of the name after ‘/’. The label namespace is\nomitted.
\n
Default: empty
\n\nExample:
\n\nnfd-master -label-whitelist='.*cpuid\\.'\nThe -extra-label-ns flag specifies a comma-separated list of allowed feature\nlabel namespaces. This option can be used to allow\nother vendor or application specific namespaces for custom labels from the\nlocal and custom feature sources, even though these labels were denied using\nthe deny-label-ns flag.
Default: empty
\n\nExample:
\n\nnfd-master -extra-label-ns=vendor-1.com,vendor-2.io\nThe -deny-label-ns flag specifies a comma-separated list of excluded\nlabel namespaces. By default, nfd-master allows creating labels in all\nnamespaces, excluding kubernetes.io namespace and its sub-namespaces\n(i.e. *.kubernetes.io). However, you should note that\nkubernetes.io and its sub-namespaces are always denied.\nFor example, nfd-master -deny-label-ns=\"\" would still disallow\nkubernetes.io and *.kubernetes.io.\nThis option can be used to exclude some vendors or application specific\nnamespaces.\nNote that the namespaces feature.node.kubernetes.io and profile.node.kubernetes.io\nand their sub-namespaces are always allowed and cannot be denied.
Default: empty
\n\nExample:
\n\nnfd-master -deny-label-ns=*.vendor.com,vendor-2.io\nThe -config flag specifies the path of the nfd-master configuration file to\nuse.
Default: /etc/kubernetes/node-feature-discovery/nfd-master.conf
\n\nExample:
\n\nnfd-master -config=/opt/nfd/master.conf\nThe -options flag may be used to specify and override configuration file\noptions directly from the command line. The required format is the same as in\nthe config file i.e. JSON or YAML. Configuration options specified via this\nflag will override those from the configuration file:
Default: empty
\n\nExample:
\n\nnfd-master -options='{\"noPublish\": true}'\nThe -nfd-api-parallelism flag can be used to specify the maximum\nnumber of concurrent node updates.
Default: 10
\n\nExample:
\n\nnfd-master -nfd-api-parallelism=1\nThe following logging-related flags are inherited from the\nklog package.
\n\nIf true, adds the file directory to the header of the log messages.
\n\nDefault: false
\n\nLog to standard error as well as files.
\n\nDefault: false
\n\nWhen logging hits line file:N, emit a stack trace.
\n\nDefault: empty
\n\nIf non-empty, write log files in this directory.
\n\nDefault: empty
\n\nIf non-empty, use this log file.
\n\nDefault: empty
\n\nDefines the maximum size a log file can grow to. Unit is megabytes. If the\nvalue is 0, the maximum file size is unlimited.
\n\nDefault: 1800
\n\nLog to standard error instead of files
\n\nDefault: true
\n\nIf true, avoid header prefixes in the log messages.
\n\nDefault: false
\n\nIf true, avoid headers when opening log files.
\n\nDefault: false
\n\nLogs at or above this threshold go to stderr.
\n\nDefault: 2
\n\nNumber for the log level verbosity.
\n\nDefault: 0
\n\nComma-separated list of pattern=N settings for file-filtered logging.
Default: empty
\n\nThe -resync-period flag specifies the NFD API controller resync period.\nThe resync means nfd-master replaying all NodeFeature and NodeFeatureRule objects,\nthus effectively re-syncing all nodes in the cluster (i.e. ensuring labels, annotations,\nextended resources and taints are in place).
Default: 1 hour.
\n\nExample:
\n\nnfd-master -resync-period=2h\nFeatures are advertised as labels in the Kubernetes Node object.
\n\nLabel creation in nfd-worker is performed by a set of separate modules called\nlabel sources. The\ncore.labelSources\nconfiguration option (or\n-label-sources\nflag) of nfd-worker controls which sources to enable for label generation.
All built-in labels use the feature.node.kubernetes.io label namespace and\nhave the following format.
feature.node.kubernetes.io/<feature> = <value>\n\n\n\nNOTE: Consecutive runs of nfd-worker will update the labels on a given\nnode. If features are not discovered on a consecutive run, the corresponding\nlabel will be removed. This includes any restrictions placed on the\nconsecutive run, such as restricting discovered features with the\n
\n-label-whitelist\nflag of nfd-master or\ncore.labelWhiteList\noption of nfd-worker.
| Feature name | \nValue | \nDescription | \n
|---|---|---|
cpu-cpuid.<cpuid-flag> | \n true | \nCPU capability is supported. NOTE: the capability might be supported but not enabled. | \n
cpu-cpuid.<cpuid-attribute> | \n string | \nCPU attribute value | \n
cpu-hardware_multithreading | \n true | \nHardware multithreading, such as Intel HTT, enabled (number of logical CPUs is greater than physical CPUs) | \n
cpu-coprocessor.nx_gzip | \n true | \nNest Accelerator for GZIP is supported(Power). | \n
cpu-power.sst_bf.enabled | \n true | \nIntel SST-BF (Intel Speed Select Technology - Base frequency) enabled | \n
cpu-pstate.status | \n string | \nThe status of the Intel pstate driver when in use and enabled, either ‘active’ or ‘passive’. | \n
cpu-pstate.turbo | \n bool | \nSet to ‘true’ if turbo frequencies are enabled in Intel pstate driver, set to ‘false’ if they have been disabled. | \n
cpu-pstate.scaling_governor | \n string | \nThe value of the Intel pstate scaling_governor when in use, either ‘powersave’ or ‘performance’. | \n
cpu-cstate.enabled | \n bool | \nSet to ‘true’ if cstates are set in the intel_idle driver, otherwise set to ‘false’. Unset if intel_idle cpuidle driver is not active. | \n
cpu-security.sgx.enabled | \n true | \nSet to ‘true’ if Intel SGX is enabled in BIOS (based on a non-zero sum value of SGX EPC section sizes). | \n
cpu-security.se.enabled | \n true | \nSet to ‘true’ if IBM Secure Execution for Linux (IBM Z & LinuxONE) is available and enabled (requires /sys/firmware/uv/prot_virt_host facility) | \n
cpu-security.tdx.enabled | \n true | \nSet to ‘true’ if Intel TDX is available on the host and has been enabled (requires /sys/module/kvm_intel/parameters/tdx). | \n
cpu-security.tdx.protected | \n true | \nSet to ‘true’ if Intel TDX was used to start the guest node, based on the existence of the “TDX_GUEST” information as part of cpuid features. | \n
cpu-security.sev.enabled | \n true | \nSet to ‘true’ if ADM SEV is available on the host and has been enabled (requires /sys/module/kvm_amd/parameters/sev). | \n
cpu-security.sev.es.enabled | \n true | \nSet to ‘true’ if ADM SEV-ES is available on the host and has been enabled (requires /sys/module/kvm_amd/parameters/sev_es). | \n
cpu-security.sev.snp.enabled | \n true | \nSet to ‘true’ if ADM SEV-SNP is available on the host and has been enabled (requires /sys/module/kvm_amd/parameters/sev_snp). | \n
cpu-model.vendor_id | \n string | \nComparable CPU vendor ID. | \n
cpu-model.family | \n int | \nCPU family. | \n
cpu-model.id | \n int | \nCPU model number. | \n
The CPU label source is configurable, see\nworker configuration and\nsources.cpu\nconfiguration options for details.
| Flag | \nDescription | \n
|---|---|
| ADX | \nMulti-Precision Add-Carry Instruction Extensions (ADX) | \n
| AESNI | \nAdvanced Encryption Standard (AES) New Instructions (AES-NI) | \n
| APX_F | \nIntel Advanced Performance Extensions (APX) | \n
| AVX10 | \nIntel Advanced Vector Extensions 10 (AVX10) | \n
| AVX10_256, AVX10_512 | \nIntel AVX10 256-bit and 512-bit vector support | \n
| AVX | \nAdvanced Vector Extensions (AVX) | \n
| AVX2 | \nAdvanced Vector Extensions 2 (AVX2) | \n
| AVXIFMA | \nAVX-IFMA instructions | \n
| AVXVNNI | \nAVX (VEX encoded) VNNI neural network instructions | \n
| AMXBF16 | \nAdvanced Matrix Extension, tile multiplication operations on BFLOAT16 numbers | \n
| AMXINT8 | \nAdvanced Matrix Extension, tile multiplication operations on 8-bit integers | \n
| AMXFP16 | \nAdvanced Matrix Extension, tile multiplication operations on FP16 numbers | \n
| AMXFP8 | \nAdvanced Matrix Extension, tile multiplication operations on FP8 numbers | \n
| AMXTILE | \nAdvanced Matrix Extension, base tile architecture support | \n
| AVX512BF16 | \nAVX-512 BFLOAT16 instructions | \n
| AVX512BITALG | \nAVX-512 bit Algorithms | \n
| AVX512BW | \nAVX-512 byte and word Instructions | \n
| AVX512CD | \nAVX-512 conflict detection instructions | \n
| AVX512DQ | \nAVX-512 doubleword and quadword instructions | \n
| AVX512ER | \nAVX-512 exponential and reciprocal instructions | \n
| AVX512F | \nAVX-512 foundation | \n
| AVX512FP16 | \nAVX-512 FP16 instructions | \n
| AVX512IFMA | \nAVX-512 integer fused multiply-add instructions | \n
| AVX512PF | \nAVX-512 prefetch instructions | \n
| AVX512VBMI | \nAVX-512 vector bit manipulation instructions | \n
| AVX512VBMI2 | \nAVX-512 vector bit manipulation instructions, version 2 | \n
| AVX512VL | \nAVX-512 vector length extensions | \n
| AVX512VNNI | \nAVX-512 vector neural network instructions | \n
| AVX512VP2INTERSECT | \nAVX-512 intersect for D/Q | \n
| AVX512VPOPCNTDQ | \nAVX-512 vector population count doubleword and quadword | \n
| AVXNECONVERT | \nAVX-NE-CONVERT instructions | \n
| AVXVNNIINT8 | \nAVX-VNNI-INT8 instructions | \n
| AVXVNNIINT16 | \nAVX-VNNI-INT16 instructions | \n
| CMPCCXADD | \nCMPCCXADD instructions | \n
| ENQCMD | \nEnqueue Command | \n
| GFNI | \nGalois Field New Instructions | \n
| HYPERVISOR | \nRunning under hypervisor | \n
| MSRLIST | \nRead/Write List of Model Specific Registers | \n
| PREFETCHI | \nPREFETCHIT0/1 instructions | \n
| VAES | \nAVX-512 vector AES instructions | \n
| VPCLMULQDQ | \nCarry-less multiplication quadword | \n
| WRMSRNS | \nNon-Serializing Write to Model Specific Register | \n
By default, the following CPUID flags have been blacklisted: AVX10 (use\nAVX10_VERSION instead), BMI1, BMI2, CLMUL, CMOV, CX16, ERMS, F16C, HTT, LZCNT,\nMMX, MMXEXT, NX, POPCNT, RDRAND, RDSEED, RDTSCP, SGX, SSE, SSE2, SSE3, SSE4,\nSSE42, SSSE3 and TDX_GUEST. See\nsources.cpu\nconfiguration options to change the behavior.
See the full list in github.com/klauspost/cpuid.
\n\n| Attribute | \nDescription | \n
|---|---|
| AVX10_VERSION | \nAVX10 vector ISA version (if supported) | \n
| Flag | \nDescription | \n
|---|---|
| IDIVA | \nInteger divide instructions available in ARM mode | \n
| IDIVT | \nInteger divide instructions available in Thumb mode | \n
| THUMB | \nThumb instructions | \n
| FASTMUL | \nFast multiplication | \n
| VFP | \nVector floating point instruction extension (VFP) | \n
| VFPv3 | \nVector floating point extension v3 | \n
| VFPv4 | \nVector floating point extension v4 | \n
| VFPD32 | \nVFP with 32 D-registers | \n
| HALF | \nHalf-word loads and stores | \n
| EDSP | \nDSP extensions | \n
| NEON | \nNEON SIMD instructions | \n
| LPAE | \nLarge Physical Address Extensions | \n
| Flag | \nDescription | \n
|---|---|
| AES | \nAnnouncing the Advanced Encryption Standard | \n
| EVSTRM | \nEvent Stream Frequency Features | \n
| FPHP | \nHalf Precision(16bit) Floating Point Data Processing Instructions | \n
| ASIMDHP | \nHalf Precision(16bit) Asimd Data Processing Instructions | \n
| ATOMICS | \nAtomic Instructions to the A64 | \n
| ASIMRDM | \nSupport for Rounding Double Multiply Add/Subtract | \n
| PMULL | \nOptional Cryptographic and CRC32 Instructions | \n
| JSCVT | \nPerform Conversion to Match Javascript | \n
| DCPOP | \nPersistent Memory Support | \n
| Feature | \nValue | \nDescription | \n
|---|---|---|
kernel-config.<option> | \n true | \nKernel config option is enabled (set ‘y’ or ‘m’). Default options are NO_HZ, NO_HZ_IDLE, NO_HZ_FULL and PREEMPT | \n
kernel-selinux.enabled | \n true | \nSelinux is enabled on the node | \n
kernel-version.full | \n string | \nFull kernel version as reported by /proc/sys/kernel/osrelease (e.g. ‘4.5.6-7-g123abcde’) | \n
kernel-version.major | \n string | \nFirst component of the kernel version (e.g. ‘4’) | \n
kernel-version.minor | \n string | \nSecond component of the kernel version (e.g. ‘5’) | \n
kernel-version.revision | \n string | \nThird component of the kernel version (e.g. ‘6’) | \n
The kernel label source is configurable, see\nworker configuration and\nsources.kernel\nconfiguration options for details.
| Feature | \nValue | \nDescription | \n
|---|---|---|
memory-numa | \n true | \nMultiple memory nodes i.e. NUMA architecture detected | \n
memory-nv.present | \n true | \nNVDIMM device(s) are present | \n
memory-nv.dax | \n true | \nNVDIMM region(s) configured in DAX mode are present | \n
memory-swap.enabled | \n true | \nSwap is enabled on the node | \n
| Feature | \nValue | \nDescription | \n
|---|---|---|
network-sriov.capable | \n true | \nSingle Root Input/Output Virtualization (SR-IOV) enabled Network Interface Card(s) present | \n
network-sriov.configured | \n true | \nSR-IOV virtual functions have been configured | \n
| Feature | \nValue | \nDescription | \n
|---|---|---|
pci-<device label>.present | \n true | \nPCI device is detected | \n
pci-<device label>.sriov.capable | \n true | \nSingle Root Input/Output Virtualization (SR-IOV) enabled PCI device present | \n
| \n | \n | \n |
<device label> is format is configurable and set to <class>_<vendor> by\ndefault. For more more details about configuration of the pci labels, see\nsources.pci options\nand worker configuration\ninstructions.
| Feature | \nValue | \nDescription | \n
|---|---|---|
usb-<device label>.present | \n true | \nUSB device is detected | \n
<device label> is format is configurable and set to\n<class>_<vendor>_<device> by default. For more more details about\nconfiguration of the usb labels, see\nsources.usb options\nand worker configuration\ninstructions.
| Feature | \nValue | \nDescription | \n
|---|---|---|
storage-nonrotationaldisk | \n true | \nNon-rotational disk, like SSD, is present in the node | \n
| Feature | \nValue | \nDescription | \n
|---|---|---|
system-os_release.ID | \n string | \nOperating system identifier | \n
system-os_release.VERSION_ID | \n string | \nOperating system version identifier (e.g. ‘6.7’) | \n
system-os_release.VERSION_ID.major | \n string | \nFirst component of the OS version id (e.g. ‘6’) | \n
system-os_release.VERSION_ID.minor | \n string | \nSecond component of the OS version id (e.g. ‘7’) | \n
The custom label source is designed for creating\nuser defined labels. However, it has a few statically\ndefined built-in labels:
\n\n| Feature | \nValue | \nDescription | \n
|---|---|---|
custom-rdma.capable | \n true | \nThe node has an RDMA capable Network adapter | \n
custom-rdma.enabled | \n true | \nThe node has the needed RDMA modules loaded to run RDMA traffic | \n
| \n | \n | \n |
NFD has many extension points for creating vendor and application specific\nlabels. See the customization guide for\ndetailed documentation.
\n\nNFD is able to create extended resources, see the\nNodeFeatureRule CRD and its\nextendedResources field for more\ndetails.
\n\nNote that NFD is not a replacement for the usage of device plugins.
\n\nAn example use-case for extended resources could be based on custom feature\n(created e.g. with feature files that\nexposes the node SGX EPC memory section size. This value will then be turned\ninto an extended resource of the node, allowing PODs to request that resource\nand the Kubernetes scheduler to schedule such PODs to only those nodes which\nhave a sufficient capacity of said resource left.
\n\n\n","dir":"/usage/","name":"features.md","path":"usage/features.md","url":"/usage/features.html"},{"title":"Deployment","layout":"default","sort":2,"content":"Node Feature Discovery can be deployed on any recent version of Kubernetes\n(v1.24+).
\n\nSee Image variants for description of the different NFD\ncontainer images available.
\n\nUsing Kustomize provides straightforward deployment with\nkubectl integration and declarative customization.
Using Helm provides easy management of NFD deployments with nice\nconfiguration management and easy upgrades.
\n\nUsing Operator provides deployment and configuration management via\nCRDs.
\n","dir":"/deployment/","name":"index.md","path":"deployment/index.md","url":"/deployment/"},{"title":"Kustomize","layout":"default","sort":2,"content":"Kustomize can be used to\ndeploy NFD. Customization of the deployment is done by maintaining\ndeclarative overlays on top of the base overlays in NFD.
\n\nTo follow the deployment instructions here,\nkubectl v1.24 or\nlater is required.
\n\nThe kustomize overlays provided in the repo can be used directly:
\n\nkubectl apply -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default?ref=master\"\nThis will required RBAC rules and deploy nfd-master (as a deployment) and\nnfd-worker (as daemonset) in the node-feature-discovery namespace.
\n\n\nNOTE: nfd-topology-updater is not deployed as part of the
\ndefault\noverlay. Refer to the Master Worker Topologyupdater\nand Topologyupdater below.
Alternatively you can clone the repository and customize the deployment by\ncreating your own overlays. See kustomize for more information\nabout managing deployment configurations.
\n\nThe NFD repository hosts a set of overlays for different usages and deployment\nscenarios under\ndeployment/overlays
default:\ndefault deployment of nfd-worker as a daemonset, described abovedefault-job:\nsee Worker one-shot belowmaster-worker-topologyupdater:\nsee Master Worker Topologyupdater belowtopologyupdater:\nsee Topology Updater belowprometheus:\nsee Metrics belowprune:\nclean up the cluster after uninstallation, see\nRemoving feature labelssamples/custom-rules:\nan example for spicing up the default deployment with a separately managed\nconfigmap of custom labeling rules, see\nCustom feature source for more information about\ncustom node labelsFeature discovery can alternatively be configured as a one-shot job.\nThe default-job overlay may be used to achieve this:
NUM_NODES=$(kubectl get no -o jsonpath='{.items[*].metadata.name}' | wc -w)\nkubectl kustomize \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default-job?ref=master\" | \\\n sed s\"/NUM_NODES/$NUM_NODES/\" | \\\n kubectl apply -f -\nThe example above launches as many jobs as there are non-master nodes. Note that\nthis approach does not guarantee running once on every node. For example,\ntainted, non-ready nodes or some other reasons in Job scheduling may cause some\nnode(s) will run extra job instance(s) to satisfy the request.
\n\nNFD-Master, nfd-worker and nfd-topology-updater can be configured to be\ndeployed as separate pods. The master-worker-topologyupdater overlay may be\nused to achieve this:
kubectl apply -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/master-worker-topologyupdater?ref=master\"\n\nTo deploy just nfd-topology-updater (without nfd-master and nfd-worker)\nuse the topologyupdater overlay:
kubectl apply -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/topologyupdater?ref=master\"\n\nNFD-Topology-Updater can be configured along with the default overlay\n(which deploys nfd-worker and nfd-master) where all the software components\nare deployed as separate pods;
\nkubectl apply -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default?ref=master\"\nkubectl apply -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/topologyupdater?ref=master\"\n\nTo allow prometheus operator\nto scrape metrics from node-feature-discovery,\nrun the following command:
\n\nkubectl apply -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default?ref=master\"\nkubectl apply -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/prometheus?ref=master\"\nSimplest way is to invoke kubectl delete on the overlay that was used for\ndeployment. Beware that this will also delete the namespace that NFD is\nrunning in. For example, in case the default overlay from the repo was used:
kubectl delete -k https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default?ref=master\nAlternatively you can delete create objects one-by-one, depending on the type\nof deployment, for example:
\n\nNFD_NS=node-feature-discovery\nkubectl -n $NFD_NS delete ds nfd-worker\nkubectl -n $NFD_NS delete deploy nfd-master\nkubectl -n $NFD_NS delete svc nfd-master\nkubectl -n $NFD_NS delete sa nfd-master\nkubectl delete clusterrole nfd-master\nkubectl delete clusterrolebinding nfd-master\nMinimal steps to deploy latest released version of NFD in your cluster.
\n\nDeploy with kustomize – creates a new namespace, service and required RBAC\nrules and deploys nfd-master and nfd-worker daemons.
\n\nkubectl apply -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default?ref=master\"\nWait until NFD master and NFD worker are running.
\n\n$ kubectl -n node-feature-discovery get ds,deploy\nNAME DESIRED CURRENT READY UP-TO-DATE AVAILABLE NODE SELECTOR AGE\ndaemonset.apps/nfd-worker 2 2 2 2 2 <none> 10s\n\nNAME READY UP-TO-DATE AVAILABLE AGE\ndeployment.apps/nfd-master 1/1 1 1 17s\n\nCheck that NFD feature labels have been created
\n\n$ kubectl get no -o json | jq \".items[].metadata.labels\"\n{\n \"kubernetes.io/arch\": \"amd64\",\n \"kubernetes.io/os\": \"linux\",\n \"feature.node.kubernetes.io/cpu-cpuid.ADX\": \"true\",\n \"feature.node.kubernetes.io/cpu-cpuid.AESNI\": \"true\",\n \"feature.node.kubernetes.io/cpu-cpuid.AVX\": \"true\",\n...\nCreate a pod targeting a distinguishing feature (select a valid feature from\nthe list printed on the previous step)
\n\n$ cat << EOF | kubectl apply -f -\napiVersion: v1\nkind: Pod\nmetadata:\n name: feature-dependent-pod\nspec:\n containers:\n - image: registry.k8s.io/pause\n name: pause\n nodeSelector:\n # Select a valid feature\n feature.node.kubernetes.io/cpu-cpuid.AESNI: 'true'\nEOF\npod/feature-dependent-pod created\nSee that the pod is running on a desired node
\n\n$ kubectl get po feature-dependent-pod -o wide\nNAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES\nfeature-dependent-pod 1/1 Running 0 23s 10.36.0.4 node-2 <none> <none>\nTo deploy nfd-topology-updater use the topologyupdater kustomize\noverlay.
kubectl apply -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/topologyupdater?ref=master\"\nWait until nfd-topology-updater is running.
\n\n$ kubectl -n node-feature-discovery get ds\nNAME DESIRED CURRENT READY UP-TO-DATE AVAILABLE NODE SELECTOR AGE\ndaemonset.apps/nfd-topology-updater 2 2 2 2 2 <none> 5s\n\nCheck that the NodeResourceTopology objects are created
\n\n$ kubectl get noderesourcetopologies.topology.node.k8s.io\nNAME AGE\nkind-control-plane 23s\nkind-worker 23s\nTo quickly view available command line flags execute nfd-worker -help.\nIn a docker container:
docker run gcr.io/k8s-staging-nfd/node-feature-discovery:master nfd-worker -help\nPrint usage and exit.
\n\nPrint version and exit.
\n\nThe -feature-gates flag is used to enable or disable non GA features.\nThe list of available feature gates can be found in the feature gates documentation.
Example:
\n\nnfd-master -feature-gates NodeFeatureGroupAPI=true\nThe -config flag specifies the path of the nfd-worker configuration file to\nuse.
Default: /etc/kubernetes/node-feature-discovery/nfd-worker.conf
\n\nExample:
\n\nnfd-worker -config=/opt/nfd/worker.conf\nThe -options flag may be used to specify and override configuration file\noptions directly from the command line. The required format is the same as in\nthe config file i.e. JSON or YAML. Configuration options specified via this\nflag will override those from the configuration file:
Default: empty
\n\nExample:
\n\nnfd-worker -options='{\"sources\":{\"cpu\":{\"cpuid\":{\"attributeWhitelist\":[\"AVX\",\"AVX2\"]}}}}'\nThe -kubeconfig flag specifies the kubeconfig to use for connecting to the\nKubernetes API server. It is needed for manipulating\nNodeFeature objects. An empty value\n(which is also the default) implies in-cluster kubeconfig.
Default: empty
\n\nExample:
\n\nnfd-worker -kubeconfig ${HOME}/.kube/config\nThe -feature-sources flag specifies a comma-separated list of enabled feature\nsources. A special value all enables all sources. Prefixing a source name\nwith - indicates that the source will be disabled instead - this is only\nmeaningful when used in conjunction with all. This command line flag allows\ncompletely disabling the feature detection so that neither standard feature\nlabels are generated nor the raw feature data is available for custom rule\nprocessing. Consider using the core.featureSources config file option,\ninstead, allowing dynamic configurability.
\n\n\nNOTE: This flag takes precedence over the
\ncore.featureSources\nconfiguration file option.
Default: all
\n\nExample:
\n\nnfd-worker -feature-sources=all,-pci\nThe -label-sources flag specifies a comma-separated list of enabled label\nsources. A special value all enables all sources. Prefixing a source name\nwith - indicates that the source will be disabled instead - this is only\nmeaningful when used in conjunction with all. Consider using the\ncore.labelSources config file option, instead, allowing dynamic\nconfigurability.
\n\n\nNOTE: This flag takes precedence over the
\ncore.labelSources\nconfiguration file option.
Default: all
\n\nExample:
\n\nnfd-worker -label-sources=kernel,system,local\nDEPRECATED: Will be removed in NFD v0.17 and replaced by -port.
The -metrics flag specifies the port on which to expose\nPrometheus metrics. Setting this to 0 disables the\nmetrics server on nfd-worker.
Default: 8081
\n\nExample:
\n\nnfd-worker -metrics=12345\nThe -no-publish flag disables all communication with the nfd-master and the\nKubernetes API server. It is effectively a “dry-run” flag for nfd-worker.\nNFD-Worker runs feature detection normally, but no labeling requests are sent\nto nfd-master and no NodeFeature objects are created or updated in the API\nserver.
\n\n\nNOTE: This flag takes precedence over the\n
\ncore.noPublish\nconfiguration file option.
Default: false
\n\nExample:
\n\nnfd-worker -no-publish\nThe -no-owner-refs flag disables setting the owner references to Pod\nof the NodeFeature object.
\n\n\nNOTE: This flag takes precedence over the\n
\ncore.noOwnerRefs\nconfiguration file option.
Default: false
\n\nExample:
\n\nnfd-worker -no-owner-refs\nThe -oneshot flag causes nfd-worker to exit after one pass of feature\ndetection.
Default: false
\n\nExample:
\n\nnfd-worker -oneshot -no-publish\nThe following logging-related flags are inherited from the\nklog package.
\n\n\n\n\nNOTE: The logger setup can also be specified via the
\ncore.klog\nconfiguration file options. However, the command line flags take precedence\nover any corresponding config file options specified.
If true, adds the file directory to the header of the log messages.
\n\nDefault: false
\n\nLog to standard error as well as files.
\n\nDefault: false
\n\nWhen logging hits line file:N, emit a stack trace.
\n\nDefault: empty
\n\nIf non-empty, write log files in this directory.
\n\nDefault: empty
\n\nIf non-empty, use this log file.
\n\nDefault: empty
\n\nDefines the maximum size a log file can grow to. Unit is megabytes. If the\nvalue is 0, the maximum file size is unlimited.
\n\nDefault: 1800
\n\nLog to standard error instead of files
\n\nDefault: true
\n\nIf true, avoid header prefixes in the log messages.
\n\nDefault: false
\n\nIf true, avoid headers when opening log files.
\n\nDefault: false
\n\nLogs at or above this threshold go to stderr.
\n\nDefault: 2
\n\nNumber for the log level verbosity.
\n\nDefault: 0
\n\nComma-separated list of pattern=N settings for file-filtered logging.
Default: empty
\n","dir":"/reference/","name":"worker-commandline-reference.md","path":"reference/worker-commandline-reference.md","url":"/reference/worker-commandline-reference.html"},{"title":"Using node labels","layout":"default","sort":2,"content":"Nodes with specific features can be targeted using the nodeSelector field. The\nfollowing example shows how to target nodes with Intel TurboBoost enabled.
apiVersion: v1\nkind: Pod\nmetadata:\n labels:\n env: test\n name: golang-test\nspec:\n containers:\n - image: golang\n name: go1\n nodeSelector:\n feature.node.kubernetes.io/cpu-pstate.turbo: 'true'\nFor more details on targeting nodes, see\nnode selection.
\n","dir":"/usage/","name":"using-labels.md","path":"usage/using-labels.md","url":"/usage/using-labels.html"},{"title":"Helm","layout":"default","sort":3,"content":"Node Feature Discovery provides a Helm chart to manage its deployment.
\n\n\n\n\nNOTE: NFD is not ideal for other Helm charts to depend on as that may\nresult in multiple parallel NFD deployments in the same cluster which is not\nfully supported by the NFD Helm chart.
\n
Helm package manager should be installed.
\n\nTo install the latest stable version:
\n\nexport NFD_NS=node-feature-discovery\nhelm repo add nfd https://kubernetes-sigs.github.io/node-feature-discovery/charts\nhelm repo update\nhelm install nfd/node-feature-discovery --namespace $NFD_NS --create-namespace --generate-name\nTo install the latest development version you need to clone the NFD Git\nrepository and install from there.
\n\ngit clone https://github.com/kubernetes-sigs/node-feature-discovery/\ncd node-feature-discovery/deployment/helm\nexport NFD_NS=node-feature-discovery\nhelm install node-feature-discovery ./node-feature-discovery/ --namespace $NFD_NS --create-namespace\nSee the configuration section below for instructions how to\nalter the deployment parameters.
\n\nYou can override values from values.yaml and provide a file with custom values:
export NFD_NS=node-feature-discovery\nhelm install nfd/node-feature-discovery -f <path/to/custom/values.yaml> --namespace $NFD_NS --create-namespace\nTo specify each parameter separately you can provide them to helm install command:
\n\nexport NFD_NS=node-feature-discovery\nhelm install nfd/node-feature-discovery --set nameOverride=NFDinstance --set master.replicaCount=2 --namespace $NFD_NS --create-namespace\nTo upgrade the node-feature-discovery deployment to master via Helm.
Please see\nthe uninstallation guide.\nAnd then follow the standard deployment instructions.
\n\nHelm deployment of NFD was introduced in v0.8.0.
\n\nexport NFD_NS=node-feature-discovery\n# Uninstall the old NFD deployment\nhelm uninstall node-feature-discovery --namespace $NFD_NS\n# Update Helm repository\nhelm repo update\n# Install the new NFD deployment\nhelm upgrade --install node-feature-discovery nfd/node-feature-discovery --namespace $NFD_NS --set master.enable=false\n# Wait for NFD Worker to be ready\nkubectl wait --timeout=-1s --for=condition=ready pod -l app.kubernetes.io/name=node-feature-discovery --namespace $NFD_NS\n# Enable the NFD Master\nhelm upgrade --install node-feature-discovery nfd/node-feature-discovery --namespace $NFD_NS --set master.enable=true\nIn v0.12 the NodeFeature CRD was introduced as experimental.\nThe API was not enabled by default.
export NFD_NS=node-feature-discovery\n# Update Helm repository\nhelm repo update\n# Install and upgrade CRD's\nkubectl apply -f https://raw.githubusercontent.com/kubernetes-sigs/node-feature-discovery/master/deployment/base/nfd-crds/nfd-api-crds.yaml\n# Install the new NFD deployment\nhelm upgrade node-feature-discovery nfd/node-feature-discovery --namespace $NFD_NS --set master.enable=false\n# Wait for NFD Worker to be ready\nkubectl wait --timeout=-1s --for=condition=ready pod -l app.kubernetes.io/name=node-feature-discovery --namespace $NFD_NS\n# Enable the NFD Master\nhelm upgrade node-feature-discovery nfd/node-feature-discovery --namespace $NFD_NS --set master.enable=true\nAs of version v0.14 the Helm chart is the primary deployment method for NFD,\nand the CRD NodeFeature is enabled by default.
export NFD_NS=node-feature-discovery\n# Update Helm repository\nhelm repo update\n# Install and upgrade CRD's\nkubectl apply -f https://raw.githubusercontent.com/kubernetes-sigs/node-feature-discovery/master/deployment/base/nfd-crds/nfd-api-crds.yaml\n# Install the new NFD deployment\nhelm upgrade node-feature-discovery nfd/node-feature-discovery --namespace $NFD_NS\nTo uninstall the node-feature-discovery deployment:
export NFD_NS=node-feature-discovery\nhelm uninstall node-feature-discovery --namespace $NFD_NS\nThe command removes all the Kubernetes components associated with the chart and\ndeletes the release. It also runs a post-delete hook that cleans up the nodes\nof all labels, annotations, taints and extended resources that were created by\nNFD.
\n\nTo tailor the deployment of the Node Feature Discovery to your needs following\nChart parameters are available.
\n\n| Name | \nType | \nDefault | \nDescription | \n
|---|---|---|---|
image.repository | \n string | \ngcr.io/k8s-staging-nfd/node-feature-discovery | \n NFD image repository | \n
image.tag | \n string | \nmaster | \n NFD image tag | \n
image.pullPolicy | \n string | \nAlways | \n Image pull policy | \n
imagePullSecrets | \n array | \n[] | \nImagePullSecrets is an optional list of references to secrets in the same namespace to use for pulling any of the images used by this PodSpec. More info. | \n
nameOverride | \n string | \n\n | Override the name of the chart | \n
fullnameOverride | \n string | \n\n | Override a default fully qualified app name | \n
featureGates.NodeFeatureAPI | \n bool | \ntrue | \nEnable the NodeFeature CRD API for communicating node features. This will automatically disable the gRPC communication. | \n
featureGates.NodeFeatureGroupAPI | \n bool | \nfalse | \nEnable the NodeFeatureGroup CRD API. | \n
featureGates.DisableAutoPrefix | \n bool | \nfalse | \nEnable DisableAutoPrefix feature gate. Disables automatic prefixing of unprefixed labels, annotations and extended resources. | \n
prometheus.enable | \n bool | \nfalse | \nSpecifies whether to expose metrics using prometheus operator | \n
prometheus.labels | \n dict | \n{} | \nSpecifies labels for use with the prometheus operator to control how it is selected | \n
prometheus.scrapeInterval | \n string | \n10s | \nSpecifies the interval by which metrics are scraped | \n
priorityClassName | \n string | \n\n | The name of the PriorityClass to be used for the NFD pods. | \n
Metrics are configured to be exposed using prometheus operator API’s by\ndefault. If you want to expose metrics using the prometheus operator\nAPI’s you need to install the prometheus operator in your cluster.
\n\n| Name | \nType | \nDefault | \nDescription | \n
|---|---|---|---|
master.* | \n dict | \n\n | NFD master deployment configuration | \n
master.enable | \n bool | \ntrue | \nSpecifies whether nfd-master should be deployed | \n
master.hostNetwork | \n bool | \nfalse | \nSpecifies whether to enable or disable running the container in the host’s network namespace | \n
master.metricsPort | \n integer | \n8081 | \nPort on which to expose metrics from components to prometheus operator. DEPRECATED: will be replaced by master.port in NFD v0.18. | \n
master.healthPort | \n integer | \n8082 | \nPort on which to expose the grpc health endpoint, will be also used for the probes. DEPRECATED: will be replaced by master.port in NFD v0.18. | \n
master.instance | \n string | \n\n | Instance name. Used to separate annotation namespaces for multiple parallel deployments | \n
master.resyncPeriod | \n string | \n\n | NFD API controller resync period. | \n
master.extraLabelNs | \n array | \n[] | \nList of allowed extra label namespaces | \n
master.enableTaints | \n bool | \nfalse | \nSpecifies whether to enable or disable node tainting | \n
master.replicaCount | \n integer | \n1 | \nNumber of desired pods. This is a pointer to distinguish between explicit zero and not specified | \n
master.podSecurityContext | \n dict | \n{} | \nPodSecurityContext holds pod-level security attributes and common container settings | \n
master.securityContext | \n dict | \n{} | \nContainer security settings | \n
master.serviceAccount.create | \n bool | \ntrue | \nSpecifies whether a service account should be created | \n
master.serviceAccount.annotations | \n dict | \n{} | \nAnnotations to add to the service account | \n
master.serviceAccount.name | \n string | \n\n | The name of the service account to use. If not set and create is true, a name is generated using the fullname template | \n
master.rbac.create | \n bool | \ntrue | \nSpecifies whether to create RBAC configuration for nfd-master | \n
master.resources.limits | \n dict | \n{memory: 4Gi} | \nNFD master pod resources limits | \n
master.resources.requests | \n dict | \n{cpu: 100m, memory: 128Mi} | \nNFD master pod resources requests. See [0] for more info | \n
master.tolerations | \n dict | \nSchedule to control-plane node | \nNFD master pod tolerations | \n
master.annotations | \n dict | \n{} | \nNFD master pod annotations | \n
master.affinity | \n dict | \n\n | NFD master pod required node affinity | \n
master.deploymentAnnotations | \n dict | \n{} | \nNFD master deployment annotations | \n
master.nfdApiParallelism | \n integer | \n10 | \nSpecifies the maximum number of concurrent node updates. | \n
master.config | \n dict | \n\n | NFD master configuration | \n
master.extraArgs | \n array | \n[] | \nAdditional command line arguments to pass to nfd-master | \n
master.extraEnvs | \n array | \n[] | \nAdditional environment variables to pass to nfd-master | \n
master.revisionHistoryLimit | \n integer | \n\n | Specify how many old ReplicaSets for this Deployment you want to retain. revisionHistoryLimit | \n
master.startupProbe.initialDelaySecond s | \n integer | \n0 (by Kubernetes) | \nSpecifies the number of seconds after the container has started before startup probes are initiated. | \n
master.startupProbe.failureThreshold | \n integer | \n30 | \nSpecifies the number of consecutive failures of startup probes before considering the pod as not ready. | \n
master.startupProbe.periodSeconds | \n integer | \n10 (by Kubernetes) | \nSpecifies how often (in seconds) to perform the startup probe. | \n
master.startupProbe.timeoutSeconds | \n integer | \n1 (by Kubernetes) | \nSpecifies the number of seconds after which the probe times out. | \n
master.livenessProbe.initialDelaySeconds | \n integer | \n0 (by Kubernetes) | \nSpecifies the number of seconds after the container has started before liveness probes are initiated. | \n
master.livenessProbe.failureThreshold | \n integer | \n3 (by Kubernetes) | \nSpecifies the number of consecutive failures of liveness probes before considering the pod as not ready. | \n
master.livenessProbe.periodSeconds | \n integer | \n10 (by Kubernetes) | \nSpecifies how often (in seconds) to perform the liveness probe. | \n
master.livenessProbe.timeoutSeconds | \n integer | \n1 (by Kubernetes) | \nSpecifies the number of seconds after which the probe times out. | \n
master.readinessProbe.initialDelaySeconds | \n integer | \n0 (by Kubernetes) | \nSpecifies the number of seconds after the container has started before readiness probes are initiated. | \n
master.readinessProbe.failureThreshold | \n integer | \n10 | \nSpecifies the number of consecutive failures of readiness probes before considering the pod as not ready. | \n
master.readinessProbe.periodSeconds | \n integer | \n10 (by Kubernetes) | \nSpecifies how often (in seconds) to perform the readiness probe. | \n
master.readinessProbe.timeoutSeconds | \n integer | \n1 (by Kubernetes) | \nSpecifies the number of seconds after which the probe times out. | \n
master.readinessProbe.successThreshold | \n integer | \n1 (by Kubernetes) | \nSpecifies the number of consecutive successes of readiness probes before considering the pod as ready. | \n
\n\n\n\n
[0]Additional info formaster.resources.requests:
\nYou may want to use the same value forrequests.memoryandlimits.memory.\nThe “requests” value affects scheduling to accommodate pods on nodes.\nIf there is a large difference between “requests” and “limits” and nodes\nexperience memory pressure, the kernel may invoke the OOM Killer, even if\nthe memory does not exceed the “limits” threshold.\nThis can cause unexpected pod evictions. Memory cannot be compressed and\nonce allocated to a pod, it can only be reclaimed by killing the pod.\nNatan Yellin 22/09/2022\nthat discusses this issue.
| Name | \nType | \nDefault | \nDescription | \n
|---|---|---|---|
worker.* | \n dict | \n\n | NFD worker daemonset configuration | \n
worker.enable | \n bool | \ntrue | \nSpecifies whether nfd-worker should be deployed | \n
worker.hostNetwork | \n bool | \nfalse | \nSpecifies whether to enable or disable running the container in the host’s network namespace | \n
worker.metricsPort | \n int | \n8081 | \nPort on which to expose metrics from components to prometheus operator. DEPRECATED: will be replaced by worker.port in NFD v0.18. | \n
worker.healthPort | \n int | \n8082 | \nPort on which to expose the grpc health endpoint, will be also used for the probes. DEPRECATED: will be replaced by worker.port in NFD v0.18. | \n
worker.config | \n dict | \n\n | NFD worker configuration | \n
worker.podSecurityContext | \n dict | \n{} | \nPodSecurityContext holds pod-level security attributes and common container settins | \n
worker.securityContext | \n dict | \n{} | \nContainer security settings | \n
worker.serviceAccount.create | \n bool | \ntrue | \nSpecifies whether a service account for nfd-worker should be created | \n
worker.serviceAccount.annotations | \n dict | \n{} | \nAnnotations to add to the service account for nfd-worker | \n
worker.serviceAccount.name | \n string | \n\n | The name of the service account to use for nfd-worker. If not set and create is true, a name is generated using the fullname template (suffixed with -worker) | \n
worker.rbac.create | \n bool | \ntrue | \nSpecifies whether to create RBAC configuration for nfd-worker | \n
worker.mountUsrSrc | \n bool | \nfalse | \nSpecifies whether to allow users to mount the hostpath /user/src. Does not work on systems without /usr/src AND a read-only /usr | \n
worker.resources.limits | \n dict | \n{memory: 512Mi} | \nNFD worker pod resources limits | \n
worker.resources.requests | \n dict | \n{cpu: 5m, memory: 64Mi} | \nNFD worker pod resources requests | \n
worker.nodeSelector | \n dict | \n{} | \nNFD worker pod node selector | \n
worker.tolerations | \n dict | \n{} | \nNFD worker pod node tolerations | \n
worker.priorityClassName | \n string | \n\n | NFD worker pod priority class | \n
worker.annotations | \n dict | \n{} | \nNFD worker pod annotations | \n
worker.daemonsetAnnotations | \n dict | \n{} | \nNFD worker daemonset annotations | \n
worker.extraArgs | \n array | \n[] | \nAdditional command line arguments to pass to nfd-worker | \n
worker.extraEnvs | \n array | \n[] | \nAdditional environment variables to pass to nfd-worker | \n
worker.revisionHistoryLimit | \n integer | \n\n | Specify how many old ControllerRevisions for this DaemonSet you want to retain. revisionHistoryLimit | \n
worker.livenessProbe.initialDelaySeconds | \n integer | \n10 | \nSpecifies the number of seconds after the container has started before liveness probes are initiated. | \n
worker.livenessProbe.failureThreshold | \n integer | \n3 (by Kubernetes) | \nSpecifies the number of consecutive failures of liveness probes before considering the pod as not ready. | \n
worker.livenessProbe.periodSeconds | \n integer | \n10 (by Kubernetes) | \nSpecifies how often (in seconds) to perform the liveness probe. | \n
worker.livenessProbe.timeoutSeconds | \n integer | \n1 (by Kubernetes) | \nSpecifies the number of seconds after which the probe times out. | \n
worker.readinessProbe.initialDelaySeconds | \n integer | \n5 | \nSpecifies the number of seconds after the container has started before readiness probes are initiated. | \n
worker.readinessProbe.failureThreshold | \n integer | \n10 | \nSpecifies the number of consecutive failures of readiness probes before considering the pod as not ready. | \n
worker.readinessProbe.periodSeconds | \n integer | \n10 (by Kubernetes) | \nSpecifies how often (in seconds) to perform the readiness probe. | \n
worker.readinessProbe.timeoutSeconds | \n integer | \n1 (by Kubernetes) | \nSpecifies the number of seconds after which the probe times out. | \n
worker.readinessProbe.successThreshold | \n integer | \n1 (by Kubernetes) | \nSpecifies the number of consecutive successes of readiness probes before considering the pod as ready. | \n
| Name | \nType | \nDefault | \nDescription | \n
|---|---|---|---|
topologyUpdater.* | \n dict | \n\n | NFD Topology Updater configuration | \n
topologyUpdater.enable | \n bool | \nfalse | \nSpecifies whether the NFD Topology Updater should be created | \n
topologyUpdater.hostNetwork | \n bool | \nfalse | \nSpecifies whether to enable or disable running the container in the host’s network namespace | \n
topologyUpdater.createCRDs | \n bool | \nfalse | \nSpecifies whether the NFD Topology Updater CRDs should be created | \n
topologyUpdater.serviceAccount.create | \n bool | \ntrue | \nSpecifies whether the service account for topology updater should be created | \n
topologyUpdater.serviceAccount.annotations | \n dict | \n{} | \nAnnotations to add to the service account for topology updater | \n
topologyUpdater.serviceAccount.name | \n string | \n\n | The name of the service account for topology updater to use. If not set and create is true, a name is generated using the fullname template and -topology-updater suffix | \n
topologyUpdater.rbac.create | \n bool | \ntrue | \nSpecifies whether to create RBAC configuration for topology updater | \n
topologyUpdater.metricsPort | \n integer | \n8081 | \nPort on which to expose prometheus metrics. DEPRECATED: will be replaced by topologyUpdater.port in NFD v0.18. | \n
topologyUpdater.healthPort | \n integer | \n8082 | \nPort on which to expose the grpc health endpoint, will be also used for the probes. DEPRECATED: will be replaced by topologyUpdater.port in NFD v0.18. | \n
topologyUpdater.kubeletConfigPath | \n string | \n”” | \nSpecifies the kubelet config host path | \n
topologyUpdater.kubeletPodResourcesSockPath | \n string | \n”” | \nSpecifies the kubelet sock path to read pod resources | \n
topologyUpdater.updateInterval | \n string | \n60s | \nTime to sleep between CR updates. Non-positive value implies no CR update. | \n
topologyUpdater.watchNamespace | \n string | \n* | \n Namespace to watch pods, * for all namespaces | \n
topologyUpdater.podSecurityContext | \n dict | \n{} | \nPodSecurityContext holds pod-level security attributes and common container sett | \n
topologyUpdater.securityContext | \n dict | \n{} | \nContainer security settings | \n
topologyUpdater.resources.limits | \n dict | \n{memory: 60Mi} | \nNFD Topology Updater pod resources limits | \n
topologyUpdater.resources.requests | \n dict | \n{cpu: 50m, memory: 40Mi} | \nNFD Topology Updater pod resources requests | \n
topologyUpdater.nodeSelector | \n dict | \n{} | \nTopology updater pod node selector | \n
topologyUpdater.tolerations | \n dict | \n{} | \nTopology updater pod node tolerations | \n
topologyUpdater.annotations | \n dict | \n{} | \nTopology updater pod annotations | \n
topologyUpdater.daemonsetAnnotations | \n dict | \n{} | \nTopology updater daemonset annotations | \n
topologyUpdater.affinity | \n dict | \n{} | \nTopology updater pod affinity | \n
topologyUpdater.config | \n dict | \n\n | configuration | \n
topologyUpdater.podSetFingerprint | \n bool | \ntrue | \nEnables compute and report of pod fingerprint in NRT objects. | \n
topologyUpdater.kubeletStateDir | \n string | \n/var/lib/kubelet | \nSpecifies kubelet state directory path for watching state and checkpoint files. Empty value disables kubelet state tracking. | \n
topologyUpdater.extraArgs | \n array | \n[] | \nAdditional command line arguments to pass to nfd-topology-updater | \n
topologyUpdater.extraEnvs | \n array | \n[] | \nAdditional environment variables to pass to nfd-topology-updater | \n
topologyUpdater.revisionHistoryLimit | \n integer | \n\n | Specify how many old ControllerRevisions for this DaemonSet you want to retain. revisionHistoryLimit | \n
topologyUpdater.livenessProbe.initialDelaySeconds | \n integer | \n10 | \nSpecifies the number of seconds after the container has started before liveness probes are initiated. | \n
topologyUpdater.livenessProbe.failureThreshold | \n integer | \n3 (by Kubernetes) | \nSpecifies the number of consecutive failures of liveness probes before considering the pod as not ready. | \n
topologyUpdater.livenessProbe.periodSeconds | \n integer | \n10 (by Kubernetes) | \nSpecifies how often (in seconds) to perform the liveness probe. | \n
topologyUpdater.livenessProbe.timeoutSeconds | \n integer | \n1 (by Kubernetes) | \nSpecifies the number of seconds after which the probe times out. | \n
topologyUpdater.readinessProbe.initialDelaySeconds | \n integer | \n5 | \nSpecifies the number of seconds after the container has started before readiness probes are initiated. | \n
topologyUpdater.readinessProbe.failureThreshold | \n integer | \n10 | \nSpecifies the number of consecutive failures of readiness probes before considering the pod as not ready. | \n
topologyUpdater.readinessProbe.periodSeconds | \n integer | \n10 (by Kubernetes) | \nSpecifies how often (in seconds) to perform the readiness probe. | \n
topologyUpdater.readinessProbe.timeoutSeconds | \n integer | \n1 (by Kubernetes) | \nSpecifies the number of seconds after which the probe times out. | \n
topologyUpdater.readinessProbe.successThreshold | \n integer | \n1 (by Kubernetes) | \nSpecifies the number of consecutive successes of readiness probes before considering the pod as ready. | \n
| Name | \nType | \nDefault | \nDescription | \n
|---|---|---|---|
gc.* | \n dict | \n\n | NFD Garbage Collector configuration | \n
gc.enable | \n bool | \ntrue | \nSpecifies whether the NFD Garbage Collector should be created | \n
gc.hostNetwork | \n bool | \nfalse | \nSpecifies whether to enable or disable running the container in the host’s network namespace | \n
gc.serviceAccount.create | \n bool | \ntrue | \nSpecifies whether the service account for garbage collector should be created | \n
gc.serviceAccount.annotations | \n dict | \n{} | \nAnnotations to add to the service account for garbage collector | \n
gc.serviceAccount.name | \n string | \n\n | The name of the service account for garbage collector to use. If not set and create is true, a name is generated using the fullname template and -gc suffix | \n
gc.rbac.create | \n bool | \ntrue | \nSpecifies whether to create RBAC configuration for garbage collector | \n
gc.interval | \n string | \n1h | \nTime between periodic garbage collector runs | \n
gc.podSecurityContext | \n dict | \n{} | \nPodSecurityContext holds pod-level security attributes and common container settings | \n
gc.resources.limits | \n dict | \n{memory: 1Gi} | \nNFD Garbage Collector pod resources limits | \n
gc.resources.requests | \n dict | \n{cpu: 10m, memory: 128Mi} | \nNFD Garbage Collector pod resources requests | \n
gc.metricsPort | \n integer | \n8081 | \nPort on which to serve Prometheus metrics. DEPRECATED: will be replaced by gc.port in NFD v0.18. | \n
gc.nodeSelector | \n dict | \n{} | \nGarbage collector pod node selector | \n
gc.tolerations | \n dict | \n{} | \nGarbage collector pod node tolerations | \n
gc.annotations | \n dict | \n{} | \nGarbage collector pod annotations | \n
gc.deploymentAnnotations | \n dict | \n{} | \nGarbage collector deployment annotations | \n
gc.affinity | \n dict | \n{} | \nGarbage collector pod affinity | \n
gc.extraArgs | \n array | \n[] | \nAdditional command line arguments to pass to nfd-gc | \n
gc.extraEnvs | \n array | \n[] | \nAdditional environment variables to pass to nfd-gc | \n
gc.revisionHistoryLimit | \n integer | \n\n | Specify how many old ReplicaSets for this Deployment you want to retain. revisionHistoryLimit | \n
See the\nsample configuration file\nfor a full example configuration.
\n\nnoPublish option disables updates to the Node objects in the Kubernetes\nAPI server, making a “dry-run” flag for nfd-master. No Labels, Annotations, Taints\nor ExtendedResources of nodes are updated.
Default: false
Example:
\n\nnoPublish: true\nextraLabelNs specifies a list of allowed feature\nlabel namespaces. This option can be used to allow\nother vendor or application specific namespaces for custom labels from the\nlocal and custom feature sources, even though these labels were denied using\nthe denyLabelNs parameter.
Default: empty
\n\nExample:
\n\nextraLabelNs: [\"added.ns.io\",\"added.kubernets.io\"]\ndenyLabelNs specifies a list of excluded\nlabel namespaces. By default, nfd-master allows creating labels in all\nnamespaces, excluding kubernetes.io namespace and its sub-namespaces\n(i.e. *.kubernetes.io). However, you should note that\nkubernetes.io and its sub-namespaces are always denied.\nThis option can be used to exclude some vendors or application specific\nnamespaces.
Default: empty
\n\nExample:
\n\ndenyLabelNs: [\"denied.ns.io\",\"denied.kubernetes.io\"]\nDEPRECATED: Will be removed in NFD v0.17. Use the\nDisableAutoPrefix feature gate instead.
\n\nThe autoDefaultNs option controls the automatic prefixing of names. When set\nto true (the default in NFD version master) nfd-master\nautomatically adds the default feature.node.kubernetes.io/ prefix to\nunprefixed labels, annotations and extended resources - this is also the\ndefault behavior in NFD v0.15 and earlier. When the option is set to false,\nno prefix will be prepended to unprefixed names, effectively causing them to be\nfiltered out (as NFD does not allow unprefixed names of labels, annotations or\nextended resources). The default will be changed to false in a future\nrelease.
For example, with the autoDefaultNs set to true, a NodeFeatureRule with
labels:\n foo: bar\nWill turn into feature.node.kubernetes.io/foo=bar node label. With\nautoDefaultNs set to false, no prefix is added and the label will be\nfiltered out.
Note that taint keys are not affected by this option.
\n\nDefault: true
Example:
\n\nautoDefaultNs: false\nenableTaints enables/disables node tainting feature of NFD.
Default: false
\n\nExample:
\n\nenableTaints: true\nlabelWhiteList specifies a regular expression for filtering feature\nlabels based on their name. Each label must match against the given regular\nexpression or it will not be published.
\n\n\n** NOTE:** The regular expression is only matches against the “basename” part\nof the label, i.e. to the part of the name after ‘/’. The label namespace is\nomitted.
\n
Default: empty
\n\nExample:
\n\nlabelWhiteList: \"foo\"\nThe resyncPeriod option specifies the NFD API controller resync period.\nThe resync means nfd-master replaying all NodeFeature and NodeFeatureRule objects,\nthus effectively re-syncing all nodes in the cluster (i.e. ensuring labels, annotations,\nextended resources and taints are in place).
Default: 1 hour.
\n\nExample:
\n\nresyncPeriod: 2h\nThe leaderElection section exposes configuration to tweak leader election.
leaderElection.leaseDuration is the duration that non-leader candidates will\nwait to force acquire leadership. This is measured against time of\nlast observed ack.
A client needs to wait a full LeaseDuration without observing a change to\nthe record before it can attempt to take over. When all clients are\nshutdown and a new set of clients are started with different names against\nthe same leader record, they must wait the full LeaseDuration before\nattempting to acquire the lease. Thus LeaseDuration should be as short as\npossible (within your tolerance for clock skew rate) to avoid a possible\nlong waits in the scenario.
\n\nDefault: 15 seconds.
\n\nExample:
\n\nleaderElection:\n leaseDurtation: 15s\nleaderElection.renewDeadline is the duration that the acting master will retry\nrefreshing leadership before giving up.
This value has to be lower than leaseDuration and greater than retryPeriod*1.2.
\n\nDefault: 10 seconds.
\n\nExample:
\n\nleaderElection:\n renewDeadline: 10s\nleaderElection.retryPeriod is the duration the LeaderElector clients should wait\nbetween tries of actions.
It has to be greater than 0.
\n\nDefault: 2 seconds.
\n\nExample:
\n\nleaderElection:\n retryPeriod: 2s\nThe nfdApiParallelism option can be used to specify the maximum\nnumber of concurrent node updates.
Default: 10
\n\nExample:
\n\nnfdApiParallelism: 1\nThe following options specify the logger configuration. Most of which can be\ndynamically adjusted at run-time.
\n\n\n\n\nNOTE: The logger options can also be specified via command line flags\nwhich take precedence over any corresponding config file options.
\n
If true, adds the file directory to the header of the log messages.
\n\nDefault: false
Run-time configurable: yes
\n\nLog to standard error as well as files.
\n\nDefault: false
Run-time configurable: yes
\n\nWhen logging hits line file:N, emit a stack trace.
\n\nDefault: empty
\n\nRun-time configurable: yes
\n\nIf non-empty, write log files in this directory.
\n\nDefault: empty
\n\nRun-time configurable: no
\n\nIf non-empty, use this log file.
\n\nDefault: empty
\n\nRun-time configurable: no
\n\nDefines the maximum size a log file can grow to. Unit is megabytes. If the\nvalue is 0, the maximum file size is unlimited.
\n\nDefault: 1800
Run-time configurable: no
\n\nLog to standard error instead of files
\n\nDefault: true
Run-time configurable: yes
\n\nIf true, avoid header prefixes in the log messages.
\n\nDefault: false
Run-time configurable: yes
\n\nIf true, avoid headers when opening log files.
\n\nDefault: false
Run-time configurable: no
\n\nLogs at or above this threshold go to stderr (default 2)
\n\nRun-time configurable: yes
\n\nNumber for the log level verbosity.
\n\nDefault: 0
Run-time configurable: yes
\n\nComma-separated list of pattern=N settings for file-filtered logging.
Default: empty
\n\nRun-time configurable: yes
\n\nThe following options specify the restrictions that can be applied by the\nnfd-master on the deployed Custom Resources in the cluster.
\n\nThe nodeFeatureNamespaceSelector option specifies the NodeFeatures namespaces\nto watch, which can be selected by using metav1.LabelSelector as a type for\nthis option. An empty value selects all namespaces to be watched.
Default: empty
\n\nExample:
\n\nrestrictions:\n nodeFeatureNamespaceSelector:\n matchLabels:\n kubernetes.io/metadata.name: \"node-feature-discovery\"\n matchExpressions:\n - key: \"kubernetes.io/metadata.name\"\n operator: \"In\"\n values:\n - \"node-feature-discovery\"\nThe disableLabels option controls whether to allow creation of node labels\nfrom NodeFeature and NodeFeatureRule CRs or not.
Default: false
\n\nExample:
\n\nrestrictions:\n disableLabels: true\nThe disableExtendedResources option controls whether to allow creation of\nnode extended resources from NodeFeatureRule CR or not.
Default: false
\n\nExample:
\n\nrestrictions:\n disableExtendedResources: true\nhe disableAnnotations option controls whether to allow creation of node annotations\nfrom NodeFeatureRule CR or not.
Default: false
\n\nExample:
\n\nrestrictions:\n disableAnnotations: true\nThe allowOverwrite option controls whether NFD is allowed to overwrite and\ntake over management of existing node labels, annotations, and extended resources.\nLabels, annotations and extended resources created by NFD itself are not affected\n(overwrite cannot be disabled). NFD tracks the labels, annotations and extended\nresources that it manages with specific\nnode annotations.
Default: true
\n\nExample:
\n\nrestrictions:\n allowOverwrite: false\nThe denyNodeFeatureLabels option specifies whether to deny labels from 3rd party\nNodeFeature objects or not. NodeFeature objects created by nfd-worker are not affected.
Default: false
\n\nExample:
\n\nrestrictions:\n denyNodeFeatureLabels: true\nUsage instructions.
\n","dir":"/usage/","name":"index.md","path":"usage/index.md","url":"/usage/"},{"title":"NFD-Master","layout":"default","sort":3,"content":"NFD-Master is responsible for connecting to the Kubernetes API server and\nupdating node objects. More specifically, it modifies node labels, taints and\nextended resources based on requests from nfd-workers and 3rd party extensions.
\n\nThe NodeFeature Controller uses NodeFeature objects as\nthe input for the NodeFeatureRule\nprocessing pipeline. In addition, any labels listed in the NodeFeature object\nare created on the node (note the allowed\nlabel namespaces are controlled).
\n\nNFD-Master acts as the controller for\nNodeFeatureRule objects.\nIt applies the rules specified in NodeFeatureRule objects on raw feature data\nand creates node labels accordingly. The feature data used as the input is\nreceived from nfd-worker instances through\nNodeFeature objects.
\n\nNFD-Master supports configuration through a configuration file. The\ndefault location is /etc/kubernetes/node-feature-discovery/nfd-master.conf,\nbut, this can be changed by specifying the-config command line flag.
Master configuration file is read inside the container, and thus, Volumes and\nVolumeMounts are needed to make your configuration available for NFD. The\npreferred method is to use a ConfigMap which provides easy deployment and\nre-configurability.
\n\nThe provided deployment methods (Helm and Kustomize) create an empty configmap\nand mount it inside the nfd-master containers.
\n\nIn Helm deployments,\nMaster pod parameter\nmaster.config can be used to edit the respective configuration.
In Kustomize deployments, modify the nfd-master-conf ConfigMap with a custom\noverlay.
\n\n\nNOTE: dynamic run-time reconfiguration was dropped in NFD v0.17.\nRe-configuration is handled by pod restarts.
\n
See\nnfd-master configuration file reference\nfor more details.\nThe (empty-by-default)\nexample config\ncontains all available configuration options and can be used as a reference\nfor creating a configuration.
\n\nNFD-Master runs as a deployment, by default\nit prefers running on the cluster’s master nodes but will run on worker\nnodes if no master nodes are found.
\n\nFor High Availability, you should increase the replica count of\nthe deployment object. You should also look into adding\ninter-pod\naffinity to prevent masters from running on the same node.\nHowever note that inter-pod affinity is costly and is not recommended\nin bigger clusters.
\n\n\n\n\nNote: When NFD-Master is intended to run with more than one replica,\nit is advised to use
\n-enable-leader-electionflag. This flag turns on\nleader election for NFD-Master and let only one replica to act on changes\nin NodeFeature and NodeFeatureRule objects.
If you have RBAC authorization enabled (as is the default e.g. with clusters\ninitialized with kubeadm) you need to configure the appropriate ClusterRoles,\nClusterRoleBindings and a ServiceAccount for NFD to create node\nlabels. The provided template will configure these for you.
\n","dir":"/usage/","name":"nfd-master.md","path":"usage/nfd-master.md","url":"/usage/nfd-master.html"},{"title":"NFD Operator","layout":"default","sort":4,"content":"The Node Feature Discovery Operator automates installation,\nconfiguration and updates of NFD using a specific NodeFeatureDiscovery custom\nresource. This also provides good support for managing NFD as a dependency of\nother operators.
\n\nDeployment using the\nNode Feature Discovery Operator\nis recommended to be done via\noperatorhub.io.
\n\nInstall the operator:
\n\nkubectl create -f https://operatorhub.io/install/nfd-operator.yaml\nCreate NodeFeatureDiscovery object (in nfd namespace here):
cat << EOF | kubectl apply -f -\napiVersion: v1\nkind: Namespace\nmetadata:\n name: nfd\n---\napiVersion: nfd.kubernetes.io/v1\nkind: NodeFeatureDiscovery\nmetadata:\n name: my-nfd-deployment\n namespace: nfd\nspec:\n operand:\n image: gcr.io/k8s-staging-nfd/node-feature-discovery:master\n imagePullPolicy: IfNotPresent\nEOF\nIf you followed the deployment instructions above you can uninstall NFD with:
\n\nkubectl -n nfd delete NodeFeatureDiscovery my-nfd-deployment\nOptionally, you can also remove the namespace:
\n\nkubectl delete ns nfd\nSee the node-feature-discovery-operator and OLM project\ndocumentation for instructions for uninstalling the operator and operator\nlifecycle manager, respectively.
\n\n\n","dir":"/deployment/","name":"operator.md","path":"deployment/operator.md","url":"/deployment/operator.html"},{"title":"Reference","layout":"default","sort":4,"content":"Command line and configuration reference.
\n","dir":"/reference/","name":"index.md","path":"reference/index.md","url":"/reference/"},{"title":"Worker config reference","layout":"default","sort":4,"content":"See the\nsample configuration file\nfor a full example configuration.
\n\nThe core section contains common configuration settings that are not specific\nto any particular feature source.
core.sleepInterval specifies the interval between consecutive passes of\nfeature (re-)detection, and thus also the interval between node re-labeling. A\nnon-positive value implies infinite sleep interval, i.e. no re-detection or\nre-labeling is done.
Default: 60s
Example:
\n\ncore:\n sleepInterval: 60s\ncore.featureSources specifies the list of enabled feature sources. A special\nvalue all enables all sources. Prefixing a source name with - indicates\nthat the source will be disabled instead - this is only meaningful when used in\nconjunction with all. This option allows completely disabling the feature\ndetection so that neither standard feature labels are generated nor the raw\nfeature data is available for custom rule processing.
Default: [all]
Example:
\n\ncore:\n # Enable all but cpu and local sources\n featureSources:\n - \"all\"\n - \"-cpu\"\n - \"-local\"\ncore:\n # Enable only cpu and local sources\n featureSources:\n - \"cpu\"\n - \"local\"\ncore.labelSources specifies the list of enabled label sources. A special\nvalue all enables all sources. Prefixing a source name with - indicates\nthat the source will be disabled instead - this is only meaningful when used in\nconjunction with all. This configuration option affects the generation of\nnode labels but not the actual discovery of the underlying feature data that is\nused e.g. in custom/NodeFeatureRule rules.
\n\n\nNOTE: Overridden by the
\n-label-sourcescommand line flag and the\ncore.sourcesconfigurations option (if either of them is specified).
Default: [all]
Example:
\n\ncore:\n # Enable all but cpu and system sources\n labelSources:\n - \"all\"\n - \"-cpu\"\n - \"-system\"\ncore:\n # Enable only cpu and system sources\n labelSources:\n - \"cpu\"\n - \"system\"\nDEPRECATED: use core.labelSources instead.
\n\n\nNOTE:
\ncore.sourcestakes precedence over thecore.labelSources\nconfiguration file option.
core.labelWhiteList specifies a regular expression for filtering feature\nlabels based on the label name. Non-matching labels are not published.
\n\n\nNOTE: The regular expression is only matches against the “basename” part\nof the label, i.e. to the part of the name after ‘/’. The label prefix (or\nnamespace) is omitted.
\n
Default: null
Example:
\n\ncore:\n labelWhiteList: '^cpu-cpuid'\nSetting core.noPublish to true disables all communication with the\nnfd-master and the Kubernetes API server. It is effectively a “dry-run” option.\nNFD-Worker runs feature detection normally, but no labeling requests are sent\nto nfd-master and no NodeFeature\nobjects are created or updated in the API server.
\n\n\nNOTE: Overridden by the\n
\n-no-publish\ncommand line flag (if specified).
Default: false
Example:
\n\ncore:\n noPublish: true\nSetting core.noOwnerRefs to true disables setting the owner references\nof the NodeFeature object created by the nfd-worker.
\n\n\nNOTE: Overridden by the\n
\n-no-owner-refs\ncommand line flag (if specified).
Default: false
Example:
\n\ncore:\n noOwnerRefs: true\nThe following options specify the logger configuration.
\n\n\n\n\nNOTE: The logger options can also be specified via command line flags\nwhich take precedence over any corresponding config file options.
\n
If true, adds the file directory to the header of the log messages.
\n\nDefault: false
Log to standard error as well as files.
\n\nDefault: false
When logging hits line file:N, emit a stack trace.
\n\nDefault: empty
\n\nIf non-empty, write log files in this directory.
\n\nDefault: empty
\n\nIf non-empty, use this log file.
\n\nDefault: empty
\n\nDefines the maximum size a log file can grow to. Unit is megabytes. If the\nvalue is 0, the maximum file size is unlimited.
\n\nDefault: 1800
Log to standard error instead of files
\n\nDefault: true
If true, avoid header prefixes in the log messages.
\n\nDefault: false
If true, avoid headers when opening log files.
\n\nDefault: false
Logs at or above this threshold go to stderr (default 2)
\n\nNumber for the log level verbosity.
\n\nDefault: 0
Comma-separated list of pattern=N settings for file-filtered logging.
Default: empty
\n\nThe sources section contains feature source specific configuration parameters.
Prevent publishing cpuid features listed in this option.
\n\n\n\n\nNOTE: overridden by
\nsources.cpu.cpuid.attributeWhitelist(if specified)
Default: [AVX10, BMI1, BMI2, CLMUL, CMOV, CX16, ERMS, F16C, HTT, LZCNT, MMX, MMXEXT,\nNX, POPCNT, RDRAND, RDSEED, RDTSCP, SGX, SGXLC, SSE, SSE2, SSE3, SSE4.1,\nSSE4.2, SSSE3, TDX_GUEST]
Example:
\n\nsources:\n cpu:\n cpuid:\n attributeBlacklist: [MMX, MMXEXT]\nOnly publish the cpuid features listed in this option.
\n\n\n\n\nNOTE: takes precedence over
\nsources.cpu.cpuid.attributeBlacklist
Default: empty
\n\nExample:
\n\nsources:\n cpu:\n cpuid:\n attributeWhitelist: [AVX512BW, AVX512CD, AVX512DQ, AVX512F, AVX512VL]\nPath of the kernel config file. If empty, NFD runs a search in the well-known\nstandard locations.
\n\nDefault: empty
\n\nExample:
\n\nsources:\n kernel:\n kconfigFile: \"/path/to/kconfig\"\nKernel configuration options to publish as feature labels.
\n\nDefault: [NO_HZ, NO_HZ_IDLE, NO_HZ_FULL, PREEMPT]
Example:
\n\nsources:\n kernel:\n configOpts: [NO_HZ, X86, DMI]\nList of PCI device class IDs for which to\npublish a label. Can be specified as a main class only (e.g. 03) or full\nclass-subclass combination (e.g. 0300) - the former implies that all\nsubclasses are accepted. The format of the labels can be further configured\nwith deviceLabelFields.
Default: [\"03\", \"0b40\", \"12\"]
Example:
\n\nsources:\n pci:\n deviceClassWhitelist: [\"0200\", \"03\"]\nThe set of PCI ID fields to use when constructing the name of the feature\nlabel. Valid fields are class, vendor, device, subsystem_vendor and\nsubsystem_device.
Default: [class, vendor]
Example:
\n\nsources:\n pci:\n deviceLabelFields: [class, vendor, device]\nWith the example config above NFD would publish labels like:\nfeature.node.kubernetes.io/pci-<class-id>_<vendor-id>_<device-id>.present=true
List of USB device class IDs for\nwhich to publish a feature label. The format of the labels can be further\nconfigured with deviceLabelFields.
\n\nDefault: [\"0e\", \"ef\", \"fe\", \"ff\"]
Example:
\n\nsources:\n usb:\n deviceClassWhitelist: [\"ef\", \"ff\"]\nThe set of USB ID fields from which to compose the name of the feature label.\nValid fields are class, vendor, device and serial.
Default: [class, vendor, device]
Example:
\n\nsources:\n pci:\n deviceLabelFields: [class, vendor]\nWith the example config above NFD would publish labels like:\nfeature.node.kubernetes.io/usb-<class-id>_<vendor-id>.present=true
List of rules to process in the custom feature source to create user-specific\nlabels. Refer to the documentation of the\ncustom feature source for\ndetails of the available rules and their configuration.
\n\nDefault: empty
\n\nExample:
\n\nsources:\n custom:\n - name: \"my custom rule\"\n labels:\n my-custom-feature: \"true\"\n matchFeatures:\n - feature: kernel.loadedmodule\n matchExpressions:\n e1000e: {op: Exists}\n - feature: pci.device\n matchExpressions:\n class: {op: In, value: [\"0200\"]}\n vendor: {op: In, value: [\"8086\"]}\nNFD-Worker is preferably run as a Kubernetes DaemonSet. This assures\nre-labeling on regular intervals capturing changes in the system configuration\nand makes sure that new nodes are labeled as they are added to the cluster.\nWorker connects to the nfd-master service to advertise hardware features.
\n\nWhen run as a daemonset, nodes are re-labeled at an default interval of 60s.\nThis can be changed by using the\ncore.sleepInterval\nconfig option.
NFD-Worker supports configuration through a configuration file. The\ndefault location is /etc/kubernetes/node-feature-discovery/nfd-worker.conf,\nbut, this can be changed by specifying the-config command line flag.\nConfiguration file is re-read whenever it is modified which makes run-time\nre-configuration of nfd-worker straightforward.
Worker configuration file is read inside the container, and thus, Volumes and\nVolumeMounts are needed to make your configuration available for NFD. The\npreferred method is to use a ConfigMap which provides easy deployment and\nre-configurability.
\n\nThe provided deployment methods (Helm and Kustomize) create an empty configmap\nand mount it inside the nfd-master containers.
\n\nIn Helm deployments,\nWorker pod parameter\nworker.config can be used to edit the respective configuration.
In Kustomize deployments, modify the nfd-worker-conf ConfigMap with a custom\noverlay.
\n\n\nNOTE: dynamic run-time reconfiguration was dropped in NFD v0.17.\nRe-configuration is handled by pod restarts.
\n
See\nnfd-worker configuration file reference\nfor more details.\nThe (empty-by-default)\nexample config\ncontains all available configuration options and can be used as a reference\nfor creating a configuration.
\n\nConfiguration options can also be specified via the -options command line\nflag, in which case no mounts need to be used. The same format as in the config\nfile must be used, i.e. JSON (or YAML). For example:
-options='{\"sources\": { \"pci\": { \"deviceClassWhitelist\": [\"12\"] } } }'\nConfiguration options specified from the command line will override those read\nfrom the config file.
\n","dir":"/usage/","name":"nfd-worker.md","path":"usage/nfd-worker.md","url":"/usage/nfd-worker.html"},{"title":"Uninstallation","layout":"default","sort":5,"content":"Follow the uninstallation instructions of the deployment method used\n(kustomize,\nhelm or\noperator).
\n\n\n\n\nNOTE: This is unnecessary when using the Helm chart for deployment as it\nwill clean up the nodes when NFD is uninstalled.
\n
NFD-Master has a special -prune command line flag for removing all\nnfd-related node labels, annotations, extended resources and taints from the\ncluster.
kubectl apply -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/prune?ref=master\"\nkubectl -n node-feature-discovery wait job.batch/nfd-master --for=condition=complete && \\\n kubectl delete -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/prune?ref=master\"\n\n\n","dir":"/deployment/","name":"uninstallation.md","path":"deployment/uninstallation.md","url":"/deployment/uninstallation.html"},{"title":"Developer guide","layout":"default","sort":5,"content":"NOTE: You must run prune before removing the RBAC rules (serviceaccount,\nclusterrole and clusterrolebinding).
\n
git clone https://github.com/kubernetes-sigs/node-feature-discovery\ncd node-feature-discovery\nSee customizing the build below for altering the\ncontainer image registry, for example.
\n\nmake\nOptional, this example with Docker.
\n\ndocker push <IMAGE_TAG>\nThe default set of architectures enabled for mulit-arch builds are linux/amd64\nand linux/arm64. If more architectures are needed one can override the\nIMAGE_ALL_PLATFORMS variable with a comma separated list of OS/ARCH tuples.
make image-all\nCurrently docker does not support loading of manifest-lists meaning the images\nare not shown when executing docker images, see:\nbuildx issue #59.
make push-all\nThe resulting container image can be used in the same way on each arch by pulling\ne.g. node-feature-discovery:master without specifying the\narchitecture. The manifest-list will take care of providing the right\narchitecture image.
To use your published image from the step above instead of the\nregistry.k8s.io/nfd/node-feature-discovery image, edit image\nattribute in the spec template(s) to the new location\n(<registry-name>/<image-name>[:<version>]).
The yamls makefile generates a kustomization.yaml matching your locally\nbuilt image and using the deploy/overlays/default deployment. See\nbuild customization below for configurability, e.g.\nchanging the deployment namespace.
K8S_NAMESPACE=my-ns make yamls\nkubectl apply -k .\nYou can use alternative deployment methods by modifying the auto-generated\nkustomization file.
\n\nYou can also build the binaries locally
\n\nmake build\nThis will compile binaries under bin/
There are several Makefile variables that control the build process and the\nname of the resulting container image. The following are targeted targeted for\nbuild customization and they can be specified via environment variables or\nmakefile overrides.
\n\n| Variable | \nDescription | \nDefault value | \n
|---|---|---|
| HOSTMOUNT_PREFIX | \nPrefix of system directories for feature discovery (local builds) | \n/ (local builds) /host- (container builds) | \n
| IMAGE_BUILD_CMD | \nCommand to build the image | \ndocker build | \n
| IMAGE_BUILD_EXTRA_OPTS | \nExtra options to pass to build command | \nempty | \n
| IMAGE_BUILDX_CMD | \nCommand to build and push multi-arch images with buildx | \nDOCKER_CLI_EXPERIMENTAL=enabled docker buildx build –platform=${IMAGE_ALL_PLATFORMS} –progress=auto –pull | \n
| IMAGE_ALL_PLATFORMS | \nComma separated list of OS/ARCH tuples for mulit-arch builds | \nlinux/amd64,linux/arm64 | \n
| IMAGE_PUSH_CMD | \nCommand to push the image to remote registry | \ndocker push | \n
| IMAGE_REGISTRY | \nContainer image registry to use | \nregistry.k8s.io/nfd | \n
| IMAGE_TAG_NAME | \nContainer image tag name | \n<nfd version> | \n
| IMAGE_EXTRA_TAG_NAMES | \nAdditional container image tag(s) to create when building image | \nempty | \n
| K8S_NAMESPACE | \nnfd-master and nfd-worker namespace | \nnode-feature-discovery | \n
For example, to use a custom registry:
\n\nmake IMAGE_REGISTRY=<my custom registry uri>\nOr to specify a build tool different from Docker, It can be done in 2 ways:
\n\nvia environment
\n\n IMAGE_BUILD_CMD=\"buildah bud\" make\nby overriding the variable value
\n\n make IMAGE_BUILD_CMD=\"buildah bud\"\nUnit tests are automatically run as part of the container image build. You can\nalso run them manually in the source code tree by running:
\n\nmake test\nEnd-to-end tests are built on top of the e2e test framework of Kubernetes, and,\nthey required a cluster to run them on. For running the tests on your test\ncluster you need to specify the kubeconfig to be used:
\n\nmake e2e-test KUBECONFIG=$HOME/.kube/config\nThere are several environment variables that can be used to customize the\ne2e-tests:
\n\n| Variable | \nDescription | \nDefault value | \n
|---|---|---|
| KUBECONFIG | \nKubeconfig for running e2e-tests | \nempty | \n
| E2E_TEST_CONFIG | \nParameterization file of e2e-tests (see example) | \nempty | \n
| E2E_PULL_IF_NOT_PRESENT | \nTrue-ish value makes the image pull policy IfNotPresent (to be used only in e2e tests) | \nfalse | \n
| E2E_TEST_FULL_IMAGE | \nRun e2e-test also against the Full Image tag | \nfalse | \n
| E2E_GINKGO_LABEL_FILTER | \nGinkgo label filter to use for running e2e tests | \nempty | \n
| OPENSHIFT | \nNon-empty value enables OpenShift specific support (only affects e2e tests) | \nempty | \n
When running as a standalone container labeling is expected to fail because\nKubernetes API is not available. Thus, it is recommended to use -no-publish.
$ export NFD_CONTAINER_IMAGE=gcr.io/k8s-staging-nfd/node-feature-discovery:master\n$ docker run --rm --name=nfd-test ${NFD_CONTAINER_IMAGE} nfd-master -no-publish -crd-controller=false -feature-gates NodeFeatureAPI=false\n2019/02/01 14:48:21 Node Feature Discovery Master <NFD_VERSION>\nTo run nfd-worker as a “stand-alone” container you need to run it in the same\nnetwork namespace as the nfd-master container:
\n\n$ docker run --rm --network=container:nfd-test ${NFD_CONTAINER_IMAGE} nfd-worker -feature-gates NodeFeatureAPI=false\n2019/02/01 14:48:56 Node Feature Discovery Worker <NFD_VERSION>\n...\nIf you just want to try out feature discovery without connecting to nfd-master,\npass the -no-publish flag to nfd-worker.
\n\n\nNOTE: Some feature sources need certain directories and/or files from the\nhost mounted inside the NFD container. Thus, you need to provide Docker with\nthe correct
\n--volumeoptions for them to work correctly when run\nstand-alone directly withdocker run. See\nthe default deployment\nfor up-to-date information about the required volume mounts.
To run nfd-topology-updater as a “stand-alone” container\nyou need to run it in with the -no-publish flag to disable communication to\nthe Kubernetes apiserver.
$ docker run --rm ${NFD_CONTAINER_IMAGE} nfd-topology-updater -no-publish\n2019/02/01 14:48:56 Node Feature Discovery Topology Updater <NFD_VERSION>\n...\nIf you just want to try out resource topology discovery without connecting to\nthe Kubernetes API, pass the -no-publish flag to nfd-topology-updater.
\n\n\nNOTE: NFD topology updater needs certain directories and/or files from\nthe host mounted inside the NFD container. Thus, you need to provide Docker\nwith the correct
\n--volumeoptions for them to work correctly when\nrun stand-alone directly withdocker run. See\nthe template spec\nfor up-to-date information about the required volume mounts.
Another option for building NFD locally is via Tilt tool, which can build container\nimages, push them to a local registry and reload your Kubernetes pods automatically.\nWhen using Tilt, you don’t have to build container images and re-deploy your pods\nmanually but instead let the Tilt take care of it. Tiltfile is a configuration file\nfor the Tilt and is located at the root directory. To develop NFD with Tilt, follow\nthe steps below.
\n\nStart up node feature discovery development environment\n To start up your Tilt development environment, run at the root of your\n local NFD codebase.
\n\n tilt up\nTilt will start a web interface in the localhost and port 10350. From the\n web interface, you are able to see how NFD worker and master are\n progressing, watch their build and runtime logs. Once your code changes are\n saved locally, Tilt will notice it and re-build the container image from\n the current code, push the image to the registry and re-deploy NFD pods\n with the latest container image.
\nTo override environment variables used in the Tiltfile during image build,\nexport them in your current terminal before starting Tilt.
\n\nexport IMAGE_TAG_NAME=\"v1\"\ntilt up\nThis will override the default value(master) of IMAGE_TAG_NAME variable defined\nin the Tiltfile.
All documentation resides under the\ndocs\ndirectory in the source tree. It is designed to be served as a html site by\nGitHub Pages.
\n\nBuilding the documentation is containerized to fix the build\nenvironment. The recommended way for developing documentation is to run:
\n\nmake site-serve\nThis will build the documentation in a container and serve it under\nlocalhost:4000/ making it easy to verify the results.\nAny changes made to the docs/ will automatically re-trigger a rebuild and are\nreflected in the served content and can be inspected with a browser refresh.
To just build the html documentation run:
\n\nmake site-build\nThis will generate html documentation under docs/_site/.
To quickly view available command line flags execute nfd-topology-updater -help.\nIn a docker container:
docker run gcr.io/k8s-staging-nfd/node-feature-discovery:master \\\nnfd-topology-updater -help\nPrint usage and exit.
\n\nPrint version and exit.
\n\nThe -config flag specifies the path of the nfd-topology-updater\nconfiguration file to use.
Default: /etc/kubernetes/node-feature-discovery/nfd-topology-updater.conf
\n\nExample:
\n\nnfd-topology-updater -config=/opt/nfd/nfd-topology-updater.conf\nThe -no-publish flag makes for a “dry-run” flag for nfd-topology-updater.\nNFD-Topology-Updater runs resource hardware topology detection normally, but\nNodeResourceTopology\nobjects are not created or updated.
Default: false
\n\nExample:
\n\nnfd-topology-updater -no-publish\nThe -oneshot flag causes nfd-topology-updater to exit after one pass of\nresource hardware topology detection.
Default: false
\n\nExample:
\n\nnfd-topology-updater -oneshot -no-publish\nDEPRECATED: Will be removed in NFD v0.17 and replaced by -port.
The -metrics flag specifies the port on which to expose\nPrometheus metrics. Setting this to 0 disables the\nmetrics server on nfd-topology-updater.
Default: 8081
\n\nExample:
\n\nnfd-topology-updater -metrics=12345\nThe -sleep-interval specifies the interval between resource hardware\ntopology re-examination (and CR updates). zero means no CR updates on interval basis.
Default: 60s
\n\nExample:
\n\nnfd-topology-updater -sleep-interval=1h\nThe -watch-namespace specifies the namespace to ensure that resource\nhardware topology examination only happens for the pods running in the\nspecified namespace. Pods that are not running in the specified namespace\nare not considered during resource accounting. This is particularly useful\nfor testing/debugging purpose. A “*” value would mean that all the pods would\nbe considered during the accounting process.
Default: “*”
\n\nExample:
\n\nnfd-topology-updater -watch-namespace=rte\nThe -kubelet-config-uri specifies the path to the Kubelet’s configuration.\nNote that the URi could either be a local host file or an HTTP endpoint.
Default: https://${NODE_ADDRESS}:10250/configz
Example:
\n\nnfd-topology-updater -kubelet-config-uri=file:///var/lib/kubelet/config.yaml\nThe -api-auth-token-file specifies the path to the api auth token file\nwhich is used to retrieve Kubelet’s configuration from Kubelet secure port,\nonly taking effect when -kubelet-config-uri is https.\nNote that this token file must bind to a role that has the get capability to\nnodes/proxy resources.
Default: /var/run/secrets/kubernetes.io/serviceaccount/token
Example:
\n\nnfd-topology-updater -token-file=/var/run/secrets/kubernetes.io/serviceaccount/token\nThe -podresources-socket specifies the path to the Unix socket where kubelet\nexports a gRPC service to enable discovery of in-use CPUs and devices, and to\nprovide metadata for them.
Default: /host-var/lib/kubelet/pod-resources/kubelet.sock
\n\nExample:
\n\nnfd-topology-updater -podresources-socket=/var/lib/kubelet/pod-resources/kubelet.sock\nEnables compute and report the pod set fingerprint in the NRT.\nA pod fingerprint is a compact representation of the “node state” regarding resources.
\n\nDefault: true
Example:
\n\nnfd-topology-updater -pods-fingerprint=false\nThe -kubelet-state-dir specifies the path to the Kubelet state directory,\nwhere state and checkpoint files are stored.\nThe files are mount as read-only and cannot be change by the updater.\nEnabled by default.\nPassing an empty string will disable the watching.
Default: /host-var/lib/kubelet
\n\nExample:
\n\nnfd-topology-updater -kubelet-state-dir=/var/lib/kubelet\nNFD-Topology-Updater is preferably run as a Kubernetes DaemonSet.\nThis assures re-examination on regular intervals\nand/or per pod life-cycle events, capturing changes in the allocated\nresources and hence the allocatable resources on a per-zone basis by updating\nNodeResourceTopology custom resources.\nIt makes sure that new NodeResourceTopology instances are created for each new\nnodes that get added to the cluster.
\n\nBecause of the design and implementation of Kubernetes, only resources exclusively\nallocated to Guaranteed Quality of Service\npods will be accounted.\nThis includes\nCPU cores,\nmemory\nand\ndevices.
\n\nWhen run as a daemonset, nodes are re-examined for the allocated resources\n(to determine the information of the allocatable resources on a per-zone basis\nwhere a zone can be a NUMA node) at an interval specified using the\n-sleep-interval\noption. The default sleep interval is set to 60s\nwhich is the value when no -sleep-interval is specified.\nThe re-examination can be disabled by setting the sleep-interval to 0.
Another option is to configure the updater to update\nthe allocated resources per pod life-cycle events.\nThe updater will monitor the checkpoint file stated in\n-kubelet-state-dir\nand triggers an update for every change occurs in the files.
In addition, it can avoid examining specific allocated resources\ngiven a configuration of resources to exclude via -excludeList
Kubelet PodResource API with the\nGetAllocatableResources functionality enabled is a\nprerequisite for nfd-topology-updater to be able to run (i.e. Kubernetes v1.21\nor later is required).
\n\nPreceding Kubernetes v1.23, the kubelet must be started with\n--feature-gates=KubeletPodResourcesGetAllocatable=true.
Starting from Kubernetes v1.23, the KubeletPodResourcesGetAllocatable\nfeature gate. is enabled by default
NFD-Topology-Updater supports configuration through a configuration file. The\ndefault location is /etc/kubernetes/node-feature-discovery/topology-updater.conf,\nbut, this can be changed by specifying the-config command line flag.
Topology-Updater configuration file is read inside the container,\nand thus, Volumes and VolumeMounts are needed\nto make your configuration available for NFD.\nThe preferred method is to use a ConfigMap\nwhich provides easy deployment and re-configurability.
\n\nThe provided deployment templates create an empty configmap\nand mount it inside the nfd-topology-updater containers.
\n\nIn Helm deployments,\nTopology Updater parameters\ntoplogyUpdater.config can be used to edit the respective configuration.
In Kustomize deployments, modify the nfd-worker-conf ConfigMap with a custom\noverlay.
See\nnfd-topology-updater configuration file reference\nfor more details.\nThe (empty-by-default)\nexample config\ncontains all available configuration options and can be used as a reference\nfor creating a configuration.
\n\n\n","dir":"/usage/","name":"nfd-topology-updater.md","path":"usage/nfd-topology-updater.md","url":"/usage/nfd-topology-updater.html"},{"title":"Contributing","layout":"default","sort":6,"content":"You can reach us via the following channels:
\n\nThis is a\nSIG-node\nsubproject, hosted under the\nKubernetes SIGs organization in Github.\nThe project was established in 2016 and was migrated to Kubernetes SIGs in 2018.
\n\nThis is open source software released under the Apache 2.0 License.
\n","dir":"/contributing/","name":"index.md","path":"contributing/index.md","url":"/contributing/"},{"title":"Metrics","layout":"default","sort":6,"content":"Metrics are configured to be exposed using prometheus operator\nAPI’s by default. If you want to expose metrics using the prometheus operator\nAPI’s you need to install the prometheus operator in your cluster.\nBy default NFD Master and Worker expose metrics on port 8081.
\n\nThe exposed metrics are
\n\n| Metric | \nType | \nDescription | \n
|---|---|---|
nfd_master_build_info | \n Gauge | \nVersion from which nfd-master was built | \n
nfd_worker_build_info | \n Gauge | \nVersion from which nfd-worker was built | \n
nfd_gc_build_info | \n Gauge | \nVersion from which nfd-gc was built | \n
nfd_topology_updater_build_info | \n Gauge | \nVersion from which nfd-topology-updater was built | \n
nfd_master_node_update_requests_total | \n Counter | \nNumber of node update requests received by the master over gRPC | \n
nfd_master_node_updates_total | \n Counter | \nNumber of nodes updated | \n
nfd_master_node_feature_group_update_requests_total | \n Counter | \nNumber of cluster feature update requests processed by the master | \n
nfd_master_node_update_failures_total | \n Counter | \nNumber of nodes update failures | \n
nfd_master_node_labels_rejected_total | \n Counter | \nNumber of nodes labels rejected by nfd-master | \n
nfd_master_node_extendedresources_rejected_total | \n Counter | \nNumber of nodes extended resources rejected by nfd-master | \n
nfd_master_node_taints_rejected_total | \n Counter | \nNumber of nodes taints rejected by nfd-master | \n
nfd_master_nodefeaturerule_processing_duration_seconds | \n Histogram | \nTime taken to process NodeFeatureRule objects | \n
nfd_master_nodefeaturerule_processing_errors_total | \n Counter | \nNumber or errors encountered while processing NodeFeatureRule objects | \n
nfd_worker_feature_discovery_duration_seconds | \n Histogram | \nTime taken to discover features on a node | \n
nfd_topology_updater_scan_errors_total | \n Counter | \nNumber of errors in scanning resource allocation of pods. | \n
nfd_gc_objects_deleted_total | \n Counter | \nNumber of NodeFeature and NodeResourceTopology objects garbage collected. | \n
nfd_gc_object_delete_failures_total | \n Counter | \nNumber of errors in deleting NodeFeature and NodeResourceTopology objects. | \n
To deploy NFD with metrics enabled using kustomize, you can use the\nprometheus overlay.
\n\nBy default metrics are enabled when deploying NFD via Helm. To enable Prometheus\nto scrape metrics from NFD, you need to pass the following values to Helm:
\n\n--set prometheus.enable=true\nFor more info on Helm deployment, see Helm.
\n\nIt is recommended to specify\n--set prometheus.prometheusSpec.podMonitorSelectorNilUsesHelmValues=false\nwhen deploying prometheus-operator via Helm to enable the prometheus-operator\nto scrape metrics from any PodMonitor.
or setting labels on the PodMonitor via the helm parameter prometheus.labels\nto control which Prometheus instances will scrape this PodMonitor.
NFD contains an example Grafana dashboard. You can import\nexamples/grafana-dashboard.json\nto your Grafana instance to visualize the NFD metrics.
See the\nsample configuration file\nfor a full example configuration.
\n\nThe excludeList specifies a key-value map of allocated resources\nthat should not be examined by the topology-updater\nagent per node.\nEach key is a node name with a value as a list of resources\nthat should not be examined by the agent for that specific node.
Default: empty
\n\nExample:
\n\nexcludeList:\n nodeA: [hugepages-2Mi]\n nodeB: [memory]\n nodeC: [cpu, hugepages-2Mi]\nexcludeList.* is a special value that use to specify all nodes.\nA resource that would be listed under this key, would be excluded from all nodes.
Default: empty
\n\nExample:
\n\nexcludeList:\n '*': [hugepages-2Mi]\nNFD-GC (NFD Garbage-Collector) is preferably run as a Kubernetes deployment\nwith one replica. It makes sure that all\nNodeFeature and\nNodeResourceTopology objects\nhave corresponding nodes and removes stale objects for non-existent nodes.
\n\nThe daemon watches for Node deletion events and removes NodeFeature and\nNodeResourceTopology objects upon them. It also runs periodically to make sure\nno node delete event was missed and to remove any NodeFeature or\nNodeResourceTopology objects that were created without corresponding node. The\ndefault garbage collector interval is set to 1h which is the value when no\n-gc-interval is specified.
\n\nIn Helm deployments see\ngarbage collector parameters\nfor altering the nfd-gc configuration.
\n","dir":"/usage/","name":"nfd-gc.md","path":"usage/nfd-gc.md","url":"/usage/nfd-gc.html"},{"title":"Garbage Collector Cmdline Reference","layout":"default","sort":7,"content":"To quickly view available command line flags execute nfd-gc -help.\nIn a docker container:
docker run gcr.io/k8s-staging-nfd/node-feature-discovery:master \\\nnfd-gc -help\nPrint usage and exit.
\n\nPrint version and exit.
\n\nThe -gc-interval specifies the interval between periodic garbage collector runs.
Default: 1h
\n\nExample:
\n\nnfd-gc -gc-interval=1h\nNFD uses some Kubernetes custom resources.
\n\nNodeFeature is an NFD-specific custom resource for communicating node\nfeatures and node labeling requests. The nfd-master pod watches for NodeFeature\nobjects, labels nodes as specified and uses the listed features as input when\nevaluating NodeFeatureRules. NodeFeature objects can be\nused for implementing 3rd party extensions (see\ncustomization guide for more\ndetails).
\n\napiVersion: nfd.k8s-sigs.io/v1alpha1\nkind: NodeFeature\nmetadata:\n labels:\n nfd.node.kubernetes.io/node-name: node-1\n name: node-1-vendor-features\nspec:\n features:\n instances:\n vendor.device:\n elements:\n - attributes:\n model: \"xpu-1\"\n memory: \"4000\"\n type: \"fast\"\n - attributes:\n model: \"xpu-2\"\n memory: \"16000\"\n type: \"slow\"\n labels:\n vendor-xpu-present: \"true\"\nNodeFeatureGroup is an NFD-specific custom resource that is designed for\ngrouping nodes based on their features. NFD-Master watches for NodeFeatureGroup\nobjects in the cluster and updates the status of the NodeFeatureGroup object\nwith the list of nodes that match the feature group rules. The NodeFeatureGroup\nrules follow the same syntax as the NodeFeatureRule rules.
\n\napiVersion: nfd.k8s-sigs.io/v1alpha1\nkind: NodeFeatureGroup\nmetadata:\n name: node-feature-group-example\nspec:\n featureGroupRules:\n - name: \"node has kernel version discovered\"\n matchFeatures:\n - feature: kernel.version\n matchExpressions:\n major: {op: Exists}\nNodeFeatureGroup API is an alpha feature and disabled by default in NFD version\nmaster. For more details and examples see the\ncustomization guide.
\n\nNodeFeatureRule is an NFD-specific custom resource that is designed for\nrule-based custom labeling of nodes. NFD-Master watches for NodeFeatureRule\nobjects in the cluster and labels nodes according to the rules within. Some use\ncases are e.g. application specific labeling in a specific environments or\nbeing distributed by hardware vendors to create specific labels for their\ndevices.
\n\napiVersion: nfd.k8s-sigs.io/v1alpha1\nkind: NodeFeatureRule\nmetadata:\n name: example-rule\nspec:\n rules:\n - name: \"example rule\"\n labels:\n \"example-custom-feature\": \"true\"\n # Label is created if all of the rules below match\n matchFeatures:\n # Match if \"veth\" kernel module is loaded\n - feature: kernel.loadedmodule\n matchExpressions:\n veth: {op: Exists}\n # Match if any PCI device with vendor 8086 exists in the system\n - feature: pci.device\n matchExpressions:\n vendor: {op: In, value: [\"8086\"]}\nSee the\nCustomization guide\nfor full documentation of the NodeFeatureRule resource and its usage.
\n\nThe\ndeployment/nodefeaturerule/samples/\ndirectory contains sample NodeFeatureRule objects that replicate the built-in\ndefault feature labels generated by NFD. The sample rules can be used as a base\nto customize NFD feature labels. To use them in place of the the NFD built-in\nlabels, the corresponding feature source(s) of nfd-worker should be disabled\nwith the\ncore.labelSources\nconfiguration option.
When run with NFD-Topology-Updater, NFD creates NodeResourceTopology objects\ncorresponding to node resource hardware topology such as:
\n\napiVersion: topology.node.k8s.io/v1alpha1\nkind: NodeResourceTopology\nmetadata:\n name: node1\ntopologyPolicies: [\"SingleNUMANodeContainerLevel\"]\nzones:\n - name: node-0\n type: Node\n resources:\n - name: cpu\n capacity: 20\n allocatable: 16\n available: 10\n - name: vendor/nic1\n capacity: 3\n allocatable: 3\n available: 3\n - name: node-1\n type: Node\n resources:\n - name: cpu\n capacity: 30\n allocatable: 30\n available: 15\n - name: vendor/nic2\n capacity: 6\n allocatable: 6\n available: 6\n - name: node-2\n type: Node\n resources:\n - name: cpu\n capacity: 30\n allocatable: 30\n available: 15\n - name: vendor/nic1\n capacity: 3\n allocatable: 3\n available: 3\nThe NodeResourceTopology objects created by NFD can be used to gain insight\ninto the allocatable resources along with the granularity of those resources at\na per-zone level (represented by node-0 and node-1 in the above example) or can\nbe used by an external entity (e.g. topology-aware scheduler plugin) to take an\naction based on the gathered information.
\n\n\n","dir":"/usage/","name":"custom-resources.md","path":"usage/custom-resources.md","url":"/usage/custom-resources.html"},{"title":"Kubectl plugin cmdline reference","layout":"default","sort":8,"content":"To quickly view available command line flags execute kubectl nfd -help.
Print usage and exit.
\n\nValidate a NodeFeatureRule file.
\n\nThe --nodefeature-file flag specifies the path to the NodeFeatureRule file\nto validate.
Test a NodeFeatureRule file against a node without applying it.
\n\nThe --kubeconfig flag specifies the path to the kubeconfig file to use for\nCLI requests.
The --namespace flag specifies the namespace to use for CLI requests.\nDefault: default.
The --nodename flag specifies the name of the node to test the\nNodeFeatureRule against.
The --nodefeaturerule-file flag specifies the path to the NodeFeatureRule file\nto test.
Process a NodeFeatureRule file against a NodeFeature file.
\n\nThe --nodefeaturerule-file flag specifies the path to the NodeFeatureRule file\nto test.
The --nodefeature-file flag specifies the path to the NodeFeature file to test.
NFD provides multiple extension points for vendor and application specific\nlabeling:
\n\nNodeFeature objects can be\nused to communicate “raw” node features and node labeling requests to\nnfd-master.NodeFeatureRule objects provide a way to\ndeploy custom labeling rules via the Kubernetes API.local feature source of nfd-worker creates\nlabels by reading text files.custom feature source of nfd-worker creates\nlabels based on user-specified rules.NodeFeature objects provide a way for 3rd party extensions to advertise custom\nfeatures, both as “raw” features that serve as input to\nNodeFeatureRule objects and as feature\nlabels directly.
\n\nNote that RBAC rules must be created for each extension for them to be able to\ncreate and manipulate NodeFeature objects in their namespace.
\n\nConsider the following referential example:
\n\napiVersion: nfd.k8s-sigs.io/v1alpha1\nkind: NodeFeature\nmetadata:\n labels:\n nfd.node.kubernetes.io/node-name: node-1\n name: vendor-features-for-node-1\nspec:\n # Features for NodeFeatureRule matching\n features:\n flags:\n vendor.flags:\n elements:\n feature-x: {}\n feature-y: {}\n attributes:\n vendor.config:\n elements:\n setting-a: \"auto\"\n knob-b: \"123\"\n instances:\n vendor.devices:\n elements:\n - attributes:\n model: \"dev-1000\"\n vendor: \"acme\"\n - attributes:\n model: \"dev-2000\"\n vendor: \"acme\"\n # Labels to be created\n labels:\n vendor.io/feature.enabled: \"true\"\nThe object targets node named node-1. It lists two “flag type” features under\nthe vendor.flags domain, two “attribute type” features and under the\nvendor.config domain and two “instance type” features under the\nvendor.devices domain. These features will not be directly affecting the node\nlabels but they will be used as input when the\nNodeFeatureRule objects are evaluated.
In addition, the example requests directly the\nvendor.io/feature.enabled=true node label to be created.
The nfd.node.kubernetes.io/node-name=<node-name> must be in place for each\nNodeFeature object as NFD uses it to determine the node which it is targeting.
Features have three different types:
\n\nNodeFeatureRule objects provide an easy way to create vendor or application\nspecific labels and taints. It uses a flexible rule-based mechanism for creating\nlabels and optionally taints based on node features.
Consider the following referential example:
\n\napiVersion: nfd.k8s-sigs.io/v1alpha1\nkind: NodeFeatureRule\nmetadata:\n name: my-sample-rule-object\nspec:\n rules:\n - name: \"my sample rule\"\n labels:\n \"feature.node.kubernetes.io/my-sample-feature\": \"true\"\n matchFeatures:\n - feature: kernel.loadedmodule\n matchExpressions:\n dummy: {op: Exists}\n - feature: kernel.config\n matchExpressions:\n X86: {op: In, value: [\"y\"]}\nIt specifies one rule which creates node label\nfeature.node.kubernetes.io/my-sample-feature=true if both of the following\nconditions are true (matchFeatures implements a logical AND over the\nmatchers):
dummy network driver module has been loaded=yCreate a NodeFeatureRule with a yaml file:
kubectl apply -f https://raw.githubusercontent.com/kubernetes-sigs/node-feature-discovery/master/examples/nodefeaturerule.yaml\nNow, on X86 platforms the feature label appears after doing modprobe dummy on\na system and correspondingly the label is removed after rmmod dummy. Note a\nre-labeling delay up to the sleep-interval of nfd-worker (1 minute by default).
See Feature rule format for detailed description of\navailable fields and how to write labeling rules.
\n\nIn some circumstances, it is desirable to keep nodes with specialized hardware\naway from running general workload and instead leave them for workloads that\nneed the specialized hardware. One way to achieve it is to taint the nodes with\nthe specialized hardware and add corresponding toleration to pods that require\nthe special hardware. NFD offers node tainting functionality which is disabled\nby default. User can define one or more custom taints via the taints field of\nthe NodeFeatureRule CR. The same rule-based mechanism is applied here and the\nNFD taints only rule matching nodes.
To enable the tainting feature, --enable-taints flag needs to be set to true.\nIf the flag --enable-taints is set to false (i.e. disabled), taints defined in\nthe NodeFeatureRule CR have no effect and will be ignored by the NFD master.
See documentation of the taints field for detailed description how\nto specify taints in the NodeFeatureRule object.
\n\n\n\n\nNOTE: Before enabling any taints, make sure to edit nfd-worker daemonset\nto tolerate the taints to be created. Otherwise, already running pods that do\nnot tolerate the taint are evicted immediately from the node including the\nnfd-worker pod.
\n
NodeFeatureGroup API is an alpha feature and disabled by default in NFD version\nmaster. Use the\nNodeFeatureAPI feature\ngate to enable it.
\n\nNodeFeatureGroup objects provide a way to create node groups that share the\nsame set of features. The NodeFeatureGroup object spec consists of a list of\nNodeFeatureRule that follow the same format as the NodeFeatureRule,\nbut the difference in this case is that nodes that match any of the rules in the\nNodeFeatureGroup will be listed in the NodeFeatureGroup status.
Consider the following referential example:
\n\napiVersion: nfd.k8s-sigs.io/v1alpha1\nkind: NodeFeatureGroup\nmetadata:\n name: node-feature-group-example\nspec:\n featureGroupRules:\n - name: \"kernel version\"\n matchFeatures:\n - feature: kernel.version\n matchExpressions:\n major: {op: In, value: [\"6\"]}\nstatus:\n nodes:\n - name: node-1\n - name: node-2\n - name: node-3\nThe object specifies a group of nodes that share the same\nkernel.version.major (Linux kernel v6.x).
Create a NodeFeatureGroup with a yaml file:
kubectl apply -f https://raw.githubusercontent.com/kubernetes-sigs/node-feature-discovery/master/examples/nodefeaturegroup.yaml\nSee Feature rule format for detailed description of\navailable fields and how to write group filtering rules.
\n\nNFD-Worker has a special feature source named local which is an integration\npoint for external feature detectors. It provides a mechanism for pluggable\nextensions, allowing the creation of new user-specific features and even\noverriding built-in labels.
The local feature source uses feature files. The features discovered by the\nlocal source can further be used in label rules specified in\nNodeFeatureRule objects and\nthe custom feature source.
\n\n\nNOTE: Be careful when creating and/or updating feature files\nwhile NFD is running. To avoid race conditions you should write\ninto a temporary file, and atomically create/update the original file by\ndoing a file rename operation. NFD ignores dot files,\nso temporary file can be written to the same directory and renamed\n(
\n.my.feature->my.feature) once file is complete. Both file names should\n(obviously) be unique for the given application.
Consider a plaintext file\n/etc/kubernetes/node-feature-discovery/features.d/my-features\nhaving the following contents:
feature.node.kubernetes.io/my-feature.1\nfeature.node.kubernetes.io/my-feature.2=myvalue\nvendor.io/my-feature.3=456\nThis will translate into the following node labels:
\n\nfeature.node.kubernetes.io/my-feature.1: \"true\"\nfeature.node.kubernetes.io/my-feature.2: \"myvalue\"\nvendor.io/my-feature.3: \"456\"\nThe local source reads files found in\n/etc/kubernetes/node-feature-discovery/features.d/. File content is parsed\nand translated into node labels, see the input format below.
The feature files are expected to contain features in simple\nkey-value pairs, separated by newlines:
\n\n# This is a comment\n<key>[=<value>]\nThe label value defaults to true, if not specified.
Label namespace must be specified with <namespace>/<name>[=<value>].
\n\n\nNOTE: The feature file size limit it 64kB. The feature file will be\nignored if the size limit is exceeded.
\n
Comment lines (starting with #) are ignored.
Adding following line anywhere to feature file defines date when\nits content expires / is ignored:
\n\n# +expiry-time=2023-07-29T11:22:33Z\nAlso, the expiry-time value would stay the same during the processing of the\nfeature file until another expiry-time directive is encountered.\nConsidering the following file:
\n\n# +expiry-time=2012-07-28T11:22:33Z\nvendor.io/feature1=featureValue\n\n# +expiry-time=2080-07-28T11:22:33Z\nvendor.io/feature2=featureValue2\n\n# +expiry-time=2070-07-28T11:22:33Z\nvendor.io/feature3=featureValue3\n\n# +expiry-time=2002-07-28T11:22:33Z\nvendor.io/feature4=featureValue4\nAfter processing the above file, only vendor.io/feature2 and\nvendor.io/feature3 would be included in the list of accepted features.
\n\n\nNOTE: The time format supported is RFC3339. Also, the
\nexpiry-time\ntag is only evaluated in each re-discovery period, and the expiration of\nnode labels is not tracked.
To exclude specific features from the local.feature Feature, you can use the\n# +no-feature directive. The # +no-label directive causes the feature to\nbe excluded from the local.label Feature and a node label not to be generated.
Considering the following file:
\n\n# +no-feature\nvendor.io/label-only=value\n\nvendor.io/my-feature=value\n\nvendor.io/foo=bar\n\n# +no-label\nfoo=baz\nProcessing the above file would result in the following Features:
\n\nlocal.features:\n foo: baz\n vendor.io/my-feature: value\nlocal.labels:\n vendor.io/label-only: value\n vendor.io/my-feature: value\nand the following labels added to the Node:
\n\nvendor.io/label-only=value\nvendor.io/my-feature=value\n\n\n\nNOTE: use of unprefixed label names (like
\nfoo=bar) should not be used.\nIn NFD master unprefixed names will be automatically prefixed\nwithfeature.node.kubernetes.io/but this will change in a future version\n(see\nautoDefaultNs config option.\nUnprefixed names for plain Features (tagged with# +no-label) can be used\nwithout restrictions, however.
The standard NFD deployments contain hostPath mounts for\n/etc/kubernetes/node-feature-discovery/features.d/, making these directories\nfrom the host available inside the nfd-worker container.
One use case for the feature files is detecting features in other\nPods outside NFD, e.g. in Kubernetes device plugins. By using the same\nhostPath mounts /etc/kubernetes/node-feature-discovery/features.d/\nin the side-car (e.g. device plugin) creates a shared area for\ndeploying feature files to NFD.
The custom feature source in nfd-worker provides a rule-based mechanism for\nlabel creation, similar to the\nNodeFeatureRule objects. The difference is\nthat the rules are specified in the worker configuration instead of a\nKubernetes API object.
See worker configuration\nfor instructions how to set-up and manage the worker configuration.
\n\nConsider the following referential configuration for nfd-worker:
\n\ncore:\n labelSources: [\"custom\"]\nsources:\n custom:\n - name: \"my sample rule\"\n labels:\n \"feature.node.kubenernetes.io/my-sample-feature\": \"true\"\n matchFeatures:\n - feature: kernel.loadedmodule\n matchExpressions:\n dummy: {op: Exists}\n - feature: kernel.config\n matchExpressions:\n X86: {op: In, value: [\"y\"]}\nIt specifies one rule which creates node label\nfeature.node.kubenernetes.io/my-sample-feature=true if both of the following\nconditions are true (matchFeatures implements a logical AND over the\nmatchers):
dummy network driver module has been loaded=yIn addition, the configuration only enables the custom source, disabling all\nbuilt-in labels.
Now, on X86 platforms the feature label appears after doing modprobe dummy on\na system and correspondingly the label is removed after rmmod dummy. Note a\nre-labeling delay up to the sleep-interval of nfd-worker (1 minute by default).
In addition to the rules defined in the nfd-worker configuration file, the\ncustom feature source can read more configuration files located in the\n/etc/kubernetes/node-feature-discovery/custom.d/ directory. This makes more\ndynamic and flexible configuration easier.
As an example, consider having file\n/etc/kubernetes/node-feature-discovery/custom.d/my-rule.yaml with the\nfollowing content:
- name: \"my e1000 rule\"\n labels:\n \"feature.node.kubenernetes.io/e1000.present\": \"true\"\n matchFeatures:\n - feature: kernel.loadedmodule\n matchExpressions:\n e1000: {op: Exists}\nThis simple rule will create feature.node.kubenernetes.io/e1000.present=true\nlabel if the e1000 kernel module has been loaded.
The\nsamples/custom-rules\nkustomize overlay sample contains an example for deploying a custom rule from a\nConfigMap.
Feature labels have the following format:
\n\n<namespace>/<name> = <value>\nThe namespace part (i.e. prefix) of the labels is controlled by nfd:
\n\nfeature.node.kubernetes.io.-deny-label-ns\ncommand line flag of nfd-master\n -extra-label-ns\ncommand line flag of nfd-master.\ne.g: nfd-master -deny-label-ns=\"*\" -extra-label-ns=example.comThis section describes the rule format used in\nNodeFeatureRule objects and in the\nconfiguration of the custom feature source.
It is based on a generic feature matcher that covers all features discovered by\nnfd-worker. The rules rely on a unified data model of the available features\nand a generic expression-based format. Features that can be used in the rules\nare described in detail in available features below.
\n\nTake this rule as a referential example:
\n\n - name: \"my feature rule\"\n labels:\n \"feature.node.kubernetes.io/my-special-feature\": \"my-value\"\n matchFeatures:\n - feature: cpu.cpuid\n matchExpressions:\n AVX512F: {op: Exists}\n - feature: kernel.version\n matchExpressions:\n major: {op: In, value: [\"5\"]}\n minor: {op: Gt, value: [\"1\"]}\n - feature: pci.device\n matchExpressions:\n vendor: {op: In, value: [\"8086\"]}\n class: {op: In, value: [\"0200\"]}\nThis will yield feature.node.kubernetes.io/my-special-feature=my-value node\nlabel if all of these are true (matchFeatures implements a logical AND over\nthe matchers):
The .name field is required and used as an identifier of the rule.
The .labels is a map of the node labels to create if the rule matches.
Take this rule as a referential example:
\n\napiVersion: nfd.k8s-sigs.io/v1alpha1\nkind: NodeFeatureRule\nmetadata:\n name: my-sample-rule-object\nspec:\n rules:\n - name: \"my dynamic label value rule\"\n labels:\n feature.node.kubernetes.io/linux-lsm-enabled: \"@kernel.config.LSM\"\n feature.node.kubernetes.io/custom-label: \"customlabel\"\nLabel linux-lsm-enabled uses the @ notation for dynamic values.\nThe value of the label will be the value of the attribute LSM\nof the feature kernel.config.
The @<feature-name>.<element-name> format can be used to inject values of\ndetected features to the label. See\navailable features for possible values to use.
This will yield into the following node label:
\n\n labels:\n ...\n feature.node.kubernetes.io/linux-lsm-enabled: apparmor\n feature.node.kubernetes.io/custom-label: \"customlabel\"\nThe .labelsTemplate field specifies a text template for dynamically creating\nlabels based on the matched features. See templating for\ndetails.
\n\n\nNOTE: The
\nlabelsfield has priority overlabelsTemplate, i.e.\nlabels specified in thelabelsfield will override anything\noriginating fromlabelsTemplate.
The .annotations field is a list of features to be advertised as node\nannotations.
Take this rule as a referential example:
\n\napiVersion: nfd.k8s-sigs.io/v1alpha1\nkind: NodeFeatureRule\nmetadata:\n name: feature-annotations-example\nspec:\n rules:\n - name: \"annotation-example\"\n annotations:\n feature.node.kubernetes.io/defaul-ns-annotation: \"foo\"\n custom.vendor.io/feature: \"baz\"\n matchFeatures:\n - feature: kernel.version\n matchExpressions:\n major: {op: Exists}\nThis will yield into the following node annotations:
\n\n annotations:\n ...\n feature.node.kubernetes.io/defaul-ns-annotation: \"foo\"\n custom.vendor.io/feature: \"baz\"\n ...\nNFD enforces some limitations to the namespace (or prefix)/ of the annotations:
\n\nkubernetes.io/ and its sub-namespaces (like sub.ns.kubernetes.io/) cannot\ngenerally be usedfeature.node.kubernetes.io/ and its sub-namespaces\n(like sub.ns.feature.node.kubernetes.io)my-annotation) should not be used. In NFD master unprefixed names will be automatically prefixed with\nfeature.node.kubernetes.io/ but this will change in a future version (see\nautoDefaultNs config option.\n\n\nNOTE: The
\nannotationsfield has will only advertise features via node\nannotations the features won’t be advertised as node labels unless they are\nspecified in thelabelsfield.
taints is a list of taint entries and each entry can have key, value and effect,\nwhere the value is optional. Effect could be NoSchedule, PreferNoSchedule\nor NoExecute. To learn more about the meaning of these effects, check out k8s documentation.
Example NodeFeatureRule with taints:
\n\napiVersion: nfd.k8s-sigs.io/v1alpha1\nkind: NodeFeatureRule\nmetadata:\n name: my-sample-rule-object\nspec:\n rules:\n - name: \"my sample taint rule\"\n taints:\n - effect: PreferNoSchedule\n key: \"feature.node.kubernetes.io/special-node\"\n value: \"true\"\n - effect: NoExecute\n key: \"feature.node.kubernetes.io/dedicated-node\"\n matchFeatures:\n - feature: kernel.loadedmodule\n matchExpressions:\n dummy: {op: Exists}\n - feature: kernel.config\n matchExpressions:\n X86: {op: In, value: [\"y\"]}\nIn this example, if the my sample taint rule rule is matched,\nfeature.node.kubernetes.io/pci-0300_1d0f.present=true:NoExecute\nand feature.node.kubernetes.io/cpu-cpuid.ADX:NoExecute taints are set on the node.
There are some limitations to the namespace part (i.e. prefix/) of the taint\nkey:
\n\nkubernetes.io/ and its sub-namespaces (like sub.ns.kubernetes.io/) cannot\ngenerally be usedfeature.node.kubernetes.io/ and its sub-namespaces\n(like sub.ns.feature.node.kubernetes.io)foo) keys are disallowed\n\n\nNOTE: taints field is not available for the custom rules of nfd-worker\nand only for NodeFeatureRule objects.
\n
The .vars field is a map of values (key-value pairs) to store for subsequent\nrules to use. In other words, these are variables that are not advertised as\nnode labels. See backreferences for more details on the\nusage of vars.
The .extendedResources field is a list of extended resources to advertise.\nSee extended resources for more details.
Take this rule as a referential example:
\n\napiVersion: nfd.k8s-sigs.io/v1alpha1\nkind: NodeFeatureRule\nmetadata:\n name: my-extended-resource-rule\nspec:\n rules:\n - name: \"my extended resource rule\"\n extendedResources:\n vendor.io/dynamic: \"@kernel.version.major\"\n vendor.io/static: \"123\"\n matchFeatures:\n - feature: kernel.version\n matchExpressions:\n major: {op: Exists}\nThe extended resource vendor.io/dynamic is defined in the form @feature.attribute.\nThe value of the extended resource will be the value of the attribute major\nof the feature kernel.version.
The @<feature-name>.<element-name> format can be used to inject values of\ndetected features to the extended resource. See\navailable features for possible values to use. Note that\nthe value must be eligible as a\nKubernetes resource quantity.
This will yield into the following node status:
\n\n allocatable:\n ...\n vendor.io/dynamic: \"5\"\n vendor.io/static: \"123\"\n ...\n capacity:\n ...\n vendor.io/dynamic: \"5\"\n vendor.io/static: \"123\"\n ...\nThere are some limitations to the namespace part (i.e. prefix)/ of the Extended\nResources names:
\n\nkubernetes.io/ and its sub-namespaces (like sub.ns.kubernetes.io/) cannot\ngenerally be usedfeature.node.kubernetes.io/ and its sub-namespaces\n(like sub.ns.feature.node.kubernetes.io)my-er) site.version }} unprefixed names will be\nautomatically prefixed with feature.node.kubernetes.io/ but this will\nchange in a future version (see\nautoDefaultNs config option.\n\n\nNOTE:
\n.extendedResourcesis not supported by the\ncustom feature source – it can only be used in\nNodeFeatureRule objects.
The .varsTemplate field specifies a text template for dynamically creating\nvars based on the matched features. See templating for details\non using templates and backreferences for more details on\nthe usage of vars.
\n\n\nNOTE: The
\nvarsfield has priority overvarsTemplate, i.e.\nvars specified in thevarsfield will override anything originating from\nvarsTemplate.
The .matchFeatures field specifies a feature matcher, consisting of a list of\nfeature matcher terms. It implements a logical AND over the terms i.e. all\nof them must match for the rule to trigger.
matchFeatures:\n - feature: <feature-name>\n matchExpressions:\n <key>:\n op: <op>\n value:\n - <value-1>\n - ...\n matchName:\n op: <op>\n value:\n - <value-1>\n - ...\nThe .matchFeatures[].feature field specifies the feature which to evaluate.
\n\n\nNOTE:If both
\nmatchExpressionsand\nmatchNameare specified, they both must match.
The .matchFeatures[].matchExpressions field is used to match against the\nvalue(s) of a feature. The matchExpressions field consists of a set of\nexpressions, each of which is evaluated against all elements of the specified\nfeature.
matchExpressions:\n <key>:\n op: <op>\n value:\n - <value-1>\n - ...\nIn each MatchExpression the key specifies the name of of the feature element\n(flag and attribute features) or name of the attribute (instance\nfeatures) which to look for. The behavior of MatchExpression depends on the\nfeature type:
<key>The op field specifies the operator to apply. Valid values are described\nbelow.
| Operator | \nNumber of values | \nMatches when | \n
|---|---|---|
In | \n 1 or greater | \nInput is equal to one of the values | \n
NotIn | \n 1 or greater | \nInput is not equal to any of the values | \n
InRegexp | \n 1 or greater | \nValues of the MatchExpression are treated as regexps and input matches one or more of them | \n
Exists | \n 0 | \nThe key exists | \n
DoesNotExist | \n 0 | \nThe key does not exists | \n
Gt | \n 1 | \nInput is greater than the value. Both the input and value must be integer numbers. | \n
Lt | \n 1 | \nInput is less than the value. Both the input and value must be integer numbers. | \n
GtLt | \n 2 | \nInput is between two values. Both the input and value must be integer numbers. | \n
IsTrue | \n 0 | \nInput is equal to “true” | \n
IsFalse | \n 0 | \nInput is equal “false” | \n
The value field of MatchExpression is a list of string arguments to the\noperator.
The .matchFeatures[].matchName field is used to match against the\nname(s) of a feature (whereas the matchExpressions field\nmatches against the value(s). The matchName field consists of a single\nexpression which is evaulated against the name of each element of the specified\nfeature.
matchName:\n op: <op>\n value:\n - <value-1>\n - ...\nThe behavior of matchName depends on the feature type:
The op field specifies the operator to apply. Same operators as for\nmatchExpressions above are available.
| Operator | \nNumber of values | \nMatches | \n
|---|---|---|
In | \n 1 or greater | \nAll name is equal to one of the values | \n
NotIn | \n 1 or greater | \nAll name that is not equal to any of the values | \n
InRegexp | \n 1 or greater | \nAll name that matches any of the values (treated as regexps) | \n
Exists | \n 0 | \nAll elements | \n
Other operators are not practical with matchName (DoesNotExist never\nmatches; Gt,Lt and GtLt are only usable if feature names are integers;\nIsTrue and IsFalse are only usable if the feature name is true or\nfalse).
The value field is a list of string arguments to the operator.
An example:
\n\n matchFeatures:\n - feature: cpu.cpuid\n matchName: {op: InRegexp, value: [\"^AVX\"]}\nThe snippet above would match if any CPUID feature starting with AVX is present\n(e.g. AVX1 or AVX2 or AVX512F etc).
\n\nThe .matchAny field is a list of of matchFeatures\nmatchers. A logical OR is applied over the matchers, i.e. at least one of them\nmust match for the rule to trigger.
Consider the following example:
\n\n matchAny:\n - matchFeatures:\n - feature: kernel.loadedmodule\n matchExpressions:\n kmod-1: {op: Exists}\n - feature: pci.device\n matchExpressions:\n vendor: {op: In, value: [\"0eee\"]}\n class: {op: In, value: [\"0200\"]}\n - matchFeatures:\n - feature: kernel.loadedmodule\n matchExpressions:\n kmod-2: {op: Exists}\n - feature: pci.device\n matchExpressions:\n vendor: {op: In, value: [\"0fff\"]}\n class: {op: In, value: [\"0200\"]}\nThis matches if kernel module kmod-1 is loaded and a network controller from\nvendor 0eee is present, OR, if kernel module kmod-2 has been loaded and a\nnetwork controller from vendor 0fff is present (OR both of these conditions are\ntrue).
\n\nThe following features are available for matching:
\n\n| Feature | \nFeature types | \nElements | \nValue type | \nDescription | \n
|---|---|---|---|---|
cpu.cpuid | \n flag | \n\n | \n | Supported CPU capabilities | \n
| \n | \n | <cpuid-flag> | \n \n | CPUID flag is present | \n
| \n | attribute | \n\n | \n | CPU capability attributes | \n
| \n | \n | AVX10_VERSION | \nint | \nAVX10 vector ISA version (if supported) | \n
cpu.cstate | \n attribute | \n\n | \n | Status of cstates in the intel_idle cpuidle driver | \n
| \n | \n | enabled | \n bool | \n‘true’ if cstates are set, otherwise ‘false’. Does not exist of intel_idle driver is not active. | \n
cpu.model | \n attribute | \n\n | \n | CPU model related attributes | \n
| \n | \n | family | \n int | \nCPU family | \n
| \n | \n | vendor_id | \n string | \nCPU vendor ID | \n
| \n | \n | id | \n int | \nCPU model ID | \n
cpu.pstate | \n attribute | \n\n | \n | State of the Intel pstate driver. Does not exist if the driver is not enabled. | \n
| \n | \n | status | \n string | \nStatus of the driver, possible values are ‘active’ and ‘passive’ | \n
| \n | \n | turbo | \n bool | \n‘true’ if turbo frequencies are enabled, otherwise ‘false’ | \n
| \n | \n | scaling | \n string | \nActive scaling_governor, possible values are ‘powersave’ or ‘performance’. | \n
cpu.rdt | \n attribute | \n\n | \n | Intel RDT capabilities supported by the system | \n
| \n | \n | <rdt-flag> | \n \n | RDT capability is supported, see RDT flags for details | \n
| \n | \n | RDTL3CA_NUM_CLOSID | \n int | \nThe number or available CLOSID (Class of service ID) for Intel L3 Cache Allocation Technology | \n
cpu.security | \n attribute | \n\n | \n | Features related to security and trusted execution environments | \n
| \n | \n | sgx.enabled | \n bool | \ntrue if Intel SGX (Software Guard Extensions) has been enabled, otherwise does not exist | \n
| \n | \n | sgx.epc | \n int | \nThe total amount Intel SGX Encrypted Page Cache memory in bytes. It’s only present if sgx.enabled is true. | \n
| \n | \n | se.enabled | \n bool | \ntrue if IBM Secure Execution for Linux is available and has been enabled, otherwise does not exist | \n
| \n | \n | tdx.enabled | \n bool | \ntrue if Intel TDX (Trusted Domain Extensions) is available on the host and has been enabled, otherwise does not exist | \n
| \n | \n | tdx.total_keys | \n int | \nThe total amount of keys an Intel TDX (Trusted Domain Extensions) host can provide. It’s only present if tdx.enabled is true. | \n
| \n | \n | tdx.protected | \n bool | \ntrue if a guest VM was started using Intel TDX (Trusted Domain Extensions), otherwise does not exist. | \n
| \n | \n | sev.enabled | \n bool | \ntrue if AMD SEV (Secure Encrypted Virtualization) is available on the host and has been enabled, otherwise does not exist | \n
| \n | \n | sev.es.enabled | \n bool | \ntrue if AMD SEV-ES (Encrypted State supported) is available on the host and has been enabled, otherwise does not exist | \n
| \n | \n | sev.snp.enabled | \n bool | \ntrue if AMD SEV-SNP (Secure Nested Paging supported) is available on the host and has been enabled, otherwise does not exist | \n
| \n | \n | sev.asids | \n int | \nThe total amount of AMD SEV address-space identifiers (ASIDs), based on the /sys/fs/cgroup/misc.capacity information. | \n
| \n | \n | sev.encrypted_state_ids | \n int | \nThe total amount of AMD SEV-ES and SEV-SNP supported, based on the /sys/fs/cgroup/misc.capacity information. | \n
cpu.sst | \n attribute | \n\n | \n | Intel SST (Speed Select Technology) capabilities | \n
| \n | \n | bf.enabled | \n bool | \ntrue if Intel SST-BF (Intel Speed Select Technology - Base frequency) has been enabled, otherwise does not exist | \n
cpu.topology | \n attribute | \n\n | \n | CPU topology related features | \n
| \n | \n | hardware_multithreading | \n bool | \nHardware multithreading, such as Intel HTT, is enabled | \n
| \n | \n | socket_count | \n int | \nNumber of CPU Sockets | \n
cpu.coprocessor | \n attribute | \n\n | \n | CPU Coprocessor related features | \n
| \n | \n | nx_gzip | \n bool | \nNest Accelerator GZIP support is enabled | \n
kernel.config | \n attribute | \n\n | \n | Kernel configuration options | \n
| \n | \n | <config-flag> | \n string | \nValue of the kconfig option | \n
kernel.loadedmodule | \n flag | \n\n | \n | Kernel modules loaded on the node as reported by /proc/modules | \n
kernel.enabledmodule | \n flag | \n\n | \n | Kernel modules loaded on the node and available as built-ins as reported by modules.builtin | \n
| \n | \n | mod-name | \n \n | Kernel module <mod-name> is loaded | \n
kernel.selinux | \n attribute | \n\n | \n | Kernel SELinux related features | \n
| \n | \n | enabled | \n bool | \ntrue if SELinux has been enabled and is in enforcing mode, otherwise false | \n
kernel.version | \n attribute | \n\n | \n | Kernel version information | \n
| \n | \n | full | \n string | \nFull kernel version (e.g. ‘4.5.6-7-g123abcde’) | \n
| \n | \n | major | \n int | \nFirst component of the kernel version (e.g. ‘4’) | \n
| \n | \n | minor | \n int | \nSecond component of the kernel version (e.g. ‘5’) | \n
| \n | \n | revision | \n int | \nThird component of the kernel version (e.g. ‘6’) | \n
local.label | \n attribute | \n\n | \n | Labels from feature files, i.e. labels from the local feature source | \n
local.feature | \n attribute | \n\n | \n | Features from feature files, i.e. features from the local feature source | \n
| \n | \n | <label-name> | \n string | \nLabel <label-name> created by the local feature source, value equals the value of the label | \n
memory.nv | \n instance | \n\n | \n | NVDIMM devices present in the system | \n
| \n | \n | <sysfs-attribute> | \n string | \nValue of the sysfs device attribute, available attributes: devtype, mode | \n
memory.numa | \n attribute | \n\n | \n | NUMA nodes | \n
| \n | \n | is_numa | \n bool | \ntrue if NUMA architecture, false otherwise | \n
| \n | \n | node_count | \n int | \nNumber of NUMA nodes | \n
memory.swap | \n attribute | \n\n | \n | Swap enabled on node | \n
| \n | \n | enabled | \n bool | \ntrue if swap partition detected, false otherwise | \n
network.device | \n instance | \n\n | \n | Physical (non-virtual) network interfaces present in the system | \n
| \n | \n | name | \n string | \nName of the network interface | \n
| \n | \n | <sysfs-attribute> | \n string | \nSysfs network interface attribute, available attributes: operstate, speed, sriov_numvfs, sriov_totalvfs | \n
network.virtual | \n instance | \n\n | \n | Virtual network interfaces present in the system | \n
| \n | \n | name | \n string | \nName of the network interface | \n
| \n | \n | <sysfs-attribute> | \n string | \nSysfs network interface attribute, available attributes: operstate, speed | \n
pci.device | \n instance | \n\n | \n | PCI devices present in the system | \n
| \n | \n | <sysfs-attribute> | \n string | \nValue of the sysfs device attribute, available attributes: class, vendor, device, subsystem_vendor, subsystem_device, sriov_totalvfs, iommu_group/type, iommu/intel-iommu/version | \n
storage.block | \n instance | \n\n | \n | Block storage devices present in the system | \n
| \n | \n | name | \n string | \nName of the block device | \n
| \n | \n | <sysfs-attribute> | \n string | \nSysfs network interface attribute, available attributes: dax, rotational, nr_zones, zoned | \n
system.osrelease | \n attribute | \n\n | \n | System identification data from /etc/os-release | \n
| \n | \n | <parameter> | \n string | \nOne parameter from /etc/os-release | \n
system.dmiid | \n attribute | \n\n | \n | DMI identification data from /sys/devices/virtual/dmi/id/ | \n
| \n | \n | sys_vendor | \n string | \nVendor name from /sys/devices/virtual/dmi/id/sys_vendor | \n
| \n | \n | product_name | \n string | \nProduct name from /sys/devices/virtual/dmi/id/product_name | \n
system.name | \n attribute | \n\n | \n | System name information | \n
| \n | \n | nodename | \n string | \nName of the kubernetes node object | \n
usb.device | \n instance | \n\n | \n | USB devices present in the system | \n
| \n | \n | <sysfs-attribute> | \n string | \nValue of the sysfs device attribute, available attributes: class, vendor, device, serial | \n
rule.matched | \n attribute | \n\n | \n | Previously matched rules | \n
| \n | \n | <label-or-var> | \n string | \nLabel or var from a preceding rule that matched | \n
| Flag | \nDescription | \n
|---|---|
| RDTMON | \nIntel RDT Monitoring Technology | \n
| RDTCMT | \nIntel Cache Monitoring (CMT) | \n
| RDTMBM | \nIntel Memory Bandwidth Monitoring (MBM) | \n
| RDTL3CA | \nIntel L3 Cache Allocation Technology | \n
| RDTl2CA | \nIntel L2 Cache Allocation Technology | \n
| RDTMBA | \nIntel Memory Bandwidth Allocation (MBA) Technology | \n
Rules support template-based creation of labels and vars with the\n.labelsTemplate and .varsTemplate fields. These makes it possible to\ndynamically generate labels and vars based on the features that matched.
The template must expand into a simple format with <key>=<value> pairs\nseparated by newline.
Consider the following example:\n
\n\n labelsTemplate: |\n {{ range .pci.device }}vendor-{{ .class }}-{{ .device }}.present=true\n {{ end }}\n matchFeatures:\n - feature: pci.device\n matchExpressions:\n class: {op: InRegexp, value: [\"^02\"]}\n vendor: [\"0fff\"]\nThe rule above will create individual labels\nfeature.node.kubernetes.io/vendor-<class-id>-<device-id>.present=true for\neach network controller device (device class starting with 02) from vendor\n0fff.
All the matched features of each feature matcher term under matchFeatures\nfields are available for the template engine. Matched features can be\nreferenced with {{ .<feature-name> }} in the template, and\nthe available data could be described in yaml as follows:
.\n <key-feature>:\n - Name: <matched-key>\n - ...\n\n <value-feature>:\n - Name: <matched-key>\n Value: <matched-value>\n - ...\n\n <instance-feature>:\n - <attribute-1-name>: <attribute-1-value>\n <attribute-2-name>: <attribute-2-value>\n ...\n - ...\nThat is, the per-feature data is a list of objects whose data fields depend on\nthe type of the feature:
\n\nA simple example of a template utilizing name and value from an attribute\nfeature:\n
\n\n labelsTemplate: |\n {{ range .system.osrelease }}system-{{ .Name }}={{ .Value }}\n {{ end }}\n matchFeatures:\n - feature: system.osRelease\n matchExpressions:\n ID: {op: Exists}\n VERSION_ID.major: {op: Exists}\n\n\n\nNOTE:If both
\nmatchExpressionsandmatchNamefor a feature matcher\nterm (seematchFeatures) is specified, the list of\nmatched features (for the template engine) is the union from both of these.\n\nNOTE: In case of matchAny is specified, the template is executed\nseparately against each individualmatchFeaturesfield and the final set of\nlabels will be superset of all these separate template expansions. E.g.\nconsider the following:
- name: <name>\n labelsTemplate: <template>\n matchFeatures: <matcher#1>\n matchAny:\n - matchFeatures: <matcher#2>\n - matchFeatures: <matcher#3>\nIn the example above (assuming the overall result is a match) the template\nwould be executed on matcher#1 as well as on matcher#2 and/or matcher#3\n(depending on whether both or only one of them match). All the labels from\nthese separate expansions would be created, i.e. the end result would be a\nunion of all the individual expansions.
\n\nRule templates use the Golang text/template\npackage and all its built-in functionality (e.g. pipelines and functions) can\nbe used. An example template taking use of the built-in len function,\nadvertising the number of PCI network controllers from a specific vendor:\n
labelsTemplate: |\n num-intel-network-controllers={{ .pci.device | len }}\n matchFeatures:\n - feature: pci.device\n matchExpressions:\n vendor: {op: In, value: [\"8086\"]}\n class: {op: In, value: [\"0200\"]}\n\nImaginative template pipelines are possible, but care must be taken to\nproduce understandable and maintainable rule sets.
\n\nRules support referencing the output of preceding rules. This enables\nsophisticated scenarios where multiple rules are combined together\nto for more complex heuristics than a single rule can provide. The labels and\nvars created by the execution of preceding rules are available as a special\nrule.matched feature.
Consider the following configuration:
\n\n - name: \"my kernel label rule\"\n labels:\n kernel-feature: \"true\"\n matchFeatures:\n - feature: kernel.version\n matchExpressions:\n major: {op: Gt, value: [\"4\"]}\n\n - name: \"my var rule\"\n vars:\n nolabel-feature: \"true\"\n matchFeatures:\n - feature: cpu.cpuid\n matchExpressions:\n AVX512F: {op: Exists}\n - feature: pci.device\n matchExpressions:\n vendor: {op: In, value: [\"0fff\"]}\n device: {op: In, value: [\"1234\", \"1235\"]}\n\n - name: \"my high level feature rule\"\n labels:\n high-level-feature: \"true\"\n matchFeatures:\n - feature: rule.matched\n matchExpressions:\n kernel-feature: {op: IsTrue}\n nolabel-feature: {op: IsTrue}\nThe feature.node.kubernetes.io/high-level-feature = true label depends on the\ntwo previous rules.
Note that when referencing rules across multiple\nNodeFeatureRule objects attention must be\npaid to the ordering. NodeFeatureRule objects are processed in alphabetical\norder (based on their .metadata.name).
Some more configuration examples below.
\n\nMatch certain CPUID features:
\n\n - name: \"example cpuid rule\"\n labels:\n my-special-cpu-feature: \"true\"\n matchFeatures:\n - feature: cpu.cpuid\n matchExpressions:\n AESNI: {op: Exists}\n AVX: {op: Exists}\nRequire a certain loaded kernel module and OS version:
\n\n - name: \"my multi-feature rule\"\n labels:\n my-special-multi-feature: \"true\"\n matchFeatures:\n - feature: kernel.loadedmodule\n matchExpressions:\n e1000: {op: Exists}\n - feature: system.osrelease\n matchExpressions:\n NAME: {op: InRegexp, values: [\"^openSUSE\"]}\n VERSION_ID.major: {op: Gt, values: [\"14\"]}\nRequire a loaded kernel module and two specific PCI devices (both of which\nmust be present):
\n\n - name: \"my multi-device rule\"\n labels:\n my-multi-device-feature: \"true\"\n matchFeatures:\n - feature: kernel.loadedmodule\n matchExpressions:\n my-driver-module: {op: Exists}\n - pci.device:\n vendor: \"0fff\"\n device: \"1234\"\n - pci.device:\n vendor: \"0fff\"\n device: \"abcd\"\nNode Feature Discovery follows semantic versioning where\nthe version number consists of three components, i.e. MAJOR.MINOR.PATCH.
\n\nThe most recent two minor releases (or release branches) of Node Feature\nDiscovery are supported. That is, with X being the latest release, X and X-1\nare supported and X-1 reaches end-of-life when X+1 is released.
\n\nBuilt-in feature labels and\nfeatures are supported\nfor 2 releases after being deprecated, at minimum. That is, if a feature label\nis deprecated in version X, it will be supported in X+1 and X+2 and\nmay be dropped in X+3.
\n\nCommand-line flags and configuration file options are supported for 1 more\nrelease after being deprecated, at minimum. That is, if option/flag is\ndeprecated in version X, it will be supported in X+1 and may be removed\nin X+2.
\n\nThe same policy (support for 1 release after deprecation) also applies to Helm\nchart parameters.
\n\nNode Feature Discovery is compatible with Kubernetes v1.24 and later.
\n","dir":"/reference/","name":"versions.md","path":"reference/versions.md","url":"/reference/versions.html"},{"title":"Examples and demos","layout":"default","sort":9,"content":"This page contains usage examples and demos.
\n\n
A demo on the benefits of using node feature discovery can be found in the\nsource code repository under\ndemo/.
\n","dir":"/usage/","name":"examples-and-demos.md","path":"usage/examples-and-demos.md","url":"/usage/examples-and-demos.html"},{"title":"Feature Gates","layout":"default","sort":10,"content":"Feature gates are a set of key-value pairs that control the behavior of NFD.\nThey are used to enable or disable certain features of NFD.\nThe feature gates are set using the -feature-gates command line flag or\nfeatureGates value in the Helm chart. The following feature gates are available:
| Name | \nDefault | \nStage | \nSince | \nUntil | \n
|---|---|---|---|---|
NodeFeatureAPI | \n true | \nBeta | \nV0.14 | \nv0.16 | \n
NodeFeatureAPI | \n true | \nGA | \nV0.17 | \n\n |
DisableAutoPrefix | \n false | \nAlpha | \nV0.16 | \n\n |
NodeFeatureGroupAPI | \n false | \nAlpha | \nV0.16 | \n\n |
The NodeFeatureAPI feature gate enables the Node Feature API.\nWhen enabled, NFD will register the Node Feature API with the Kubernetes API\nserver. The Node Feature API is used to expose node-specific hardware and\nsoftware features to the Kubernetes scheduler. The Node Feature API is a beta\nfeature and is enabled by default.
The NodeFeatureGroupAPI feature gate enables the Node Feature Group API.\nWhen enabled, NFD will register the Node Feature Group API with the Kubernetes API\nserver. The Node Feature Group API is used to create node groups based on\nhardware and software features. The Node Feature Group API is an alpha feature\nand is disabled by default.
The DisableAutoPrefix feature gate controls the automatic prefixing of names.\nWhen enabled nfd-master does not automatically add the default\nfeature.node.kubernetes.io/ prefix to unprefixed labels, annotations and\nextended resources. Automatic prefixing is the default behavior in NFD v0.16\nand earlier.
Note that enabling the feature gate effectively causes unprefixed names to be\nfiltered out as NFD does not allow unprefixed names of labels, annotations or\nextended resources. For example, with the DisableAutoPrefix feature gate set\nto false, a NodeFeatureRule with
labels:\n foo: bar\nwill turn into feature.node.kubernetes.io/foo=bar node label. With\nDisableAutoPrefix set to true, no prefix is added and the label will be\nfiltered out.
Note that taint keys are not affected by this feature gate.
\n","dir":"/reference/","name":"feature-gates.md","path":"reference/feature-gates.md","url":"/reference/feature-gates.html"},{"title":"Kubectl plugin","layout":"default","sort":10,"content":"\n\n\nDeveloper Preview This feature is currently in developer preview and\nsubject to change. It is not recommended to use it in production\nenvironments.
\n
The kubectl plugin kubectl nfd can be used to validate/dryrun and test\nNodeFeatureRule objects. It can be installed with the following command:
git clone https://github.com/kubernetes-sigs/node-feature-discovery\ncd node-feature-discovery\nmake build-kubectl-nfd\nKUBECTL_PATH=/usr/local/bin/\nmv ./bin/kubectl-nfd ${KUBECTL_PATH}\nThe plugin can be used to validate a NodeFeatureRule object:
\n\nkubectl nfd validate -f <nodefeaturerule.yaml>\nThe plugin can be used to test a NodeFeatureRule object against a node:
\n\nkubectl nfd test -f <nodefeaturerule.yaml> -n <node-name>\nThe plugin can be used to DryRun a NodeFeatureRule object against a NodeFeature\nfile:
\n\nkubectl get -n node-feature-discovery nodefeature <nodename> -o yaml > <nodefeature.yaml>\nkubectl nfd dryrun -f <nodefeaturerule.yaml> -n <nodefeature.yaml>\nOr you can use the example NodeFeature file(it is a minimal NodeFeature file):
\n\n$ kubectl nfd dryrun -f examples/nodefeaturerule.yaml -n examples/nodefeature.yaml\nEvaluating NodeFeatureRule \"examples/nodefeaturerule.yaml\" against NodeFeature \"examples/nodefeature.yaml\"\nProcessing rule: my sample rule\n*** Labels ***\nvendor.io/my-sample-feature=true\nNodeFeatureRule \"examples/nodefeaturerule.yaml\" is valid for NodeFeature \"examples/nodefeature.yaml\"\nNFD offers two variants of the container image. Released container images are\navailable for x86_64 and Arm64 architectures.
\n\nThe default is a minimal image based on\nscratch\nand only supports running statically linked binaries.
\n\nFor backwards compatibility a container image tag with suffix -minimal\n(e.g. gcr.io/k8s-staging-nfd/node-feature-discovery:master-minimal) is provided.
This image is based on debian:bookworm-slim\nand contains a full Linux system for doing live debugging and diagnosis\nof the NFD images.
\n\nThe container image tag has suffix -full\n(e.g. gcr.io/k8s-staging-nfd/node-feature-discovery:master-full).
Welcome to Node Feature Discovery – a Kubernetes add-on for detecting hardware\nfeatures and system configuration!
\n\nContinue to:
\n\nIntroduction for more details on the\nproject.
\nQuick start for quick step-by-step\ninstructions on how to get NFD running on your cluster.
\n$ kubectl apply -k https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default?ref=master\n namespace/node-feature-discovery created\n serviceaccount/nfd-master created\n clusterrole.rbac.authorization.k8s.io/nfd-master created\n clusterrolebinding.rbac.authorization.k8s.io/nfd-master created\n configmap/nfd-worker-conf created\n deployment.apps/nfd-master created\n daemonset.apps/nfd-worker created\n\n$ kubectl -n node-feature-discovery get all\n NAME READY STATUS RESTARTS AGE\n pod/nfd-master-555458dbbc-sxg6w 1/1 Running 0 56s\n pod/nfd-worker-mjg9f 1/1 Running 0 17s\n...\n\n$ kubectl get nodes -o json | jq '.items[].metadata.labels'\n {\n \"kubernetes.io/arch\": \"amd64\",\n \"kubernetes.io/os\": \"linux\",\n \"feature.node.kubernetes.io/cpu-cpuid.ADX\": \"true\",\n \"feature.node.kubernetes.io/cpu-cpuid.AESNI\": \"true\",\n...\n\nThis software enables node feature discovery for Kubernetes. It detects\nhardware features available on each node in a Kubernetes cluster, and\nadvertises those features using node labels and optionally node extended\nresources, annotations and node taints. Node Feature Discovery is compatible\nwith any recent version of Kubernetes (v1.24+).
\n\nNFD consists of four software components:
\n\nNFD-Master is the daemon responsible for communication towards the Kubernetes\nAPI. That is, it receives labeling requests from the worker and modifies node\nobjects accordingly.
\n\nNFD-Worker is a daemon responsible for feature detection. It then communicates\nthe information to nfd-master which does the actual node labeling. One\ninstance of nfd-worker is supposed to be running on each node of the cluster,
\n\nNFD-Topology-Updater is a daemon responsible for examining allocated\nresources on a worker node to account for resources available to be allocated\nto new pod on a per-zone basis (where a zone can be a NUMA node). It then\ncreates or updates a\nNodeResourceTopology custom\nresource object specific to this node. One instance of nfd-topology-updater is\nsupposed to be running on each node of the cluster.
\n\nNFD-GC is a daemon responsible for cleaning obsolete\nNodeFeature and\nNodeResourceTopology objects.
\n\nOne instance of nfd-gc is supposed to be running in the cluster.
\n\nFeature discovery is divided into domain-specific feature sources:
\n\nEach feature source is responsible for detecting a set of features which. in\nturn, are turned into node feature labels. Feature labels are prefixed with\nfeature.node.kubernetes.io/ and also contain the name of the feature source.\nNon-standard user-specific feature labels can be created with the local and\ncustom feature sources.
An overview of the default feature labels:
\n\n{\n \"feature.node.kubernetes.io/cpu-<feature-name>\": \"true\",\n \"feature.node.kubernetes.io/custom-<feature-name>\": \"true\",\n \"feature.node.kubernetes.io/kernel-<feature name>\": \"<feature value>\",\n \"feature.node.kubernetes.io/memory-<feature-name>\": \"true\",\n \"feature.node.kubernetes.io/network-<feature-name>\": \"true\",\n \"feature.node.kubernetes.io/pci-<device label>.present\": \"true\",\n \"feature.node.kubernetes.io/storage-<feature-name>\": \"true\",\n \"feature.node.kubernetes.io/system-<feature name>\": \"<feature value>\",\n \"feature.node.kubernetes.io/usb-<device label>.present\": \"<feature value>\",\n \"feature.node.kubernetes.io/<file name>-<feature name>\": \"<feature value>\"\n}\nNFD also annotates nodes it is running on:
\n\n| Annotation | \nDescription | \n
|---|---|
| [<instance>.]nfd.node.kubernetes.io/feature-labels | \nComma-separated list of node labels managed by NFD. NFD uses this internally so must not be edited by users. | \n
| [<instance>.]nfd.node.kubernetes.io/feature-annotations | \nComma-separated list of node annotations managed by NFD. NFD uses this internally so must not be edited by users. | \n
| [<instance>.]nfd.node.kubernetes.io/extended-resources | \nComma-separated list of node extended resources managed by NFD. NFD uses this internally so must not be edited by users. | \n
| [<instance>.]nfd.node.kubernetes.io/taints | \nComma-separated list of node taints managed by NFD. NFD uses this internally so must not be edited by users. | \n
\n\n\nNOTE: the
\n-instance\ncommand line flag affects the annotation names
Unapplicable annotations are not created, i.e. for example\nnfd.node.kubernetes.io/extended-resources is only placed if some extended\nresources were created by NFD.
NFD takes use of some Kubernetes Custom Resources.
\n\nNodeFeatures\nis be used for representing node features and requesting node labels to be\ngenerated.
\n\nNFD-Master uses NodeFeatureRules\nfor custom labeling of nodes.
\n\nNFD-Topology-Updater creates\nNodeResourceTopology objects\nthat describe the hardware topology of node resources.
\n","dir":"/get-started/","name":"introduction.md","path":"get-started/introduction.md","url":"/get-started/introduction.html"},{"title":"Master cmdline reference","layout":"default","sort":1,"content":"To quickly view available command line flags execute nfd-master -help.\nIn a docker container:
docker run gcr.io/k8s-staging-nfd/node-feature-discovery:master nfd-master -help\nPrint usage and exit.
\n\nPrint version and exit.
\n\nThe -feature-gates flag is used to enable or disable non GA features.\nThe list of available feature gates can be found in the feature gates documentation.
Example:
\n\nnfd-master -feature-gates NodeFeatureGroupAPI=true\nThe -prune flag is a sub-command like option for cleaning up the cluster. It\ncauses nfd-master to remove all NFD related labels, annotations and extended\nresources from all Node objects of the cluster and exit.
DEPRECATED: Will be removed in NFD v0.17 and replaced by -port.
The -metrics flag specifies the port on which to expose\nPrometheus metrics. Setting this to 0 disables the\nmetrics server on nfd-master.
Default: 8081
\n\nExample:
\n\nnfd-master -metrics=12345\nThe -instance flag makes it possible to run multiple NFD deployments in\nparallel. In practice, it separates the node annotations between deployments so\nthat each of them can store metadata independently. The instance name must\nstart and end with an alphanumeric character and may only contain alphanumeric\ncharacters, -, _ or ..
Default: empty
\n\nExample:
\n\nnfd-master -instance=network\nThe -enable-leader-election flag enables leader election for NFD-Master.\nIt is advised to turn on this flag when running more than one instance of\nNFD-Master.
Default: false
\n\nnfd-master -enable-leader-election\nThe -enable-taints flag enables/disables node tainting feature of NFD.
Default: false
\n\nExample:
\n\nnfd-master -enable-taints=true\nThe -no-publish flag disables updates to the Node objects in the Kubernetes\nAPI server, making a “dry-run” flag for nfd-master. No Labels, Annotations or\nExtendedResources of nodes are updated.
Default: false
\n\nExample:
\n\nnfd-master -no-publish\nThe -label-whitelist specifies a regular expression for filtering feature\nlabels based on their name. Each label must match against the given regular\nexpression or it will not be published.
\n\n\nNOTE: The regular expression is only matches against the “basename” part\nof the label, i.e. to the part of the name after ‘/’. The label namespace is\nomitted.
\n
Default: empty
\n\nExample:
\n\nnfd-master -label-whitelist='.*cpuid\\.'\nThe -extra-label-ns flag specifies a comma-separated list of allowed feature\nlabel namespaces. This option can be used to allow\nother vendor or application specific namespaces for custom labels from the\nlocal and custom feature sources, even though these labels were denied using\nthe deny-label-ns flag.
Default: empty
\n\nExample:
\n\nnfd-master -extra-label-ns=vendor-1.com,vendor-2.io\nThe -deny-label-ns flag specifies a comma-separated list of excluded\nlabel namespaces. By default, nfd-master allows creating labels in all\nnamespaces, excluding kubernetes.io namespace and its sub-namespaces\n(i.e. *.kubernetes.io). However, you should note that\nkubernetes.io and its sub-namespaces are always denied.\nFor example, nfd-master -deny-label-ns=\"\" would still disallow\nkubernetes.io and *.kubernetes.io.\nThis option can be used to exclude some vendors or application specific\nnamespaces.\nNote that the namespaces feature.node.kubernetes.io and profile.node.kubernetes.io\nand their sub-namespaces are always allowed and cannot be denied.
Default: empty
\n\nExample:
\n\nnfd-master -deny-label-ns=*.vendor.com,vendor-2.io\nThe -config flag specifies the path of the nfd-master configuration file to\nuse.
Default: /etc/kubernetes/node-feature-discovery/nfd-master.conf
\n\nExample:
\n\nnfd-master -config=/opt/nfd/master.conf\nThe -options flag may be used to specify and override configuration file\noptions directly from the command line. The required format is the same as in\nthe config file i.e. JSON or YAML. Configuration options specified via this\nflag will override those from the configuration file:
Default: empty
\n\nExample:
\n\nnfd-master -options='{\"noPublish\": true}'\nThe -nfd-api-parallelism flag can be used to specify the maximum\nnumber of concurrent node updates.
Default: 10
\n\nExample:
\n\nnfd-master -nfd-api-parallelism=1\nThe following logging-related flags are inherited from the\nklog package.
\n\nIf true, adds the file directory to the header of the log messages.
\n\nDefault: false
\n\nLog to standard error as well as files.
\n\nDefault: false
\n\nWhen logging hits line file:N, emit a stack trace.
\n\nDefault: empty
\n\nIf non-empty, write log files in this directory.
\n\nDefault: empty
\n\nIf non-empty, use this log file.
\n\nDefault: empty
\n\nDefines the maximum size a log file can grow to. Unit is megabytes. If the\nvalue is 0, the maximum file size is unlimited.
\n\nDefault: 1800
\n\nLog to standard error instead of files
\n\nDefault: true
\n\nIf true, avoid header prefixes in the log messages.
\n\nDefault: false
\n\nIf true, avoid headers when opening log files.
\n\nDefault: false
\n\nLogs at or above this threshold go to stderr.
\n\nDefault: 2
\n\nNumber for the log level verbosity.
\n\nDefault: 0
\n\nComma-separated list of pattern=N settings for file-filtered logging.
Default: empty
\n\nThe -resync-period flag specifies the NFD API controller resync period.\nThe resync means nfd-master replaying all NodeFeature and NodeFeatureRule objects,\nthus effectively re-syncing all nodes in the cluster (i.e. ensuring labels, annotations,\nextended resources and taints are in place).
Default: 1 hour.
\n\nExample:
\n\nnfd-master -resync-period=2h\nFeatures are advertised as labels in the Kubernetes Node object.
\n\nLabel creation in nfd-worker is performed by a set of separate modules called\nlabel sources. The\ncore.labelSources\nconfiguration option (or\n-label-sources\nflag) of nfd-worker controls which sources to enable for label generation.
All built-in labels use the feature.node.kubernetes.io label namespace and\nhave the following format.
feature.node.kubernetes.io/<feature> = <value>\n\n\n\nNOTE: Consecutive runs of nfd-worker will update the labels on a given\nnode. If features are not discovered on a consecutive run, the corresponding\nlabel will be removed. This includes any restrictions placed on the\nconsecutive run, such as restricting discovered features with the\n
\n-label-whitelist\nflag of nfd-master or\ncore.labelWhiteList\noption of nfd-worker.
| Feature name | \nValue | \nDescription | \n
|---|---|---|
cpu-cpuid.<cpuid-flag> | \n true | \nCPU capability is supported. NOTE: the capability might be supported but not enabled. | \n
cpu-cpuid.<cpuid-attribute> | \n string | \nCPU attribute value | \n
cpu-hardware_multithreading | \n true | \nHardware multithreading, such as Intel HTT, enabled (number of logical CPUs is greater than physical CPUs) | \n
cpu-coprocessor.nx_gzip | \n true | \nNest Accelerator for GZIP is supported(Power). | \n
cpu-power.sst_bf.enabled | \n true | \nIntel SST-BF (Intel Speed Select Technology - Base frequency) enabled | \n
cpu-pstate.status | \n string | \nThe status of the Intel pstate driver when in use and enabled, either ‘active’ or ‘passive’. | \n
cpu-pstate.turbo | \n bool | \nSet to ‘true’ if turbo frequencies are enabled in Intel pstate driver, set to ‘false’ if they have been disabled. | \n
cpu-pstate.scaling_governor | \n string | \nThe value of the Intel pstate scaling_governor when in use, either ‘powersave’ or ‘performance’. | \n
cpu-cstate.enabled | \n bool | \nSet to ‘true’ if cstates are set in the intel_idle driver, otherwise set to ‘false’. Unset if intel_idle cpuidle driver is not active. | \n
cpu-security.sgx.enabled | \n true | \nSet to ‘true’ if Intel SGX is enabled in BIOS (based on a non-zero sum value of SGX EPC section sizes). | \n
cpu-security.se.enabled | \n true | \nSet to ‘true’ if IBM Secure Execution for Linux (IBM Z & LinuxONE) is available and enabled (requires /sys/firmware/uv/prot_virt_host facility) | \n
cpu-security.tdx.enabled | \n true | \nSet to ‘true’ if Intel TDX is available on the host and has been enabled (requires /sys/module/kvm_intel/parameters/tdx). | \n
cpu-security.tdx.protected | \n true | \nSet to ‘true’ if Intel TDX was used to start the guest node, based on the existence of the “TDX_GUEST” information as part of cpuid features. | \n
cpu-security.sev.enabled | \n true | \nSet to ‘true’ if ADM SEV is available on the host and has been enabled (requires /sys/module/kvm_amd/parameters/sev). | \n
cpu-security.sev.es.enabled | \n true | \nSet to ‘true’ if ADM SEV-ES is available on the host and has been enabled (requires /sys/module/kvm_amd/parameters/sev_es). | \n
cpu-security.sev.snp.enabled | \n true | \nSet to ‘true’ if ADM SEV-SNP is available on the host and has been enabled (requires /sys/module/kvm_amd/parameters/sev_snp). | \n
cpu-model.vendor_id | \n string | \nComparable CPU vendor ID. | \n
cpu-model.family | \n int | \nCPU family. | \n
cpu-model.id | \n int | \nCPU model number. | \n
The CPU label source is configurable, see\nworker configuration and\nsources.cpu\nconfiguration options for details.
| Flag | \nDescription | \n
|---|---|
| ADX | \nMulti-Precision Add-Carry Instruction Extensions (ADX) | \n
| AESNI | \nAdvanced Encryption Standard (AES) New Instructions (AES-NI) | \n
| APX_F | \nIntel Advanced Performance Extensions (APX) | \n
| AVX10 | \nIntel Advanced Vector Extensions 10 (AVX10) | \n
| AVX10_256, AVX10_512 | \nIntel AVX10 256-bit and 512-bit vector support | \n
| AVX | \nAdvanced Vector Extensions (AVX) | \n
| AVX2 | \nAdvanced Vector Extensions 2 (AVX2) | \n
| AVXIFMA | \nAVX-IFMA instructions | \n
| AVXVNNI | \nAVX (VEX encoded) VNNI neural network instructions | \n
| AMXBF16 | \nAdvanced Matrix Extension, tile multiplication operations on BFLOAT16 numbers | \n
| AMXINT8 | \nAdvanced Matrix Extension, tile multiplication operations on 8-bit integers | \n
| AMXFP16 | \nAdvanced Matrix Extension, tile multiplication operations on FP16 numbers | \n
| AMXFP8 | \nAdvanced Matrix Extension, tile multiplication operations on FP8 numbers | \n
| AMXTILE | \nAdvanced Matrix Extension, base tile architecture support | \n
| AVX512BF16 | \nAVX-512 BFLOAT16 instructions | \n
| AVX512BITALG | \nAVX-512 bit Algorithms | \n
| AVX512BW | \nAVX-512 byte and word Instructions | \n
| AVX512CD | \nAVX-512 conflict detection instructions | \n
| AVX512DQ | \nAVX-512 doubleword and quadword instructions | \n
| AVX512ER | \nAVX-512 exponential and reciprocal instructions | \n
| AVX512F | \nAVX-512 foundation | \n
| AVX512FP16 | \nAVX-512 FP16 instructions | \n
| AVX512IFMA | \nAVX-512 integer fused multiply-add instructions | \n
| AVX512PF | \nAVX-512 prefetch instructions | \n
| AVX512VBMI | \nAVX-512 vector bit manipulation instructions | \n
| AVX512VBMI2 | \nAVX-512 vector bit manipulation instructions, version 2 | \n
| AVX512VL | \nAVX-512 vector length extensions | \n
| AVX512VNNI | \nAVX-512 vector neural network instructions | \n
| AVX512VP2INTERSECT | \nAVX-512 intersect for D/Q | \n
| AVX512VPOPCNTDQ | \nAVX-512 vector population count doubleword and quadword | \n
| AVXNECONVERT | \nAVX-NE-CONVERT instructions | \n
| AVXVNNIINT8 | \nAVX-VNNI-INT8 instructions | \n
| AVXVNNIINT16 | \nAVX-VNNI-INT16 instructions | \n
| CMPCCXADD | \nCMPCCXADD instructions | \n
| ENQCMD | \nEnqueue Command | \n
| GFNI | \nGalois Field New Instructions | \n
| HYPERVISOR | \nRunning under hypervisor | \n
| MSRLIST | \nRead/Write List of Model Specific Registers | \n
| PREFETCHI | \nPREFETCHIT0/1 instructions | \n
| VAES | \nAVX-512 vector AES instructions | \n
| VPCLMULQDQ | \nCarry-less multiplication quadword | \n
| WRMSRNS | \nNon-Serializing Write to Model Specific Register | \n
By default, the following CPUID flags have been blacklisted: AVX10 (use\nAVX10_VERSION instead), BMI1, BMI2, CLMUL, CMOV, CX16, ERMS, F16C, HTT, LZCNT,\nMMX, MMXEXT, NX, POPCNT, RDRAND, RDSEED, RDTSCP, SGX, SSE, SSE2, SSE3, SSE4,\nSSE42, SSSE3 and TDX_GUEST. See\nsources.cpu\nconfiguration options to change the behavior.
See the full list in github.com/klauspost/cpuid.
\n\n| Attribute | \nDescription | \n
|---|---|
| AVX10_VERSION | \nAVX10 vector ISA version (if supported) | \n
| Flag | \nDescription | \n
|---|---|
| IDIVA | \nInteger divide instructions available in ARM mode | \n
| IDIVT | \nInteger divide instructions available in Thumb mode | \n
| THUMB | \nThumb instructions | \n
| FASTMUL | \nFast multiplication | \n
| VFP | \nVector floating point instruction extension (VFP) | \n
| VFPv3 | \nVector floating point extension v3 | \n
| VFPv4 | \nVector floating point extension v4 | \n
| VFPD32 | \nVFP with 32 D-registers | \n
| HALF | \nHalf-word loads and stores | \n
| EDSP | \nDSP extensions | \n
| NEON | \nNEON SIMD instructions | \n
| LPAE | \nLarge Physical Address Extensions | \n
| Flag | \nDescription | \n
|---|---|
| AES | \nAnnouncing the Advanced Encryption Standard | \n
| EVSTRM | \nEvent Stream Frequency Features | \n
| FPHP | \nHalf Precision(16bit) Floating Point Data Processing Instructions | \n
| ASIMDHP | \nHalf Precision(16bit) Asimd Data Processing Instructions | \n
| ATOMICS | \nAtomic Instructions to the A64 | \n
| ASIMRDM | \nSupport for Rounding Double Multiply Add/Subtract | \n
| PMULL | \nOptional Cryptographic and CRC32 Instructions | \n
| JSCVT | \nPerform Conversion to Match Javascript | \n
| DCPOP | \nPersistent Memory Support | \n
| Feature | \nValue | \nDescription | \n
|---|---|---|
kernel-config.<option> | \n true | \nKernel config option is enabled (set ‘y’ or ‘m’). Default options are NO_HZ, NO_HZ_IDLE, NO_HZ_FULL and PREEMPT | \n
kernel-selinux.enabled | \n true | \nSelinux is enabled on the node | \n
kernel-version.full | \n string | \nFull kernel version as reported by /proc/sys/kernel/osrelease (e.g. ‘4.5.6-7-g123abcde’) | \n
kernel-version.major | \n string | \nFirst component of the kernel version (e.g. ‘4’) | \n
kernel-version.minor | \n string | \nSecond component of the kernel version (e.g. ‘5’) | \n
kernel-version.revision | \n string | \nThird component of the kernel version (e.g. ‘6’) | \n
The kernel label source is configurable, see\nworker configuration and\nsources.kernel\nconfiguration options for details.
| Feature | \nValue | \nDescription | \n
|---|---|---|
memory-numa | \n true | \nMultiple memory nodes i.e. NUMA architecture detected | \n
memory-nv.present | \n true | \nNVDIMM device(s) are present | \n
memory-nv.dax | \n true | \nNVDIMM region(s) configured in DAX mode are present | \n
memory-swap.enabled | \n true | \nSwap is enabled on the node | \n
| Feature | \nValue | \nDescription | \n
|---|---|---|
network-sriov.capable | \n true | \nSingle Root Input/Output Virtualization (SR-IOV) enabled Network Interface Card(s) present | \n
network-sriov.configured | \n true | \nSR-IOV virtual functions have been configured | \n
| Feature | \nValue | \nDescription | \n
|---|---|---|
pci-<device label>.present | \n true | \nPCI device is detected | \n
pci-<device label>.sriov.capable | \n true | \nSingle Root Input/Output Virtualization (SR-IOV) enabled PCI device present | \n
| \n | \n | \n |
<device label> is format is configurable and set to <class>_<vendor> by\ndefault. For more more details about configuration of the pci labels, see\nsources.pci options\nand worker configuration\ninstructions.
| Feature | \nValue | \nDescription | \n
|---|---|---|
usb-<device label>.present | \n true | \nUSB device is detected | \n
<device label> is format is configurable and set to\n<class>_<vendor>_<device> by default. For more more details about\nconfiguration of the usb labels, see\nsources.usb options\nand worker configuration\ninstructions.
| Feature | \nValue | \nDescription | \n
|---|---|---|
storage-nonrotationaldisk | \n true | \nNon-rotational disk, like SSD, is present in the node | \n
| Feature | \nValue | \nDescription | \n
|---|---|---|
system-os_release.ID | \n string | \nOperating system identifier | \n
system-os_release.VERSION_ID | \n string | \nOperating system version identifier (e.g. ‘6.7’) | \n
system-os_release.VERSION_ID.major | \n string | \nFirst component of the OS version id (e.g. ‘6’) | \n
system-os_release.VERSION_ID.minor | \n string | \nSecond component of the OS version id (e.g. ‘7’) | \n
The custom label source is designed for creating\nuser defined labels. However, it has a few statically\ndefined built-in labels:
\n\n| Feature | \nValue | \nDescription | \n
|---|---|---|
custom-rdma.capable | \n true | \nThe node has an RDMA capable Network adapter | \n
custom-rdma.enabled | \n true | \nThe node has the needed RDMA modules loaded to run RDMA traffic | \n
| \n | \n | \n |
NFD has many extension points for creating vendor and application specific\nlabels. See the customization guide for\ndetailed documentation.
\n\nNFD is able to create extended resources, see the\nNodeFeatureRule CRD and its\nextendedResources field for more\ndetails.
\n\nNote that NFD is not a replacement for the usage of device plugins.
\n\nAn example use-case for extended resources could be based on custom feature\n(created e.g. with feature files that\nexposes the node SGX EPC memory section size. This value will then be turned\ninto an extended resource of the node, allowing PODs to request that resource\nand the Kubernetes scheduler to schedule such PODs to only those nodes which\nhave a sufficient capacity of said resource left.
\n\n\n","dir":"/usage/","name":"features.md","path":"usage/features.md","url":"/usage/features.html"},{"title":"Deployment","layout":"default","sort":2,"content":"Node Feature Discovery can be deployed on any recent version of Kubernetes\n(v1.24+).
\n\nSee Image variants for description of the different NFD\ncontainer images available.
\n\nUsing Kustomize provides straightforward deployment with\nkubectl integration and declarative customization.
Using Helm provides easy management of NFD deployments with nice\nconfiguration management and easy upgrades.
\n\nUsing Operator provides deployment and configuration management via\nCRDs.
\n","dir":"/deployment/","name":"index.md","path":"deployment/index.md","url":"/deployment/"},{"title":"Kustomize","layout":"default","sort":2,"content":"Kustomize can be used to\ndeploy NFD. Customization of the deployment is done by maintaining\ndeclarative overlays on top of the base overlays in NFD.
\n\nTo follow the deployment instructions here,\nkubectl v1.24 or\nlater is required.
\n\nThe kustomize overlays provided in the repo can be used directly:
\n\nkubectl apply -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default?ref=master\"\nThis will required RBAC rules and deploy nfd-master (as a deployment) and\nnfd-worker (as daemonset) in the node-feature-discovery namespace.
\n\n\nNOTE: nfd-topology-updater is not deployed as part of the
\ndefault\noverlay. Refer to the Master Worker Topologyupdater\nand Topologyupdater below.
Alternatively you can clone the repository and customize the deployment by\ncreating your own overlays. See kustomize for more information\nabout managing deployment configurations.
\n\nThe NFD repository hosts a set of overlays for different usages and deployment\nscenarios under\ndeployment/overlays
default:\ndefault deployment of nfd-worker as a daemonset, described abovedefault-job:\nsee Worker one-shot belowmaster-worker-topologyupdater:\nsee Master Worker Topologyupdater belowtopologyupdater:\nsee Topology Updater belowprometheus:\nsee Metrics belowprune:\nclean up the cluster after uninstallation, see\nRemoving feature labelssamples/custom-rules:\nan example for spicing up the default deployment with a separately managed\nconfigmap of custom labeling rules, see\nCustom feature source for more information about\ncustom node labelsFeature discovery can alternatively be configured as a one-shot job.\nThe default-job overlay may be used to achieve this:
NUM_NODES=$(kubectl get no -o jsonpath='{.items[*].metadata.name}' | wc -w)\nkubectl kustomize \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default-job?ref=master\" | \\\n sed s\"/NUM_NODES/$NUM_NODES/\" | \\\n kubectl apply -f -\nThe example above launches as many jobs as there are non-master nodes. Note that\nthis approach does not guarantee running once on every node. For example,\ntainted, non-ready nodes or some other reasons in Job scheduling may cause some\nnode(s) will run extra job instance(s) to satisfy the request.
\n\nNFD-Master, nfd-worker and nfd-topology-updater can be configured to be\ndeployed as separate pods. The master-worker-topologyupdater overlay may be\nused to achieve this:
kubectl apply -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/master-worker-topologyupdater?ref=master\"\n\nTo deploy just nfd-topology-updater (without nfd-master and nfd-worker)\nuse the topologyupdater overlay:
kubectl apply -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/topologyupdater?ref=master\"\n\nNFD-Topology-Updater can be configured along with the default overlay\n(which deploys nfd-worker and nfd-master) where all the software components\nare deployed as separate pods;
\nkubectl apply -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default?ref=master\"\nkubectl apply -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/topologyupdater?ref=master\"\n\nTo allow prometheus operator\nto scrape metrics from node-feature-discovery,\nrun the following command:
\n\nkubectl apply -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default?ref=master\"\nkubectl apply -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/prometheus?ref=master\"\nSimplest way is to invoke kubectl delete on the overlay that was used for\ndeployment. Beware that this will also delete the namespace that NFD is\nrunning in. For example, in case the default overlay from the repo was used:
kubectl delete -k https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default?ref=master\nAlternatively you can delete create objects one-by-one, depending on the type\nof deployment, for example:
\n\nNFD_NS=node-feature-discovery\nkubectl -n $NFD_NS delete ds nfd-worker\nkubectl -n $NFD_NS delete deploy nfd-master\nkubectl -n $NFD_NS delete svc nfd-master\nkubectl -n $NFD_NS delete sa nfd-master\nkubectl delete clusterrole nfd-master\nkubectl delete clusterrolebinding nfd-master\nMinimal steps to deploy latest released version of NFD in your cluster.
\n\nDeploy with kustomize – creates a new namespace, service and required RBAC\nrules and deploys nfd-master and nfd-worker daemons.
\n\nkubectl apply -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default?ref=master\"\nWait until NFD master and NFD worker are running.
\n\n$ kubectl -n node-feature-discovery get ds,deploy\nNAME DESIRED CURRENT READY UP-TO-DATE AVAILABLE NODE SELECTOR AGE\ndaemonset.apps/nfd-worker 2 2 2 2 2 <none> 10s\n\nNAME READY UP-TO-DATE AVAILABLE AGE\ndeployment.apps/nfd-master 1/1 1 1 17s\n\nCheck that NFD feature labels have been created
\n\n$ kubectl get no -o json | jq \".items[].metadata.labels\"\n{\n \"kubernetes.io/arch\": \"amd64\",\n \"kubernetes.io/os\": \"linux\",\n \"feature.node.kubernetes.io/cpu-cpuid.ADX\": \"true\",\n \"feature.node.kubernetes.io/cpu-cpuid.AESNI\": \"true\",\n \"feature.node.kubernetes.io/cpu-cpuid.AVX\": \"true\",\n...\nCreate a pod targeting a distinguishing feature (select a valid feature from\nthe list printed on the previous step)
\n\n$ cat << EOF | kubectl apply -f -\napiVersion: v1\nkind: Pod\nmetadata:\n name: feature-dependent-pod\nspec:\n containers:\n - image: registry.k8s.io/pause\n name: pause\n nodeSelector:\n # Select a valid feature\n feature.node.kubernetes.io/cpu-cpuid.AESNI: 'true'\nEOF\npod/feature-dependent-pod created\nSee that the pod is running on a desired node
\n\n$ kubectl get po feature-dependent-pod -o wide\nNAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES\nfeature-dependent-pod 1/1 Running 0 23s 10.36.0.4 node-2 <none> <none>\nTo deploy nfd-topology-updater use the topologyupdater kustomize\noverlay.
kubectl apply -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/topologyupdater?ref=master\"\nWait until nfd-topology-updater is running.
\n\n$ kubectl -n node-feature-discovery get ds\nNAME DESIRED CURRENT READY UP-TO-DATE AVAILABLE NODE SELECTOR AGE\ndaemonset.apps/nfd-topology-updater 2 2 2 2 2 <none> 5s\n\nCheck that the NodeResourceTopology objects are created
\n\n$ kubectl get noderesourcetopologies.topology.node.k8s.io\nNAME AGE\nkind-control-plane 23s\nkind-worker 23s\nTo quickly view available command line flags execute nfd-worker -help.\nIn a docker container:
docker run gcr.io/k8s-staging-nfd/node-feature-discovery:master nfd-worker -help\nPrint usage and exit.
\n\nPrint version and exit.
\n\nThe -feature-gates flag is used to enable or disable non GA features.\nThe list of available feature gates can be found in the feature gates documentation.
Example:
\n\nnfd-master -feature-gates NodeFeatureGroupAPI=true\nThe -config flag specifies the path of the nfd-worker configuration file to\nuse.
Default: /etc/kubernetes/node-feature-discovery/nfd-worker.conf
\n\nExample:
\n\nnfd-worker -config=/opt/nfd/worker.conf\nThe -options flag may be used to specify and override configuration file\noptions directly from the command line. The required format is the same as in\nthe config file i.e. JSON or YAML. Configuration options specified via this\nflag will override those from the configuration file:
Default: empty
\n\nExample:
\n\nnfd-worker -options='{\"sources\":{\"cpu\":{\"cpuid\":{\"attributeWhitelist\":[\"AVX\",\"AVX2\"]}}}}'\nThe -kubeconfig flag specifies the kubeconfig to use for connecting to the\nKubernetes API server. It is needed for manipulating\nNodeFeature objects. An empty value\n(which is also the default) implies in-cluster kubeconfig.
Default: empty
\n\nExample:
\n\nnfd-worker -kubeconfig ${HOME}/.kube/config\nThe -feature-sources flag specifies a comma-separated list of enabled feature\nsources. A special value all enables all sources. Prefixing a source name\nwith - indicates that the source will be disabled instead - this is only\nmeaningful when used in conjunction with all. This command line flag allows\ncompletely disabling the feature detection so that neither standard feature\nlabels are generated nor the raw feature data is available for custom rule\nprocessing. Consider using the core.featureSources config file option,\ninstead, allowing dynamic configurability.
\n\n\nNOTE: This flag takes precedence over the
\ncore.featureSources\nconfiguration file option.
Default: all
\n\nExample:
\n\nnfd-worker -feature-sources=all,-pci\nThe -label-sources flag specifies a comma-separated list of enabled label\nsources. A special value all enables all sources. Prefixing a source name\nwith - indicates that the source will be disabled instead - this is only\nmeaningful when used in conjunction with all. Consider using the\ncore.labelSources config file option, instead, allowing dynamic\nconfigurability.
\n\n\nNOTE: This flag takes precedence over the
\ncore.labelSources\nconfiguration file option.
Default: all
\n\nExample:
\n\nnfd-worker -label-sources=kernel,system,local\nDEPRECATED: Will be removed in NFD v0.17 and replaced by -port.
The -metrics flag specifies the port on which to expose\nPrometheus metrics. Setting this to 0 disables the\nmetrics server on nfd-worker.
Default: 8081
\n\nExample:
\n\nnfd-worker -metrics=12345\nThe -no-publish flag disables all communication with the nfd-master and the\nKubernetes API server. It is effectively a “dry-run” flag for nfd-worker.\nNFD-Worker runs feature detection normally, but no labeling requests are sent\nto nfd-master and no NodeFeature objects are created or updated in the API\nserver.
\n\n\nNOTE: This flag takes precedence over the\n
\ncore.noPublish\nconfiguration file option.
Default: false
\n\nExample:
\n\nnfd-worker -no-publish\nThe -no-owner-refs flag disables setting the owner references to Pod\nof the NodeFeature object.
\n\n\nNOTE: This flag takes precedence over the\n
\ncore.noOwnerRefs\nconfiguration file option.
Default: false
\n\nExample:
\n\nnfd-worker -no-owner-refs\nThe -oneshot flag causes nfd-worker to exit after one pass of feature\ndetection.
Default: false
\n\nExample:
\n\nnfd-worker -oneshot -no-publish\nThe following logging-related flags are inherited from the\nklog package.
\n\n\n\n\nNOTE: The logger setup can also be specified via the
\ncore.klog\nconfiguration file options. However, the command line flags take precedence\nover any corresponding config file options specified.
If true, adds the file directory to the header of the log messages.
\n\nDefault: false
\n\nLog to standard error as well as files.
\n\nDefault: false
\n\nWhen logging hits line file:N, emit a stack trace.
\n\nDefault: empty
\n\nIf non-empty, write log files in this directory.
\n\nDefault: empty
\n\nIf non-empty, use this log file.
\n\nDefault: empty
\n\nDefines the maximum size a log file can grow to. Unit is megabytes. If the\nvalue is 0, the maximum file size is unlimited.
\n\nDefault: 1800
\n\nLog to standard error instead of files
\n\nDefault: true
\n\nIf true, avoid header prefixes in the log messages.
\n\nDefault: false
\n\nIf true, avoid headers when opening log files.
\n\nDefault: false
\n\nLogs at or above this threshold go to stderr.
\n\nDefault: 2
\n\nNumber for the log level verbosity.
\n\nDefault: 0
\n\nComma-separated list of pattern=N settings for file-filtered logging.
Default: empty
\n","dir":"/reference/","name":"worker-commandline-reference.md","path":"reference/worker-commandline-reference.md","url":"/reference/worker-commandline-reference.html"},{"title":"Using node labels","layout":"default","sort":2,"content":"Nodes with specific features can be targeted using the nodeSelector field. The\nfollowing example shows how to target nodes with Intel TurboBoost enabled.
apiVersion: v1\nkind: Pod\nmetadata:\n labels:\n env: test\n name: golang-test\nspec:\n containers:\n - image: golang\n name: go1\n nodeSelector:\n feature.node.kubernetes.io/cpu-pstate.turbo: 'true'\nFor more details on targeting nodes, see\nnode selection.
\n","dir":"/usage/","name":"using-labels.md","path":"usage/using-labels.md","url":"/usage/using-labels.html"},{"title":"Helm","layout":"default","sort":3,"content":"Node Feature Discovery provides a Helm chart to manage its deployment.
\n\n\n\n\nNOTE: NFD is not ideal for other Helm charts to depend on as that may\nresult in multiple parallel NFD deployments in the same cluster which is not\nfully supported by the NFD Helm chart.
\n
Helm package manager should be installed.
\n\nTo install the latest stable version:
\n\nexport NFD_NS=node-feature-discovery\nhelm repo add nfd https://kubernetes-sigs.github.io/node-feature-discovery/charts\nhelm repo update\nhelm install nfd/node-feature-discovery --namespace $NFD_NS --create-namespace --generate-name\nTo install the latest development version you need to clone the NFD Git\nrepository and install from there.
\n\ngit clone https://github.com/kubernetes-sigs/node-feature-discovery/\ncd node-feature-discovery/deployment/helm\nexport NFD_NS=node-feature-discovery\nhelm install node-feature-discovery ./node-feature-discovery/ --namespace $NFD_NS --create-namespace\nSee the configuration section below for instructions how to\nalter the deployment parameters.
\n\nYou can override values from values.yaml and provide a file with custom values:
export NFD_NS=node-feature-discovery\nhelm install nfd/node-feature-discovery -f <path/to/custom/values.yaml> --namespace $NFD_NS --create-namespace\nTo specify each parameter separately you can provide them to helm install command:
\n\nexport NFD_NS=node-feature-discovery\nhelm install nfd/node-feature-discovery --set nameOverride=NFDinstance --set master.replicaCount=2 --namespace $NFD_NS --create-namespace\nTo upgrade the node-feature-discovery deployment to master via Helm.
Please see\nthe uninstallation guide.\nAnd then follow the standard deployment instructions.
\n\nHelm deployment of NFD was introduced in v0.8.0.
\n\nexport NFD_NS=node-feature-discovery\n# Uninstall the old NFD deployment\nhelm uninstall node-feature-discovery --namespace $NFD_NS\n# Update Helm repository\nhelm repo update\n# Install the new NFD deployment\nhelm upgrade --install node-feature-discovery nfd/node-feature-discovery --namespace $NFD_NS --set master.enable=false\n# Wait for NFD Worker to be ready\nkubectl wait --timeout=-1s --for=condition=ready pod -l app.kubernetes.io/name=node-feature-discovery --namespace $NFD_NS\n# Enable the NFD Master\nhelm upgrade --install node-feature-discovery nfd/node-feature-discovery --namespace $NFD_NS --set master.enable=true\nIn v0.12 the NodeFeature CRD was introduced as experimental.\nThe API was not enabled by default.
export NFD_NS=node-feature-discovery\n# Update Helm repository\nhelm repo update\n# Install and upgrade CRD's\nkubectl apply -f https://raw.githubusercontent.com/kubernetes-sigs/node-feature-discovery/master/deployment/base/nfd-crds/nfd-api-crds.yaml\n# Install the new NFD deployment\nhelm upgrade node-feature-discovery nfd/node-feature-discovery --namespace $NFD_NS --set master.enable=false\n# Wait for NFD Worker to be ready\nkubectl wait --timeout=-1s --for=condition=ready pod -l app.kubernetes.io/name=node-feature-discovery --namespace $NFD_NS\n# Enable the NFD Master\nhelm upgrade node-feature-discovery nfd/node-feature-discovery --namespace $NFD_NS --set master.enable=true\nAs of version v0.14 the Helm chart is the primary deployment method for NFD,\nand the CRD NodeFeature is enabled by default.
export NFD_NS=node-feature-discovery\n# Update Helm repository\nhelm repo update\n# Install and upgrade CRD's\nkubectl apply -f https://raw.githubusercontent.com/kubernetes-sigs/node-feature-discovery/master/deployment/base/nfd-crds/nfd-api-crds.yaml\n# Install the new NFD deployment\nhelm upgrade node-feature-discovery nfd/node-feature-discovery --namespace $NFD_NS\nTo uninstall the node-feature-discovery deployment:
export NFD_NS=node-feature-discovery\nhelm uninstall node-feature-discovery --namespace $NFD_NS\nThe command removes all the Kubernetes components associated with the chart and\ndeletes the release. It also runs a post-delete hook that cleans up the nodes\nof all labels, annotations, taints and extended resources that were created by\nNFD.
\n\nTo tailor the deployment of the Node Feature Discovery to your needs following\nChart parameters are available.
\n\n| Name | \nType | \nDefault | \nDescription | \n
|---|---|---|---|
image.repository | \n string | \ngcr.io/k8s-staging-nfd/node-feature-discovery | \n NFD image repository | \n
image.tag | \n string | \nmaster | \n NFD image tag | \n
image.pullPolicy | \n string | \nAlways | \n Image pull policy | \n
imagePullSecrets | \n array | \n[] | \nImagePullSecrets is an optional list of references to secrets in the same namespace to use for pulling any of the images used by this PodSpec. More info. | \n
nameOverride | \n string | \n\n | Override the name of the chart | \n
fullnameOverride | \n string | \n\n | Override a default fully qualified app name | \n
featureGates.NodeFeatureGroupAPI | \n bool | \nfalse | \nEnable the NodeFeatureGroup CRD API. | \n
featureGates.DisableAutoPrefix | \n bool | \nfalse | \nEnable DisableAutoPrefix feature gate. Disables automatic prefixing of unprefixed labels, annotations and extended resources. | \n
prometheus.enable | \n bool | \nfalse | \nSpecifies whether to expose metrics using prometheus operator | \n
prometheus.labels | \n dict | \n{} | \nSpecifies labels for use with the prometheus operator to control how it is selected | \n
prometheus.scrapeInterval | \n string | \n10s | \nSpecifies the interval by which metrics are scraped | \n
priorityClassName | \n string | \n\n | The name of the PriorityClass to be used for the NFD pods. | \n
Metrics are configured to be exposed using prometheus operator API’s by\ndefault. If you want to expose metrics using the prometheus operator\nAPI’s you need to install the prometheus operator in your cluster.
\n\n| Name | \nType | \nDefault | \nDescription | \n
|---|---|---|---|
master.* | \n dict | \n\n | NFD master deployment configuration | \n
master.enable | \n bool | \ntrue | \nSpecifies whether nfd-master should be deployed | \n
master.hostNetwork | \n bool | \nfalse | \nSpecifies whether to enable or disable running the container in the host’s network namespace | \n
master.metricsPort | \n integer | \n8081 | \nPort on which to expose metrics from components to prometheus operator. DEPRECATED: will be replaced by master.port in NFD v0.18. | \n
master.healthPort | \n integer | \n8082 | \nPort on which to expose the grpc health endpoint, will be also used for the probes. DEPRECATED: will be replaced by master.port in NFD v0.18. | \n
master.instance | \n string | \n\n | Instance name. Used to separate annotation namespaces for multiple parallel deployments | \n
master.resyncPeriod | \n string | \n\n | NFD API controller resync period. | \n
master.extraLabelNs | \n array | \n[] | \nList of allowed extra label namespaces | \n
master.enableTaints | \n bool | \nfalse | \nSpecifies whether to enable or disable node tainting | \n
master.replicaCount | \n integer | \n1 | \nNumber of desired pods. This is a pointer to distinguish between explicit zero and not specified | \n
master.podSecurityContext | \n dict | \n{} | \nPodSecurityContext holds pod-level security attributes and common container settings | \n
master.securityContext | \n dict | \n{} | \nContainer security settings | \n
master.serviceAccount.create | \n bool | \ntrue | \nSpecifies whether a service account should be created | \n
master.serviceAccount.annotations | \n dict | \n{} | \nAnnotations to add to the service account | \n
master.serviceAccount.name | \n string | \n\n | The name of the service account to use. If not set and create is true, a name is generated using the fullname template | \n
master.rbac.create | \n bool | \ntrue | \nSpecifies whether to create RBAC configuration for nfd-master | \n
master.resources.limits | \n dict | \n{memory: 4Gi} | \nNFD master pod resources limits | \n
master.resources.requests | \n dict | \n{cpu: 100m, memory: 128Mi} | \nNFD master pod resources requests. See [0] for more info | \n
master.tolerations | \n dict | \nSchedule to control-plane node | \nNFD master pod tolerations | \n
master.annotations | \n dict | \n{} | \nNFD master pod annotations | \n
master.affinity | \n dict | \n\n | NFD master pod required node affinity | \n
master.deploymentAnnotations | \n dict | \n{} | \nNFD master deployment annotations | \n
master.nfdApiParallelism | \n integer | \n10 | \nSpecifies the maximum number of concurrent node updates. | \n
master.config | \n dict | \n\n | NFD master configuration | \n
master.extraArgs | \n array | \n[] | \nAdditional command line arguments to pass to nfd-master | \n
master.extraEnvs | \n array | \n[] | \nAdditional environment variables to pass to nfd-master | \n
master.revisionHistoryLimit | \n integer | \n\n | Specify how many old ReplicaSets for this Deployment you want to retain. revisionHistoryLimit | \n
master.startupProbe.initialDelaySecond s | \n integer | \n0 (by Kubernetes) | \nSpecifies the number of seconds after the container has started before startup probes are initiated. | \n
master.startupProbe.failureThreshold | \n integer | \n30 | \nSpecifies the number of consecutive failures of startup probes before considering the pod as not ready. | \n
master.startupProbe.periodSeconds | \n integer | \n10 (by Kubernetes) | \nSpecifies how often (in seconds) to perform the startup probe. | \n
master.startupProbe.timeoutSeconds | \n integer | \n1 (by Kubernetes) | \nSpecifies the number of seconds after which the probe times out. | \n
master.livenessProbe.initialDelaySeconds | \n integer | \n0 (by Kubernetes) | \nSpecifies the number of seconds after the container has started before liveness probes are initiated. | \n
master.livenessProbe.failureThreshold | \n integer | \n3 (by Kubernetes) | \nSpecifies the number of consecutive failures of liveness probes before considering the pod as not ready. | \n
master.livenessProbe.periodSeconds | \n integer | \n10 (by Kubernetes) | \nSpecifies how often (in seconds) to perform the liveness probe. | \n
master.livenessProbe.timeoutSeconds | \n integer | \n1 (by Kubernetes) | \nSpecifies the number of seconds after which the probe times out. | \n
master.readinessProbe.initialDelaySeconds | \n integer | \n0 (by Kubernetes) | \nSpecifies the number of seconds after the container has started before readiness probes are initiated. | \n
master.readinessProbe.failureThreshold | \n integer | \n10 | \nSpecifies the number of consecutive failures of readiness probes before considering the pod as not ready. | \n
master.readinessProbe.periodSeconds | \n integer | \n10 (by Kubernetes) | \nSpecifies how often (in seconds) to perform the readiness probe. | \n
master.readinessProbe.timeoutSeconds | \n integer | \n1 (by Kubernetes) | \nSpecifies the number of seconds after which the probe times out. | \n
master.readinessProbe.successThreshold | \n integer | \n1 (by Kubernetes) | \nSpecifies the number of consecutive successes of readiness probes before considering the pod as ready. | \n
\n\n\n\n
[0]Additional info formaster.resources.requests:
\nYou may want to use the same value forrequests.memoryandlimits.memory.\nThe “requests” value affects scheduling to accommodate pods on nodes.\nIf there is a large difference between “requests” and “limits” and nodes\nexperience memory pressure, the kernel may invoke the OOM Killer, even if\nthe memory does not exceed the “limits” threshold.\nThis can cause unexpected pod evictions. Memory cannot be compressed and\nonce allocated to a pod, it can only be reclaimed by killing the pod.\nNatan Yellin 22/09/2022\nthat discusses this issue.
| Name | \nType | \nDefault | \nDescription | \n
|---|---|---|---|
worker.* | \n dict | \n\n | NFD worker daemonset configuration | \n
worker.enable | \n bool | \ntrue | \nSpecifies whether nfd-worker should be deployed | \n
worker.hostNetwork | \n bool | \nfalse | \nSpecifies whether to enable or disable running the container in the host’s network namespace | \n
worker.metricsPort | \n int | \n8081 | \nPort on which to expose metrics from components to prometheus operator. DEPRECATED: will be replaced by worker.port in NFD v0.18. | \n
worker.healthPort | \n int | \n8082 | \nPort on which to expose the grpc health endpoint, will be also used for the probes. DEPRECATED: will be replaced by worker.port in NFD v0.18. | \n
worker.config | \n dict | \n\n | NFD worker configuration | \n
worker.podSecurityContext | \n dict | \n{} | \nPodSecurityContext holds pod-level security attributes and common container settins | \n
worker.securityContext | \n dict | \n{} | \nContainer security settings | \n
worker.serviceAccount.create | \n bool | \ntrue | \nSpecifies whether a service account for nfd-worker should be created | \n
worker.serviceAccount.annotations | \n dict | \n{} | \nAnnotations to add to the service account for nfd-worker | \n
worker.serviceAccount.name | \n string | \n\n | The name of the service account to use for nfd-worker. If not set and create is true, a name is generated using the fullname template (suffixed with -worker) | \n
worker.rbac.create | \n bool | \ntrue | \nSpecifies whether to create RBAC configuration for nfd-worker | \n
worker.mountUsrSrc | \n bool | \nfalse | \nSpecifies whether to allow users to mount the hostpath /user/src. Does not work on systems without /usr/src AND a read-only /usr | \n
worker.resources.limits | \n dict | \n{memory: 512Mi} | \nNFD worker pod resources limits | \n
worker.resources.requests | \n dict | \n{cpu: 5m, memory: 64Mi} | \nNFD worker pod resources requests | \n
worker.nodeSelector | \n dict | \n{} | \nNFD worker pod node selector | \n
worker.tolerations | \n dict | \n{} | \nNFD worker pod node tolerations | \n
worker.priorityClassName | \n string | \n\n | NFD worker pod priority class | \n
worker.annotations | \n dict | \n{} | \nNFD worker pod annotations | \n
worker.daemonsetAnnotations | \n dict | \n{} | \nNFD worker daemonset annotations | \n
worker.extraArgs | \n array | \n[] | \nAdditional command line arguments to pass to nfd-worker | \n
worker.extraEnvs | \n array | \n[] | \nAdditional environment variables to pass to nfd-worker | \n
worker.revisionHistoryLimit | \n integer | \n\n | Specify how many old ControllerRevisions for this DaemonSet you want to retain. revisionHistoryLimit | \n
worker.livenessProbe.initialDelaySeconds | \n integer | \n10 | \nSpecifies the number of seconds after the container has started before liveness probes are initiated. | \n
worker.livenessProbe.failureThreshold | \n integer | \n3 (by Kubernetes) | \nSpecifies the number of consecutive failures of liveness probes before considering the pod as not ready. | \n
worker.livenessProbe.periodSeconds | \n integer | \n10 (by Kubernetes) | \nSpecifies how often (in seconds) to perform the liveness probe. | \n
worker.livenessProbe.timeoutSeconds | \n integer | \n1 (by Kubernetes) | \nSpecifies the number of seconds after which the probe times out. | \n
worker.readinessProbe.initialDelaySeconds | \n integer | \n5 | \nSpecifies the number of seconds after the container has started before readiness probes are initiated. | \n
worker.readinessProbe.failureThreshold | \n integer | \n10 | \nSpecifies the number of consecutive failures of readiness probes before considering the pod as not ready. | \n
worker.readinessProbe.periodSeconds | \n integer | \n10 (by Kubernetes) | \nSpecifies how often (in seconds) to perform the readiness probe. | \n
worker.readinessProbe.timeoutSeconds | \n integer | \n1 (by Kubernetes) | \nSpecifies the number of seconds after which the probe times out. | \n
worker.readinessProbe.successThreshold | \n integer | \n1 (by Kubernetes) | \nSpecifies the number of consecutive successes of readiness probes before considering the pod as ready. | \n
| Name | \nType | \nDefault | \nDescription | \n
|---|---|---|---|
topologyUpdater.* | \n dict | \n\n | NFD Topology Updater configuration | \n
topologyUpdater.enable | \n bool | \nfalse | \nSpecifies whether the NFD Topology Updater should be created | \n
topologyUpdater.hostNetwork | \n bool | \nfalse | \nSpecifies whether to enable or disable running the container in the host’s network namespace | \n
topologyUpdater.createCRDs | \n bool | \nfalse | \nSpecifies whether the NFD Topology Updater CRDs should be created | \n
topologyUpdater.serviceAccount.create | \n bool | \ntrue | \nSpecifies whether the service account for topology updater should be created | \n
topologyUpdater.serviceAccount.annotations | \n dict | \n{} | \nAnnotations to add to the service account for topology updater | \n
topologyUpdater.serviceAccount.name | \n string | \n\n | The name of the service account for topology updater to use. If not set and create is true, a name is generated using the fullname template and -topology-updater suffix | \n
topologyUpdater.rbac.create | \n bool | \ntrue | \nSpecifies whether to create RBAC configuration for topology updater | \n
topologyUpdater.metricsPort | \n integer | \n8081 | \nPort on which to expose prometheus metrics. DEPRECATED: will be replaced by topologyUpdater.port in NFD v0.18. | \n
topologyUpdater.healthPort | \n integer | \n8082 | \nPort on which to expose the grpc health endpoint, will be also used for the probes. DEPRECATED: will be replaced by topologyUpdater.port in NFD v0.18. | \n
topologyUpdater.kubeletConfigPath | \n string | \n”” | \nSpecifies the kubelet config host path | \n
topologyUpdater.kubeletPodResourcesSockPath | \n string | \n”” | \nSpecifies the kubelet sock path to read pod resources | \n
topologyUpdater.updateInterval | \n string | \n60s | \nTime to sleep between CR updates. Non-positive value implies no CR update. | \n
topologyUpdater.watchNamespace | \n string | \n* | \n Namespace to watch pods, * for all namespaces | \n
topologyUpdater.podSecurityContext | \n dict | \n{} | \nPodSecurityContext holds pod-level security attributes and common container sett | \n
topologyUpdater.securityContext | \n dict | \n{} | \nContainer security settings | \n
topologyUpdater.resources.limits | \n dict | \n{memory: 60Mi} | \nNFD Topology Updater pod resources limits | \n
topologyUpdater.resources.requests | \n dict | \n{cpu: 50m, memory: 40Mi} | \nNFD Topology Updater pod resources requests | \n
topologyUpdater.nodeSelector | \n dict | \n{} | \nTopology updater pod node selector | \n
topologyUpdater.tolerations | \n dict | \n{} | \nTopology updater pod node tolerations | \n
topologyUpdater.annotations | \n dict | \n{} | \nTopology updater pod annotations | \n
topologyUpdater.daemonsetAnnotations | \n dict | \n{} | \nTopology updater daemonset annotations | \n
topologyUpdater.affinity | \n dict | \n{} | \nTopology updater pod affinity | \n
topologyUpdater.config | \n dict | \n\n | configuration | \n
topologyUpdater.podSetFingerprint | \n bool | \ntrue | \nEnables compute and report of pod fingerprint in NRT objects. | \n
topologyUpdater.kubeletStateDir | \n string | \n/var/lib/kubelet | \nSpecifies kubelet state directory path for watching state and checkpoint files. Empty value disables kubelet state tracking. | \n
topologyUpdater.extraArgs | \n array | \n[] | \nAdditional command line arguments to pass to nfd-topology-updater | \n
topologyUpdater.extraEnvs | \n array | \n[] | \nAdditional environment variables to pass to nfd-topology-updater | \n
topologyUpdater.revisionHistoryLimit | \n integer | \n\n | Specify how many old ControllerRevisions for this DaemonSet you want to retain. revisionHistoryLimit | \n
topologyUpdater.livenessProbe.initialDelaySeconds | \n integer | \n10 | \nSpecifies the number of seconds after the container has started before liveness probes are initiated. | \n
topologyUpdater.livenessProbe.failureThreshold | \n integer | \n3 (by Kubernetes) | \nSpecifies the number of consecutive failures of liveness probes before considering the pod as not ready. | \n
topologyUpdater.livenessProbe.periodSeconds | \n integer | \n10 (by Kubernetes) | \nSpecifies how often (in seconds) to perform the liveness probe. | \n
topologyUpdater.livenessProbe.timeoutSeconds | \n integer | \n1 (by Kubernetes) | \nSpecifies the number of seconds after which the probe times out. | \n
topologyUpdater.readinessProbe.initialDelaySeconds | \n integer | \n5 | \nSpecifies the number of seconds after the container has started before readiness probes are initiated. | \n
topologyUpdater.readinessProbe.failureThreshold | \n integer | \n10 | \nSpecifies the number of consecutive failures of readiness probes before considering the pod as not ready. | \n
topologyUpdater.readinessProbe.periodSeconds | \n integer | \n10 (by Kubernetes) | \nSpecifies how often (in seconds) to perform the readiness probe. | \n
topologyUpdater.readinessProbe.timeoutSeconds | \n integer | \n1 (by Kubernetes) | \nSpecifies the number of seconds after which the probe times out. | \n
topologyUpdater.readinessProbe.successThreshold | \n integer | \n1 (by Kubernetes) | \nSpecifies the number of consecutive successes of readiness probes before considering the pod as ready. | \n
| Name | \nType | \nDefault | \nDescription | \n
|---|---|---|---|
gc.* | \n dict | \n\n | NFD Garbage Collector configuration | \n
gc.enable | \n bool | \ntrue | \nSpecifies whether the NFD Garbage Collector should be created | \n
gc.hostNetwork | \n bool | \nfalse | \nSpecifies whether to enable or disable running the container in the host’s network namespace | \n
gc.serviceAccount.create | \n bool | \ntrue | \nSpecifies whether the service account for garbage collector should be created | \n
gc.serviceAccount.annotations | \n dict | \n{} | \nAnnotations to add to the service account for garbage collector | \n
gc.serviceAccount.name | \n string | \n\n | The name of the service account for garbage collector to use. If not set and create is true, a name is generated using the fullname template and -gc suffix | \n
gc.rbac.create | \n bool | \ntrue | \nSpecifies whether to create RBAC configuration for garbage collector | \n
gc.interval | \n string | \n1h | \nTime between periodic garbage collector runs | \n
gc.podSecurityContext | \n dict | \n{} | \nPodSecurityContext holds pod-level security attributes and common container settings | \n
gc.resources.limits | \n dict | \n{memory: 1Gi} | \nNFD Garbage Collector pod resources limits | \n
gc.resources.requests | \n dict | \n{cpu: 10m, memory: 128Mi} | \nNFD Garbage Collector pod resources requests | \n
gc.metricsPort | \n integer | \n8081 | \nPort on which to serve Prometheus metrics. DEPRECATED: will be replaced by gc.port in NFD v0.18. | \n
gc.nodeSelector | \n dict | \n{} | \nGarbage collector pod node selector | \n
gc.tolerations | \n dict | \n{} | \nGarbage collector pod node tolerations | \n
gc.annotations | \n dict | \n{} | \nGarbage collector pod annotations | \n
gc.deploymentAnnotations | \n dict | \n{} | \nGarbage collector deployment annotations | \n
gc.affinity | \n dict | \n{} | \nGarbage collector pod affinity | \n
gc.extraArgs | \n array | \n[] | \nAdditional command line arguments to pass to nfd-gc | \n
gc.extraEnvs | \n array | \n[] | \nAdditional environment variables to pass to nfd-gc | \n
gc.revisionHistoryLimit | \n integer | \n\n | Specify how many old ReplicaSets for this Deployment you want to retain. revisionHistoryLimit | \n
See the\nsample configuration file\nfor a full example configuration.
\n\nnoPublish option disables updates to the Node objects in the Kubernetes\nAPI server, making a “dry-run” flag for nfd-master. No Labels, Annotations, Taints\nor ExtendedResources of nodes are updated.
Default: false
Example:
\n\nnoPublish: true\nextraLabelNs specifies a list of allowed feature\nlabel namespaces. This option can be used to allow\nother vendor or application specific namespaces for custom labels from the\nlocal and custom feature sources, even though these labels were denied using\nthe denyLabelNs parameter.
Default: empty
\n\nExample:
\n\nextraLabelNs: [\"added.ns.io\",\"added.kubernets.io\"]\ndenyLabelNs specifies a list of excluded\nlabel namespaces. By default, nfd-master allows creating labels in all\nnamespaces, excluding kubernetes.io namespace and its sub-namespaces\n(i.e. *.kubernetes.io). However, you should note that\nkubernetes.io and its sub-namespaces are always denied.\nThis option can be used to exclude some vendors or application specific\nnamespaces.
Default: empty
\n\nExample:
\n\ndenyLabelNs: [\"denied.ns.io\",\"denied.kubernetes.io\"]\nDEPRECATED: Will be removed in NFD v0.17. Use the\nDisableAutoPrefix feature gate instead.
\n\nThe autoDefaultNs option controls the automatic prefixing of names. When set\nto true (the default in NFD version master) nfd-master\nautomatically adds the default feature.node.kubernetes.io/ prefix to\nunprefixed labels, annotations and extended resources - this is also the\ndefault behavior in NFD v0.15 and earlier. When the option is set to false,\nno prefix will be prepended to unprefixed names, effectively causing them to be\nfiltered out (as NFD does not allow unprefixed names of labels, annotations or\nextended resources). The default will be changed to false in a future\nrelease.
For example, with the autoDefaultNs set to true, a NodeFeatureRule with
labels:\n foo: bar\nWill turn into feature.node.kubernetes.io/foo=bar node label. With\nautoDefaultNs set to false, no prefix is added and the label will be\nfiltered out.
Note that taint keys are not affected by this option.
\n\nDefault: true
Example:
\n\nautoDefaultNs: false\nenableTaints enables/disables node tainting feature of NFD.
Default: false
\n\nExample:
\n\nenableTaints: true\nlabelWhiteList specifies a regular expression for filtering feature\nlabels based on their name. Each label must match against the given regular\nexpression or it will not be published.
\n\n\n** NOTE:** The regular expression is only matches against the “basename” part\nof the label, i.e. to the part of the name after ‘/’. The label namespace is\nomitted.
\n
Default: empty
\n\nExample:
\n\nlabelWhiteList: \"foo\"\nThe resyncPeriod option specifies the NFD API controller resync period.\nThe resync means nfd-master replaying all NodeFeature and NodeFeatureRule objects,\nthus effectively re-syncing all nodes in the cluster (i.e. ensuring labels, annotations,\nextended resources and taints are in place).
Default: 1 hour.
\n\nExample:
\n\nresyncPeriod: 2h\nThe leaderElection section exposes configuration to tweak leader election.
leaderElection.leaseDuration is the duration that non-leader candidates will\nwait to force acquire leadership. This is measured against time of\nlast observed ack.
A client needs to wait a full LeaseDuration without observing a change to\nthe record before it can attempt to take over. When all clients are\nshutdown and a new set of clients are started with different names against\nthe same leader record, they must wait the full LeaseDuration before\nattempting to acquire the lease. Thus LeaseDuration should be as short as\npossible (within your tolerance for clock skew rate) to avoid a possible\nlong waits in the scenario.
\n\nDefault: 15 seconds.
\n\nExample:
\n\nleaderElection:\n leaseDurtation: 15s\nleaderElection.renewDeadline is the duration that the acting master will retry\nrefreshing leadership before giving up.
This value has to be lower than leaseDuration and greater than retryPeriod*1.2.
\n\nDefault: 10 seconds.
\n\nExample:
\n\nleaderElection:\n renewDeadline: 10s\nleaderElection.retryPeriod is the duration the LeaderElector clients should wait\nbetween tries of actions.
It has to be greater than 0.
\n\nDefault: 2 seconds.
\n\nExample:
\n\nleaderElection:\n retryPeriod: 2s\nThe nfdApiParallelism option can be used to specify the maximum\nnumber of concurrent node updates.
Default: 10
\n\nExample:
\n\nnfdApiParallelism: 1\nThe following options specify the logger configuration. Most of which can be\ndynamically adjusted at run-time.
\n\n\n\n\nNOTE: The logger options can also be specified via command line flags\nwhich take precedence over any corresponding config file options.
\n
If true, adds the file directory to the header of the log messages.
\n\nDefault: false
Run-time configurable: yes
\n\nLog to standard error as well as files.
\n\nDefault: false
Run-time configurable: yes
\n\nWhen logging hits line file:N, emit a stack trace.
\n\nDefault: empty
\n\nRun-time configurable: yes
\n\nIf non-empty, write log files in this directory.
\n\nDefault: empty
\n\nRun-time configurable: no
\n\nIf non-empty, use this log file.
\n\nDefault: empty
\n\nRun-time configurable: no
\n\nDefines the maximum size a log file can grow to. Unit is megabytes. If the\nvalue is 0, the maximum file size is unlimited.
\n\nDefault: 1800
Run-time configurable: no
\n\nLog to standard error instead of files
\n\nDefault: true
Run-time configurable: yes
\n\nIf true, avoid header prefixes in the log messages.
\n\nDefault: false
Run-time configurable: yes
\n\nIf true, avoid headers when opening log files.
\n\nDefault: false
Run-time configurable: no
\n\nLogs at or above this threshold go to stderr (default 2)
\n\nRun-time configurable: yes
\n\nNumber for the log level verbosity.
\n\nDefault: 0
Run-time configurable: yes
\n\nComma-separated list of pattern=N settings for file-filtered logging.
Default: empty
\n\nRun-time configurable: yes
\n\nThe following options specify the restrictions that can be applied by the\nnfd-master on the deployed Custom Resources in the cluster.
\n\nThe nodeFeatureNamespaceSelector option specifies the NodeFeatures namespaces\nto watch, which can be selected by using metav1.LabelSelector as a type for\nthis option. An empty value selects all namespaces to be watched.
Default: empty
\n\nExample:
\n\nrestrictions:\n nodeFeatureNamespaceSelector:\n matchLabels:\n kubernetes.io/metadata.name: \"node-feature-discovery\"\n matchExpressions:\n - key: \"kubernetes.io/metadata.name\"\n operator: \"In\"\n values:\n - \"node-feature-discovery\"\nThe disableLabels option controls whether to allow creation of node labels\nfrom NodeFeature and NodeFeatureRule CRs or not.
Default: false
\n\nExample:
\n\nrestrictions:\n disableLabels: true\nThe disableExtendedResources option controls whether to allow creation of\nnode extended resources from NodeFeatureRule CR or not.
Default: false
\n\nExample:
\n\nrestrictions:\n disableExtendedResources: true\nhe disableAnnotations option controls whether to allow creation of node annotations\nfrom NodeFeatureRule CR or not.
Default: false
\n\nExample:
\n\nrestrictions:\n disableAnnotations: true\nThe allowOverwrite option controls whether NFD is allowed to overwrite and\ntake over management of existing node labels, annotations, and extended resources.\nLabels, annotations and extended resources created by NFD itself are not affected\n(overwrite cannot be disabled). NFD tracks the labels, annotations and extended\nresources that it manages with specific\nnode annotations.
Default: true
\n\nExample:
\n\nrestrictions:\n allowOverwrite: false\nThe denyNodeFeatureLabels option specifies whether to deny labels from 3rd party\nNodeFeature objects or not. NodeFeature objects created by nfd-worker are not affected.
Default: false
\n\nExample:
\n\nrestrictions:\n denyNodeFeatureLabels: true\nUsage instructions.
\n","dir":"/usage/","name":"index.md","path":"usage/index.md","url":"/usage/"},{"title":"NFD-Master","layout":"default","sort":3,"content":"NFD-Master is responsible for connecting to the Kubernetes API server and\nupdating node objects. More specifically, it modifies node labels, taints and\nextended resources based on requests from nfd-workers and 3rd party extensions.
\n\nThe NodeFeature Controller uses NodeFeature objects as\nthe input for the NodeFeatureRule\nprocessing pipeline. In addition, any labels listed in the NodeFeature object\nare created on the node (note the allowed\nlabel namespaces are controlled).
\n\nNFD-Master acts as the controller for\nNodeFeatureRule objects.\nIt applies the rules specified in NodeFeatureRule objects on raw feature data\nand creates node labels accordingly. The feature data used as the input is\nreceived from nfd-worker instances through\nNodeFeature objects.
\n\nNFD-Master supports configuration through a configuration file. The\ndefault location is /etc/kubernetes/node-feature-discovery/nfd-master.conf,\nbut, this can be changed by specifying the-config command line flag.
Master configuration file is read inside the container, and thus, Volumes and\nVolumeMounts are needed to make your configuration available for NFD. The\npreferred method is to use a ConfigMap which provides easy deployment and\nre-configurability.
\n\nThe provided deployment methods (Helm and Kustomize) create an empty configmap\nand mount it inside the nfd-master containers.
\n\nIn Helm deployments,\nMaster pod parameter\nmaster.config can be used to edit the respective configuration.
In Kustomize deployments, modify the nfd-master-conf ConfigMap with a custom\noverlay.
\n\n\nNOTE: dynamic run-time reconfiguration was dropped in NFD v0.17.\nRe-configuration is handled by pod restarts.
\n
See\nnfd-master configuration file reference\nfor more details.\nThe (empty-by-default)\nexample config\ncontains all available configuration options and can be used as a reference\nfor creating a configuration.
\n\nNFD-Master runs as a deployment, by default\nit prefers running on the cluster’s master nodes but will run on worker\nnodes if no master nodes are found.
\n\nFor High Availability, you should increase the replica count of\nthe deployment object. You should also look into adding\ninter-pod\naffinity to prevent masters from running on the same node.\nHowever note that inter-pod affinity is costly and is not recommended\nin bigger clusters.
\n\n\n\n\nNote: When NFD-Master is intended to run with more than one replica,\nit is advised to use
\n-enable-leader-electionflag. This flag turns on\nleader election for NFD-Master and let only one replica to act on changes\nin NodeFeature and NodeFeatureRule objects.
If you have RBAC authorization enabled (as is the default e.g. with clusters\ninitialized with kubeadm) you need to configure the appropriate ClusterRoles,\nClusterRoleBindings and a ServiceAccount for NFD to create node\nlabels. The provided template will configure these for you.
\n","dir":"/usage/","name":"nfd-master.md","path":"usage/nfd-master.md","url":"/usage/nfd-master.html"},{"title":"NFD Operator","layout":"default","sort":4,"content":"The Node Feature Discovery Operator automates installation,\nconfiguration and updates of NFD using a specific NodeFeatureDiscovery custom\nresource. This also provides good support for managing NFD as a dependency of\nother operators.
\n\nDeployment using the\nNode Feature Discovery Operator\nis recommended to be done via\noperatorhub.io.
\n\nInstall the operator:
\n\nkubectl create -f https://operatorhub.io/install/nfd-operator.yaml\nCreate NodeFeatureDiscovery object (in nfd namespace here):
cat << EOF | kubectl apply -f -\napiVersion: v1\nkind: Namespace\nmetadata:\n name: nfd\n---\napiVersion: nfd.kubernetes.io/v1\nkind: NodeFeatureDiscovery\nmetadata:\n name: my-nfd-deployment\n namespace: nfd\nspec:\n operand:\n image: gcr.io/k8s-staging-nfd/node-feature-discovery:master\n imagePullPolicy: IfNotPresent\nEOF\nIf you followed the deployment instructions above you can uninstall NFD with:
\n\nkubectl -n nfd delete NodeFeatureDiscovery my-nfd-deployment\nOptionally, you can also remove the namespace:
\n\nkubectl delete ns nfd\nSee the node-feature-discovery-operator and OLM project\ndocumentation for instructions for uninstalling the operator and operator\nlifecycle manager, respectively.
\n\n\n","dir":"/deployment/","name":"operator.md","path":"deployment/operator.md","url":"/deployment/operator.html"},{"title":"Reference","layout":"default","sort":4,"content":"Command line and configuration reference.
\n","dir":"/reference/","name":"index.md","path":"reference/index.md","url":"/reference/"},{"title":"Worker config reference","layout":"default","sort":4,"content":"See the\nsample configuration file\nfor a full example configuration.
\n\nThe core section contains common configuration settings that are not specific\nto any particular feature source.
core.sleepInterval specifies the interval between consecutive passes of\nfeature (re-)detection, and thus also the interval between node re-labeling. A\nnon-positive value implies infinite sleep interval, i.e. no re-detection or\nre-labeling is done.
Default: 60s
Example:
\n\ncore:\n sleepInterval: 60s\ncore.featureSources specifies the list of enabled feature sources. A special\nvalue all enables all sources. Prefixing a source name with - indicates\nthat the source will be disabled instead - this is only meaningful when used in\nconjunction with all. This option allows completely disabling the feature\ndetection so that neither standard feature labels are generated nor the raw\nfeature data is available for custom rule processing.
Default: [all]
Example:
\n\ncore:\n # Enable all but cpu and local sources\n featureSources:\n - \"all\"\n - \"-cpu\"\n - \"-local\"\ncore:\n # Enable only cpu and local sources\n featureSources:\n - \"cpu\"\n - \"local\"\ncore.labelSources specifies the list of enabled label sources. A special\nvalue all enables all sources. Prefixing a source name with - indicates\nthat the source will be disabled instead - this is only meaningful when used in\nconjunction with all. This configuration option affects the generation of\nnode labels but not the actual discovery of the underlying feature data that is\nused e.g. in custom/NodeFeatureRule rules.
\n\n\nNOTE: Overridden by the
\n-label-sourcescommand line flag and the\ncore.sourcesconfigurations option (if either of them is specified).
Default: [all]
Example:
\n\ncore:\n # Enable all but cpu and system sources\n labelSources:\n - \"all\"\n - \"-cpu\"\n - \"-system\"\ncore:\n # Enable only cpu and system sources\n labelSources:\n - \"cpu\"\n - \"system\"\nDEPRECATED: use core.labelSources instead.
\n\n\nNOTE:
\ncore.sourcestakes precedence over thecore.labelSources\nconfiguration file option.
core.labelWhiteList specifies a regular expression for filtering feature\nlabels based on the label name. Non-matching labels are not published.
\n\n\nNOTE: The regular expression is only matches against the “basename” part\nof the label, i.e. to the part of the name after ‘/’. The label prefix (or\nnamespace) is omitted.
\n
Default: null
Example:
\n\ncore:\n labelWhiteList: '^cpu-cpuid'\nSetting core.noPublish to true disables all communication with the\nnfd-master and the Kubernetes API server. It is effectively a “dry-run” option.\nNFD-Worker runs feature detection normally, but no labeling requests are sent\nto nfd-master and no NodeFeature\nobjects are created or updated in the API server.
\n\n\nNOTE: Overridden by the\n
\n-no-publish\ncommand line flag (if specified).
Default: false
Example:
\n\ncore:\n noPublish: true\nSetting core.noOwnerRefs to true disables setting the owner references\nof the NodeFeature object created by the nfd-worker.
\n\n\nNOTE: Overridden by the\n
\n-no-owner-refs\ncommand line flag (if specified).
Default: false
Example:
\n\ncore:\n noOwnerRefs: true\nThe following options specify the logger configuration.
\n\n\n\n\nNOTE: The logger options can also be specified via command line flags\nwhich take precedence over any corresponding config file options.
\n
If true, adds the file directory to the header of the log messages.
\n\nDefault: false
Log to standard error as well as files.
\n\nDefault: false
When logging hits line file:N, emit a stack trace.
\n\nDefault: empty
\n\nIf non-empty, write log files in this directory.
\n\nDefault: empty
\n\nIf non-empty, use this log file.
\n\nDefault: empty
\n\nDefines the maximum size a log file can grow to. Unit is megabytes. If the\nvalue is 0, the maximum file size is unlimited.
\n\nDefault: 1800
Log to standard error instead of files
\n\nDefault: true
If true, avoid header prefixes in the log messages.
\n\nDefault: false
If true, avoid headers when opening log files.
\n\nDefault: false
Logs at or above this threshold go to stderr (default 2)
\n\nNumber for the log level verbosity.
\n\nDefault: 0
Comma-separated list of pattern=N settings for file-filtered logging.
Default: empty
\n\nThe sources section contains feature source specific configuration parameters.
Prevent publishing cpuid features listed in this option.
\n\n\n\n\nNOTE: overridden by
\nsources.cpu.cpuid.attributeWhitelist(if specified)
Default: [AVX10, BMI1, BMI2, CLMUL, CMOV, CX16, ERMS, F16C, HTT, LZCNT, MMX, MMXEXT,\nNX, POPCNT, RDRAND, RDSEED, RDTSCP, SGX, SGXLC, SSE, SSE2, SSE3, SSE4.1,\nSSE4.2, SSSE3, TDX_GUEST]
Example:
\n\nsources:\n cpu:\n cpuid:\n attributeBlacklist: [MMX, MMXEXT]\nOnly publish the cpuid features listed in this option.
\n\n\n\n\nNOTE: takes precedence over
\nsources.cpu.cpuid.attributeBlacklist
Default: empty
\n\nExample:
\n\nsources:\n cpu:\n cpuid:\n attributeWhitelist: [AVX512BW, AVX512CD, AVX512DQ, AVX512F, AVX512VL]\nPath of the kernel config file. If empty, NFD runs a search in the well-known\nstandard locations.
\n\nDefault: empty
\n\nExample:
\n\nsources:\n kernel:\n kconfigFile: \"/path/to/kconfig\"\nKernel configuration options to publish as feature labels.
\n\nDefault: [NO_HZ, NO_HZ_IDLE, NO_HZ_FULL, PREEMPT]
Example:
\n\nsources:\n kernel:\n configOpts: [NO_HZ, X86, DMI]\nList of PCI device class IDs for which to\npublish a label. Can be specified as a main class only (e.g. 03) or full\nclass-subclass combination (e.g. 0300) - the former implies that all\nsubclasses are accepted. The format of the labels can be further configured\nwith deviceLabelFields.
Default: [\"03\", \"0b40\", \"12\"]
Example:
\n\nsources:\n pci:\n deviceClassWhitelist: [\"0200\", \"03\"]\nThe set of PCI ID fields to use when constructing the name of the feature\nlabel. Valid fields are class, vendor, device, subsystem_vendor and\nsubsystem_device.
Default: [class, vendor]
Example:
\n\nsources:\n pci:\n deviceLabelFields: [class, vendor, device]\nWith the example config above NFD would publish labels like:\nfeature.node.kubernetes.io/pci-<class-id>_<vendor-id>_<device-id>.present=true
List of USB device class IDs for\nwhich to publish a feature label. The format of the labels can be further\nconfigured with deviceLabelFields.
\n\nDefault: [\"0e\", \"ef\", \"fe\", \"ff\"]
Example:
\n\nsources:\n usb:\n deviceClassWhitelist: [\"ef\", \"ff\"]\nThe set of USB ID fields from which to compose the name of the feature label.\nValid fields are class, vendor, device and serial.
Default: [class, vendor, device]
Example:
\n\nsources:\n pci:\n deviceLabelFields: [class, vendor]\nWith the example config above NFD would publish labels like:\nfeature.node.kubernetes.io/usb-<class-id>_<vendor-id>.present=true
List of rules to process in the custom feature source to create user-specific\nlabels. Refer to the documentation of the\ncustom feature source for\ndetails of the available rules and their configuration.
\n\nDefault: empty
\n\nExample:
\n\nsources:\n custom:\n - name: \"my custom rule\"\n labels:\n my-custom-feature: \"true\"\n matchFeatures:\n - feature: kernel.loadedmodule\n matchExpressions:\n e1000e: {op: Exists}\n - feature: pci.device\n matchExpressions:\n class: {op: In, value: [\"0200\"]}\n vendor: {op: In, value: [\"8086\"]}\nNFD-Worker is preferably run as a Kubernetes DaemonSet. This assures\nre-labeling on regular intervals capturing changes in the system configuration\nand makes sure that new nodes are labeled as they are added to the cluster.\nWorker connects to the nfd-master service to advertise hardware features.
\n\nWhen run as a daemonset, nodes are re-labeled at an default interval of 60s.\nThis can be changed by using the\ncore.sleepInterval\nconfig option.
NFD-Worker supports configuration through a configuration file. The\ndefault location is /etc/kubernetes/node-feature-discovery/nfd-worker.conf,\nbut, this can be changed by specifying the-config command line flag.\nConfiguration file is re-read whenever it is modified which makes run-time\nre-configuration of nfd-worker straightforward.
Worker configuration file is read inside the container, and thus, Volumes and\nVolumeMounts are needed to make your configuration available for NFD. The\npreferred method is to use a ConfigMap which provides easy deployment and\nre-configurability.
\n\nThe provided deployment methods (Helm and Kustomize) create an empty configmap\nand mount it inside the nfd-master containers.
\n\nIn Helm deployments,\nWorker pod parameter\nworker.config can be used to edit the respective configuration.
In Kustomize deployments, modify the nfd-worker-conf ConfigMap with a custom\noverlay.
\n\n\nNOTE: dynamic run-time reconfiguration was dropped in NFD v0.17.\nRe-configuration is handled by pod restarts.
\n
See\nnfd-worker configuration file reference\nfor more details.\nThe (empty-by-default)\nexample config\ncontains all available configuration options and can be used as a reference\nfor creating a configuration.
\n\nConfiguration options can also be specified via the -options command line\nflag, in which case no mounts need to be used. The same format as in the config\nfile must be used, i.e. JSON (or YAML). For example:
-options='{\"sources\": { \"pci\": { \"deviceClassWhitelist\": [\"12\"] } } }'\nConfiguration options specified from the command line will override those read\nfrom the config file.
\n","dir":"/usage/","name":"nfd-worker.md","path":"usage/nfd-worker.md","url":"/usage/nfd-worker.html"},{"title":"Uninstallation","layout":"default","sort":5,"content":"Follow the uninstallation instructions of the deployment method used\n(kustomize,\nhelm or\noperator).
\n\n\n\n\nNOTE: This is unnecessary when using the Helm chart for deployment as it\nwill clean up the nodes when NFD is uninstalled.
\n
NFD-Master has a special -prune command line flag for removing all\nnfd-related node labels, annotations, extended resources and taints from the\ncluster.
kubectl apply -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/prune?ref=master\"\nkubectl -n node-feature-discovery wait job.batch/nfd-master --for=condition=complete && \\\n kubectl delete -k \"https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/prune?ref=master\"\n\n\n","dir":"/deployment/","name":"uninstallation.md","path":"deployment/uninstallation.md","url":"/deployment/uninstallation.html"},{"title":"Developer guide","layout":"default","sort":5,"content":"NOTE: You must run prune before removing the RBAC rules (serviceaccount,\nclusterrole and clusterrolebinding).
\n
git clone https://github.com/kubernetes-sigs/node-feature-discovery\ncd node-feature-discovery\nSee customizing the build below for altering the\ncontainer image registry, for example.
\n\nmake\nOptional, this example with Docker.
\n\ndocker push <IMAGE_TAG>\nThe default set of architectures enabled for mulit-arch builds are linux/amd64\nand linux/arm64. If more architectures are needed one can override the\nIMAGE_ALL_PLATFORMS variable with a comma separated list of OS/ARCH tuples.
make image-all\nCurrently docker does not support loading of manifest-lists meaning the images\nare not shown when executing docker images, see:\nbuildx issue #59.
make push-all\nThe resulting container image can be used in the same way on each arch by pulling\ne.g. node-feature-discovery:master without specifying the\narchitecture. The manifest-list will take care of providing the right\narchitecture image.
To use your published image from the step above instead of the\nregistry.k8s.io/nfd/node-feature-discovery image, edit image\nattribute in the spec template(s) to the new location\n(<registry-name>/<image-name>[:<version>]).
The yamls makefile generates a kustomization.yaml matching your locally\nbuilt image and using the deploy/overlays/default deployment. See\nbuild customization below for configurability, e.g.\nchanging the deployment namespace.
K8S_NAMESPACE=my-ns make yamls\nkubectl apply -k .\nYou can use alternative deployment methods by modifying the auto-generated\nkustomization file.
\n\nYou can also build the binaries locally
\n\nmake build\nThis will compile binaries under bin/
There are several Makefile variables that control the build process and the\nname of the resulting container image. The following are targeted targeted for\nbuild customization and they can be specified via environment variables or\nmakefile overrides.
\n\n| Variable | \nDescription | \nDefault value | \n
|---|---|---|
| HOSTMOUNT_PREFIX | \nPrefix of system directories for feature discovery (local builds) | \n/ (local builds) /host- (container builds) | \n
| IMAGE_BUILD_CMD | \nCommand to build the image | \ndocker build | \n
| IMAGE_BUILD_EXTRA_OPTS | \nExtra options to pass to build command | \nempty | \n
| IMAGE_BUILDX_CMD | \nCommand to build and push multi-arch images with buildx | \nDOCKER_CLI_EXPERIMENTAL=enabled docker buildx build –platform=${IMAGE_ALL_PLATFORMS} –progress=auto –pull | \n
| IMAGE_ALL_PLATFORMS | \nComma separated list of OS/ARCH tuples for mulit-arch builds | \nlinux/amd64,linux/arm64 | \n
| IMAGE_PUSH_CMD | \nCommand to push the image to remote registry | \ndocker push | \n
| IMAGE_REGISTRY | \nContainer image registry to use | \nregistry.k8s.io/nfd | \n
| IMAGE_TAG_NAME | \nContainer image tag name | \n<nfd version> | \n
| IMAGE_EXTRA_TAG_NAMES | \nAdditional container image tag(s) to create when building image | \nempty | \n
| K8S_NAMESPACE | \nnfd-master and nfd-worker namespace | \nnode-feature-discovery | \n
For example, to use a custom registry:
\n\nmake IMAGE_REGISTRY=<my custom registry uri>\nOr to specify a build tool different from Docker, It can be done in 2 ways:
\n\nvia environment
\n\n IMAGE_BUILD_CMD=\"buildah bud\" make\nby overriding the variable value
\n\n make IMAGE_BUILD_CMD=\"buildah bud\"\nUnit tests are automatically run as part of the container image build. You can\nalso run them manually in the source code tree by running:
\n\nmake test\nEnd-to-end tests are built on top of the e2e test framework of Kubernetes, and,\nthey required a cluster to run them on. For running the tests on your test\ncluster you need to specify the kubeconfig to be used:
\n\nmake e2e-test KUBECONFIG=$HOME/.kube/config\nThere are several environment variables that can be used to customize the\ne2e-tests:
\n\n| Variable | \nDescription | \nDefault value | \n
|---|---|---|
| KUBECONFIG | \nKubeconfig for running e2e-tests | \nempty | \n
| E2E_TEST_CONFIG | \nParameterization file of e2e-tests (see example) | \nempty | \n
| E2E_PULL_IF_NOT_PRESENT | \nTrue-ish value makes the image pull policy IfNotPresent (to be used only in e2e tests) | \nfalse | \n
| E2E_TEST_FULL_IMAGE | \nRun e2e-test also against the Full Image tag | \nfalse | \n
| E2E_GINKGO_LABEL_FILTER | \nGinkgo label filter to use for running e2e tests | \nempty | \n
| OPENSHIFT | \nNon-empty value enables OpenShift specific support (only affects e2e tests) | \nempty | \n
For development and debugging it is possible to run nfd-master as a stand-alone\nbinary outside the cluster. The -no-publish flag can be used to prevent\nnfd-master making changes to the nodes. If -no-publish is not set, nfd-master\nalso requires the NODE_NAME environment variable to be set for cleaning up\nstale annotations.
make build\nNODE_NAME=<EXISTING_NODE> ./nfd-master -no-publish -kubeconfig ~/.kube/config\nFor development and debugging it is possible to run nfd-worker as a stand-alone\nbinary outside the cluster. The -no-publish flag can be used to prevent\nnfd-worker from creating NodeFeature objects in the target cluster. If the\n-no-publish is not set, nfd-worker also requires the NODE_NAME and\nKUBERNETES_NAMESPACE environment variables to be defined to create the\nNodeFeature object in the target cluster.
make build\nKUBERNETES_NAMESPACE=default NODE_NAME=nonexistent-node ./bin/nfd-worker -kubeconfig ~/.kube/config\n\n\n\nNOTE: Running nfd-worker locally this way discovers and publishes\nfeatures of the local development system you’re running nfd-worker on.
\n
For development and debugging it is possible to run nfd-topology-updater as a\nstand-alone binary outside the cluster. However, it requires access to the\nkubelet’s local pod-resources socket and the kubelet http api so in practice it\nneeds to be run on a host acting as a Kubernetes node and thus running\nkubelet. Running kubelet with --read-only-port=10255 (or readOnlyPort:\n10255 in config) makes it possible to connect to kubelet without auth-token\n(never do this in a production cluster). Also, the -no-publish flag can be\nused to prevent nfd-topology-updater from creating NodeResourceTopology objects\nin the target cluster. If the -no-publish is not set, nfd-topology-updater\nalso requires the NODE_NAME and KUBERNETES_NAMESPACE environment variables\nto be defined.
make build\nKUBERNETES_NAMESPACE=default NODE_NAME=nonexistent-node ./bin/nfd-topology-updater -kubeconfig ~/.kube/config -kubelet-config-uri http://127.0.0.1:10255\nAnother option for building NFD locally is via Tilt tool, which can build container\nimages, push them to a local registry and reload your Kubernetes pods automatically.\nWhen using Tilt, you don’t have to build container images and re-deploy your pods\nmanually but instead let the Tilt take care of it. Tiltfile is a configuration file\nfor the Tilt and is located at the root directory. To develop NFD with Tilt, follow\nthe steps below.
\n\nStart up node feature discovery development environment\n To start up your Tilt development environment, run at the root of your\n local NFD codebase.
\n\n tilt up\nTilt will start a web interface in the localhost and port 10350. From the\n web interface, you are able to see how NFD worker and master are\n progressing, watch their build and runtime logs. Once your code changes are\n saved locally, Tilt will notice it and re-build the container image from\n the current code, push the image to the registry and re-deploy NFD pods\n with the latest container image.
\nTo override environment variables used in the Tiltfile during image build,\nexport them in your current terminal before starting Tilt.
\n\nexport IMAGE_TAG_NAME=\"v1\"\ntilt up\nThis will override the default value(master) of IMAGE_TAG_NAME variable defined\nin the Tiltfile.
All documentation resides under the\ndocs\ndirectory in the source tree. It is designed to be served as a html site by\nGitHub Pages.
\n\nBuilding the documentation is containerized to fix the build\nenvironment. The recommended way for developing documentation is to run:
\n\nmake site-serve\nThis will build the documentation in a container and serve it under\nlocalhost:4000/ making it easy to verify the results.\nAny changes made to the docs/ will automatically re-trigger a rebuild and are\nreflected in the served content and can be inspected with a browser refresh.
To just build the html documentation run:
\n\nmake site-build\nThis will generate html documentation under docs/_site/.
To quickly view available command line flags execute nfd-topology-updater -help.\nIn a docker container:
docker run gcr.io/k8s-staging-nfd/node-feature-discovery:master \\\nnfd-topology-updater -help\nPrint usage and exit.
\n\nPrint version and exit.
\n\nThe -config flag specifies the path of the nfd-topology-updater\nconfiguration file to use.
Default: /etc/kubernetes/node-feature-discovery/nfd-topology-updater.conf
\n\nExample:
\n\nnfd-topology-updater -config=/opt/nfd/nfd-topology-updater.conf\nThe -no-publish flag makes for a “dry-run” flag for nfd-topology-updater.\nNFD-Topology-Updater runs resource hardware topology detection normally, but\nNodeResourceTopology\nobjects are not created or updated.
Default: false
\n\nExample:
\n\nnfd-topology-updater -no-publish\nThe -oneshot flag causes nfd-topology-updater to exit after one pass of\nresource hardware topology detection.
Default: false
\n\nExample:
\n\nnfd-topology-updater -oneshot -no-publish\nDEPRECATED: Will be removed in NFD v0.17 and replaced by -port.
The -metrics flag specifies the port on which to expose\nPrometheus metrics. Setting this to 0 disables the\nmetrics server on nfd-topology-updater.
Default: 8081
\n\nExample:
\n\nnfd-topology-updater -metrics=12345\nThe -sleep-interval specifies the interval between resource hardware\ntopology re-examination (and CR updates). zero means no CR updates on interval basis.
Default: 60s
\n\nExample:
\n\nnfd-topology-updater -sleep-interval=1h\nThe -watch-namespace specifies the namespace to ensure that resource\nhardware topology examination only happens for the pods running in the\nspecified namespace. Pods that are not running in the specified namespace\nare not considered during resource accounting. This is particularly useful\nfor testing/debugging purpose. A “*” value would mean that all the pods would\nbe considered during the accounting process.
Default: “*”
\n\nExample:
\n\nnfd-topology-updater -watch-namespace=rte\nThe -kubelet-config-uri specifies the path to the Kubelet’s configuration.\nNote that the URi could either be a local host file or an HTTP endpoint.
Default: https://${NODE_ADDRESS}:10250/configz
Example:
\n\nnfd-topology-updater -kubelet-config-uri=file:///var/lib/kubelet/config.yaml\nThe -api-auth-token-file specifies the path to the api auth token file\nwhich is used to retrieve Kubelet’s configuration from Kubelet secure port,\nonly taking effect when -kubelet-config-uri is https.\nNote that this token file must bind to a role that has the get capability to\nnodes/proxy resources.
Default: /var/run/secrets/kubernetes.io/serviceaccount/token
Example:
\n\nnfd-topology-updater -token-file=/var/run/secrets/kubernetes.io/serviceaccount/token\nThe -podresources-socket specifies the path to the Unix socket where kubelet\nexports a gRPC service to enable discovery of in-use CPUs and devices, and to\nprovide metadata for them.
Default: /host-var/lib/kubelet/pod-resources/kubelet.sock
\n\nExample:
\n\nnfd-topology-updater -podresources-socket=/var/lib/kubelet/pod-resources/kubelet.sock\nEnables compute and report the pod set fingerprint in the NRT.\nA pod fingerprint is a compact representation of the “node state” regarding resources.
\n\nDefault: true
Example:
\n\nnfd-topology-updater -pods-fingerprint=false\nThe -kubelet-state-dir specifies the path to the Kubelet state directory,\nwhere state and checkpoint files are stored.\nThe files are mount as read-only and cannot be change by the updater.\nEnabled by default.\nPassing an empty string will disable the watching.
Default: /host-var/lib/kubelet
\n\nExample:
\n\nnfd-topology-updater -kubelet-state-dir=/var/lib/kubelet\nNFD-Topology-Updater is preferably run as a Kubernetes DaemonSet.\nThis assures re-examination on regular intervals\nand/or per pod life-cycle events, capturing changes in the allocated\nresources and hence the allocatable resources on a per-zone basis by updating\nNodeResourceTopology custom resources.\nIt makes sure that new NodeResourceTopology instances are created for each new\nnodes that get added to the cluster.
\n\nBecause of the design and implementation of Kubernetes, only resources exclusively\nallocated to Guaranteed Quality of Service\npods will be accounted.\nThis includes\nCPU cores,\nmemory\nand\ndevices.
\n\nWhen run as a daemonset, nodes are re-examined for the allocated resources\n(to determine the information of the allocatable resources on a per-zone basis\nwhere a zone can be a NUMA node) at an interval specified using the\n-sleep-interval\noption. The default sleep interval is set to 60s\nwhich is the value when no -sleep-interval is specified.\nThe re-examination can be disabled by setting the sleep-interval to 0.
Another option is to configure the updater to update\nthe allocated resources per pod life-cycle events.\nThe updater will monitor the checkpoint file stated in\n-kubelet-state-dir\nand triggers an update for every change occurs in the files.
In addition, it can avoid examining specific allocated resources\ngiven a configuration of resources to exclude via -excludeList
Kubelet PodResource API with the\nGetAllocatableResources functionality enabled is a\nprerequisite for nfd-topology-updater to be able to run (i.e. Kubernetes v1.21\nor later is required).
\n\nPreceding Kubernetes v1.23, the kubelet must be started with\n--feature-gates=KubeletPodResourcesGetAllocatable=true.
Starting from Kubernetes v1.23, the KubeletPodResourcesGetAllocatable\nfeature gate. is enabled by default
NFD-Topology-Updater supports configuration through a configuration file. The\ndefault location is /etc/kubernetes/node-feature-discovery/topology-updater.conf,\nbut, this can be changed by specifying the-config command line flag.
Topology-Updater configuration file is read inside the container,\nand thus, Volumes and VolumeMounts are needed\nto make your configuration available for NFD.\nThe preferred method is to use a ConfigMap\nwhich provides easy deployment and re-configurability.
\n\nThe provided deployment templates create an empty configmap\nand mount it inside the nfd-topology-updater containers.
\n\nIn Helm deployments,\nTopology Updater parameters\ntoplogyUpdater.config can be used to edit the respective configuration.
In Kustomize deployments, modify the nfd-worker-conf ConfigMap with a custom\noverlay.
See\nnfd-topology-updater configuration file reference\nfor more details.\nThe (empty-by-default)\nexample config\ncontains all available configuration options and can be used as a reference\nfor creating a configuration.
\n\n\n","dir":"/usage/","name":"nfd-topology-updater.md","path":"usage/nfd-topology-updater.md","url":"/usage/nfd-topology-updater.html"},{"title":"Contributing","layout":"default","sort":6,"content":"You can reach us via the following channels:
\n\nThis is a\nSIG-node\nsubproject, hosted under the\nKubernetes SIGs organization in Github.\nThe project was established in 2016 and was migrated to Kubernetes SIGs in 2018.
\n\nThis is open source software released under the Apache 2.0 License.
\n","dir":"/contributing/","name":"index.md","path":"contributing/index.md","url":"/contributing/"},{"title":"Metrics","layout":"default","sort":6,"content":"Metrics are configured to be exposed using prometheus operator\nAPI’s by default. If you want to expose metrics using the prometheus operator\nAPI’s you need to install the prometheus operator in your cluster.\nBy default NFD Master and Worker expose metrics on port 8081.
\n\nThe exposed metrics are
\n\n| Metric | \nType | \nDescription | \n
|---|---|---|
nfd_master_build_info | \n Gauge | \nVersion from which nfd-master was built | \n
nfd_worker_build_info | \n Gauge | \nVersion from which nfd-worker was built | \n
nfd_gc_build_info | \n Gauge | \nVersion from which nfd-gc was built | \n
nfd_topology_updater_build_info | \n Gauge | \nVersion from which nfd-topology-updater was built | \n
nfd_master_node_update_requests_total | \n Counter | \nNumber of node update requests received by the master over gRPC | \n
nfd_master_node_updates_total | \n Counter | \nNumber of nodes updated | \n
nfd_master_node_feature_group_update_requests_total | \n Counter | \nNumber of cluster feature update requests processed by the master | \n
nfd_master_node_update_failures_total | \n Counter | \nNumber of nodes update failures | \n
nfd_master_node_labels_rejected_total | \n Counter | \nNumber of nodes labels rejected by nfd-master | \n
nfd_master_node_extendedresources_rejected_total | \n Counter | \nNumber of nodes extended resources rejected by nfd-master | \n
nfd_master_node_taints_rejected_total | \n Counter | \nNumber of nodes taints rejected by nfd-master | \n
nfd_master_nodefeaturerule_processing_duration_seconds | \n Histogram | \nTime taken to process NodeFeatureRule objects | \n
nfd_master_nodefeaturerule_processing_errors_total | \n Counter | \nNumber or errors encountered while processing NodeFeatureRule objects | \n
nfd_worker_feature_discovery_duration_seconds | \n Histogram | \nTime taken to discover features on a node | \n
nfd_topology_updater_scan_errors_total | \n Counter | \nNumber of errors in scanning resource allocation of pods. | \n
nfd_gc_objects_deleted_total | \n Counter | \nNumber of NodeFeature and NodeResourceTopology objects garbage collected. | \n
nfd_gc_object_delete_failures_total | \n Counter | \nNumber of errors in deleting NodeFeature and NodeResourceTopology objects. | \n
To deploy NFD with metrics enabled using kustomize, you can use the\nprometheus overlay.
\n\nBy default metrics are enabled when deploying NFD via Helm. To enable Prometheus\nto scrape metrics from NFD, you need to pass the following values to Helm:
\n\n--set prometheus.enable=true\nFor more info on Helm deployment, see Helm.
\n\nIt is recommended to specify\n--set prometheus.prometheusSpec.podMonitorSelectorNilUsesHelmValues=false\nwhen deploying prometheus-operator via Helm to enable the prometheus-operator\nto scrape metrics from any PodMonitor.
or setting labels on the PodMonitor via the helm parameter prometheus.labels\nto control which Prometheus instances will scrape this PodMonitor.
NFD contains an example Grafana dashboard. You can import\nexamples/grafana-dashboard.json\nto your Grafana instance to visualize the NFD metrics.
See the\nsample configuration file\nfor a full example configuration.
\n\nThe excludeList specifies a key-value map of allocated resources\nthat should not be examined by the topology-updater\nagent per node.\nEach key is a node name with a value as a list of resources\nthat should not be examined by the agent for that specific node.
Default: empty
\n\nExample:
\n\nexcludeList:\n nodeA: [hugepages-2Mi]\n nodeB: [memory]\n nodeC: [cpu, hugepages-2Mi]\nexcludeList.* is a special value that use to specify all nodes.\nA resource that would be listed under this key, would be excluded from all nodes.
Default: empty
\n\nExample:
\n\nexcludeList:\n '*': [hugepages-2Mi]\nNFD-GC (NFD Garbage-Collector) is preferably run as a Kubernetes deployment\nwith one replica. It makes sure that all\nNodeFeature and\nNodeResourceTopology objects\nhave corresponding nodes and removes stale objects for non-existent nodes.
\n\nThe daemon watches for Node deletion events and removes NodeFeature and\nNodeResourceTopology objects upon them. It also runs periodically to make sure\nno node delete event was missed and to remove any NodeFeature or\nNodeResourceTopology objects that were created without corresponding node. The\ndefault garbage collector interval is set to 1h which is the value when no\n-gc-interval is specified.
\n\nIn Helm deployments see\ngarbage collector parameters\nfor altering the nfd-gc configuration.
\n","dir":"/usage/","name":"nfd-gc.md","path":"usage/nfd-gc.md","url":"/usage/nfd-gc.html"},{"title":"Garbage Collector Cmdline Reference","layout":"default","sort":7,"content":"To quickly view available command line flags execute nfd-gc -help.\nIn a docker container:
docker run gcr.io/k8s-staging-nfd/node-feature-discovery:master \\\nnfd-gc -help\nPrint usage and exit.
\n\nPrint version and exit.
\n\nThe -gc-interval specifies the interval between periodic garbage collector runs.
Default: 1h
\n\nExample:
\n\nnfd-gc -gc-interval=1h\nNFD uses some Kubernetes custom resources.
\n\nNodeFeature is an NFD-specific custom resource for communicating node\nfeatures and node labeling requests. The nfd-master pod watches for NodeFeature\nobjects, labels nodes as specified and uses the listed features as input when\nevaluating NodeFeatureRules. NodeFeature objects can be\nused for implementing 3rd party extensions (see\ncustomization guide for more\ndetails).
\n\napiVersion: nfd.k8s-sigs.io/v1alpha1\nkind: NodeFeature\nmetadata:\n labels:\n nfd.node.kubernetes.io/node-name: node-1\n name: node-1-vendor-features\nspec:\n features:\n instances:\n vendor.device:\n elements:\n - attributes:\n model: \"xpu-1\"\n memory: \"4000\"\n type: \"fast\"\n - attributes:\n model: \"xpu-2\"\n memory: \"16000\"\n type: \"slow\"\n labels:\n vendor-xpu-present: \"true\"\nNodeFeatureGroup is an NFD-specific custom resource that is designed for\ngrouping nodes based on their features. NFD-Master watches for NodeFeatureGroup\nobjects in the cluster and updates the status of the NodeFeatureGroup object\nwith the list of nodes that match the feature group rules. The NodeFeatureGroup\nrules follow the same syntax as the NodeFeatureRule rules.
\n\napiVersion: nfd.k8s-sigs.io/v1alpha1\nkind: NodeFeatureGroup\nmetadata:\n name: node-feature-group-example\nspec:\n featureGroupRules:\n - name: \"node has kernel version discovered\"\n matchFeatures:\n - feature: kernel.version\n matchExpressions:\n major: {op: Exists}\nNodeFeatureGroup API is an alpha feature and disabled by default in NFD version\nmaster. For more details and examples see the\ncustomization guide.
\n\nNodeFeatureRule is an NFD-specific custom resource that is designed for\nrule-based custom labeling of nodes. NFD-Master watches for NodeFeatureRule\nobjects in the cluster and labels nodes according to the rules within. Some use\ncases are e.g. application specific labeling in a specific environments or\nbeing distributed by hardware vendors to create specific labels for their\ndevices.
\n\napiVersion: nfd.k8s-sigs.io/v1alpha1\nkind: NodeFeatureRule\nmetadata:\n name: example-rule\nspec:\n rules:\n - name: \"example rule\"\n labels:\n \"example-custom-feature\": \"true\"\n # Label is created if all of the rules below match\n matchFeatures:\n # Match if \"veth\" kernel module is loaded\n - feature: kernel.loadedmodule\n matchExpressions:\n veth: {op: Exists}\n # Match if any PCI device with vendor 8086 exists in the system\n - feature: pci.device\n matchExpressions:\n vendor: {op: In, value: [\"8086\"]}\nSee the\nCustomization guide\nfor full documentation of the NodeFeatureRule resource and its usage.
\n\nThe\ndeployment/nodefeaturerule/samples/\ndirectory contains sample NodeFeatureRule objects that replicate the built-in\ndefault feature labels generated by NFD. The sample rules can be used as a base\nto customize NFD feature labels. To use them in place of the the NFD built-in\nlabels, the corresponding feature source(s) of nfd-worker should be disabled\nwith the\ncore.labelSources\nconfiguration option.
When run with NFD-Topology-Updater, NFD creates NodeResourceTopology objects\ncorresponding to node resource hardware topology such as:
\n\napiVersion: topology.node.k8s.io/v1alpha1\nkind: NodeResourceTopology\nmetadata:\n name: node1\ntopologyPolicies: [\"SingleNUMANodeContainerLevel\"]\nzones:\n - name: node-0\n type: Node\n resources:\n - name: cpu\n capacity: 20\n allocatable: 16\n available: 10\n - name: vendor/nic1\n capacity: 3\n allocatable: 3\n available: 3\n - name: node-1\n type: Node\n resources:\n - name: cpu\n capacity: 30\n allocatable: 30\n available: 15\n - name: vendor/nic2\n capacity: 6\n allocatable: 6\n available: 6\n - name: node-2\n type: Node\n resources:\n - name: cpu\n capacity: 30\n allocatable: 30\n available: 15\n - name: vendor/nic1\n capacity: 3\n allocatable: 3\n available: 3\nThe NodeResourceTopology objects created by NFD can be used to gain insight\ninto the allocatable resources along with the granularity of those resources at\na per-zone level (represented by node-0 and node-1 in the above example) or can\nbe used by an external entity (e.g. topology-aware scheduler plugin) to take an\naction based on the gathered information.
\n\n\n","dir":"/usage/","name":"custom-resources.md","path":"usage/custom-resources.md","url":"/usage/custom-resources.html"},{"title":"Kubectl plugin cmdline reference","layout":"default","sort":8,"content":"To quickly view available command line flags execute kubectl nfd -help.
Print usage and exit.
\n\nValidate a NodeFeatureRule file.
\n\nThe --nodefeature-file flag specifies the path to the NodeFeatureRule file\nto validate.
Test a NodeFeatureRule file against a node without applying it.
\n\nThe --kubeconfig flag specifies the path to the kubeconfig file to use for\nCLI requests.
The --namespace flag specifies the namespace to use for CLI requests.\nDefault: default.
The --nodename flag specifies the name of the node to test the\nNodeFeatureRule against.
The --nodefeaturerule-file flag specifies the path to the NodeFeatureRule file\nto test.
Process a NodeFeatureRule file against a NodeFeature file.
\n\nThe --nodefeaturerule-file flag specifies the path to the NodeFeatureRule file\nto test.
The --nodefeature-file flag specifies the path to the NodeFeature file to test.
NFD provides multiple extension points for vendor and application specific\nlabeling:
\n\nNodeFeature objects can be\nused to communicate “raw” node features and node labeling requests to\nnfd-master.NodeFeatureRule objects provide a way to\ndeploy custom labeling rules via the Kubernetes API.local feature source of nfd-worker creates\nlabels by reading text files.custom feature source of nfd-worker creates\nlabels based on user-specified rules.NodeFeature objects provide a way for 3rd party extensions to advertise custom\nfeatures, both as “raw” features that serve as input to\nNodeFeatureRule objects and as feature\nlabels directly.
\n\nNote that RBAC rules must be created for each extension for them to be able to\ncreate and manipulate NodeFeature objects in their namespace.
\n\nConsider the following referential example:
\n\napiVersion: nfd.k8s-sigs.io/v1alpha1\nkind: NodeFeature\nmetadata:\n labels:\n nfd.node.kubernetes.io/node-name: node-1\n name: vendor-features-for-node-1\nspec:\n # Features for NodeFeatureRule matching\n features:\n flags:\n vendor.flags:\n elements:\n feature-x: {}\n feature-y: {}\n attributes:\n vendor.config:\n elements:\n setting-a: \"auto\"\n knob-b: \"123\"\n instances:\n vendor.devices:\n elements:\n - attributes:\n model: \"dev-1000\"\n vendor: \"acme\"\n - attributes:\n model: \"dev-2000\"\n vendor: \"acme\"\n # Labels to be created\n labels:\n vendor.io/feature.enabled: \"true\"\nThe object targets node named node-1. It lists two “flag type” features under\nthe vendor.flags domain, two “attribute type” features and under the\nvendor.config domain and two “instance type” features under the\nvendor.devices domain. These features will not be directly affecting the node\nlabels but they will be used as input when the\nNodeFeatureRule objects are evaluated.
In addition, the example requests directly the\nvendor.io/feature.enabled=true node label to be created.
The nfd.node.kubernetes.io/node-name=<node-name> must be in place for each\nNodeFeature object as NFD uses it to determine the node which it is targeting.
Features have three different types:
\n\nNodeFeatureRule objects provide an easy way to create vendor or application\nspecific labels and taints. It uses a flexible rule-based mechanism for creating\nlabels and optionally taints based on node features.
Consider the following referential example:
\n\napiVersion: nfd.k8s-sigs.io/v1alpha1\nkind: NodeFeatureRule\nmetadata:\n name: my-sample-rule-object\nspec:\n rules:\n - name: \"my sample rule\"\n labels:\n \"feature.node.kubernetes.io/my-sample-feature\": \"true\"\n matchFeatures:\n - feature: kernel.loadedmodule\n matchExpressions:\n dummy: {op: Exists}\n - feature: kernel.config\n matchExpressions:\n X86: {op: In, value: [\"y\"]}\nIt specifies one rule which creates node label\nfeature.node.kubernetes.io/my-sample-feature=true if both of the following\nconditions are true (matchFeatures implements a logical AND over the\nmatchers):
dummy network driver module has been loaded=yCreate a NodeFeatureRule with a yaml file:
kubectl apply -f https://raw.githubusercontent.com/kubernetes-sigs/node-feature-discovery/master/examples/nodefeaturerule.yaml\nNow, on X86 platforms the feature label appears after doing modprobe dummy on\na system and correspondingly the label is removed after rmmod dummy. Note a\nre-labeling delay up to the sleep-interval of nfd-worker (1 minute by default).
See Feature rule format for detailed description of\navailable fields and how to write labeling rules.
\n\nIn some circumstances, it is desirable to keep nodes with specialized hardware\naway from running general workload and instead leave them for workloads that\nneed the specialized hardware. One way to achieve it is to taint the nodes with\nthe specialized hardware and add corresponding toleration to pods that require\nthe special hardware. NFD offers node tainting functionality which is disabled\nby default. User can define one or more custom taints via the taints field of\nthe NodeFeatureRule CR. The same rule-based mechanism is applied here and the\nNFD taints only rule matching nodes.
To enable the tainting feature, --enable-taints flag needs to be set to true.\nIf the flag --enable-taints is set to false (i.e. disabled), taints defined in\nthe NodeFeatureRule CR have no effect and will be ignored by the NFD master.
See documentation of the taints field for detailed description how\nto specify taints in the NodeFeatureRule object.
\n\n\n\n\nNOTE: Before enabling any taints, make sure to edit nfd-worker daemonset\nto tolerate the taints to be created. Otherwise, already running pods that do\nnot tolerate the taint are evicted immediately from the node including the\nnfd-worker pod.
\n
NodeFeatureGroup API is an alpha feature and disabled by default in NFD version\nmaster. Use the\nNodeFeatureAPI feature\ngate to enable it.
\n\nNodeFeatureGroup objects provide a way to create node groups that share the\nsame set of features. The NodeFeatureGroup object spec consists of a list of\nNodeFeatureRule that follow the same format as the NodeFeatureRule,\nbut the difference in this case is that nodes that match any of the rules in the\nNodeFeatureGroup will be listed in the NodeFeatureGroup status.
Consider the following referential example:
\n\napiVersion: nfd.k8s-sigs.io/v1alpha1\nkind: NodeFeatureGroup\nmetadata:\n name: node-feature-group-example\nspec:\n featureGroupRules:\n - name: \"kernel version\"\n matchFeatures:\n - feature: kernel.version\n matchExpressions:\n major: {op: In, value: [\"6\"]}\nstatus:\n nodes:\n - name: node-1\n - name: node-2\n - name: node-3\nThe object specifies a group of nodes that share the same\nkernel.version.major (Linux kernel v6.x).
Create a NodeFeatureGroup with a yaml file:
kubectl apply -f https://raw.githubusercontent.com/kubernetes-sigs/node-feature-discovery/master/examples/nodefeaturegroup.yaml\nSee Feature rule format for detailed description of\navailable fields and how to write group filtering rules.
\n\nNFD-Worker has a special feature source named local which is an integration\npoint for external feature detectors. It provides a mechanism for pluggable\nextensions, allowing the creation of new user-specific features and even\noverriding built-in labels.
The local feature source uses feature files. The features discovered by the\nlocal source can further be used in label rules specified in\nNodeFeatureRule objects and\nthe custom feature source.
\n\n\nNOTE: Be careful when creating and/or updating feature files\nwhile NFD is running. To avoid race conditions you should write\ninto a temporary file, and atomically create/update the original file by\ndoing a file rename operation. NFD ignores dot files,\nso temporary file can be written to the same directory and renamed\n(
\n.my.feature->my.feature) once file is complete. Both file names should\n(obviously) be unique for the given application.
Consider a plaintext file\n/etc/kubernetes/node-feature-discovery/features.d/my-features\nhaving the following contents:
feature.node.kubernetes.io/my-feature.1\nfeature.node.kubernetes.io/my-feature.2=myvalue\nvendor.io/my-feature.3=456\nThis will translate into the following node labels:
\n\nfeature.node.kubernetes.io/my-feature.1: \"true\"\nfeature.node.kubernetes.io/my-feature.2: \"myvalue\"\nvendor.io/my-feature.3: \"456\"\nThe local source reads files found in\n/etc/kubernetes/node-feature-discovery/features.d/. File content is parsed\nand translated into node labels, see the input format below.
The feature files are expected to contain features in simple\nkey-value pairs, separated by newlines:
\n\n# This is a comment\n<key>[=<value>]\nThe label value defaults to true, if not specified.
Label namespace must be specified with <namespace>/<name>[=<value>].
\n\n\nNOTE: The feature file size limit it 64kB. The feature file will be\nignored if the size limit is exceeded.
\n
Comment lines (starting with #) are ignored.
Adding following line anywhere to feature file defines date when\nits content expires / is ignored:
\n\n# +expiry-time=2023-07-29T11:22:33Z\nAlso, the expiry-time value would stay the same during the processing of the\nfeature file until another expiry-time directive is encountered.\nConsidering the following file:
\n\n# +expiry-time=2012-07-28T11:22:33Z\nvendor.io/feature1=featureValue\n\n# +expiry-time=2080-07-28T11:22:33Z\nvendor.io/feature2=featureValue2\n\n# +expiry-time=2070-07-28T11:22:33Z\nvendor.io/feature3=featureValue3\n\n# +expiry-time=2002-07-28T11:22:33Z\nvendor.io/feature4=featureValue4\nAfter processing the above file, only vendor.io/feature2 and\nvendor.io/feature3 would be included in the list of accepted features.
\n\n\nNOTE: The time format supported is RFC3339. Also, the
\nexpiry-time\ntag is only evaluated in each re-discovery period, and the expiration of\nnode labels is not tracked.
To exclude specific features from the local.feature Feature, you can use the\n# +no-feature directive. The # +no-label directive causes the feature to\nbe excluded from the local.label Feature and a node label not to be generated.
Considering the following file:
\n\n# +no-feature\nvendor.io/label-only=value\n\nvendor.io/my-feature=value\n\nvendor.io/foo=bar\n\n# +no-label\nfoo=baz\nProcessing the above file would result in the following Features:
\n\nlocal.features:\n foo: baz\n vendor.io/my-feature: value\nlocal.labels:\n vendor.io/label-only: value\n vendor.io/my-feature: value\nand the following labels added to the Node:
\n\nvendor.io/label-only=value\nvendor.io/my-feature=value\n\n\n\nNOTE: use of unprefixed label names (like
\nfoo=bar) should not be used.\nIn NFD master unprefixed names will be automatically prefixed\nwithfeature.node.kubernetes.io/but this will change in a future version\n(see\nautoDefaultNs config option.\nUnprefixed names for plain Features (tagged with# +no-label) can be used\nwithout restrictions, however.
The standard NFD deployments contain hostPath mounts for\n/etc/kubernetes/node-feature-discovery/features.d/, making these directories\nfrom the host available inside the nfd-worker container.
One use case for the feature files is detecting features in other\nPods outside NFD, e.g. in Kubernetes device plugins. By using the same\nhostPath mounts /etc/kubernetes/node-feature-discovery/features.d/\nin the side-car (e.g. device plugin) creates a shared area for\ndeploying feature files to NFD.
The custom feature source in nfd-worker provides a rule-based mechanism for\nlabel creation, similar to the\nNodeFeatureRule objects. The difference is\nthat the rules are specified in the worker configuration instead of a\nKubernetes API object.
See worker configuration\nfor instructions how to set-up and manage the worker configuration.
\n\nConsider the following referential configuration for nfd-worker:
\n\ncore:\n labelSources: [\"custom\"]\nsources:\n custom:\n - name: \"my sample rule\"\n labels:\n \"feature.node.kubenernetes.io/my-sample-feature\": \"true\"\n matchFeatures:\n - feature: kernel.loadedmodule\n matchExpressions:\n dummy: {op: Exists}\n - feature: kernel.config\n matchExpressions:\n X86: {op: In, value: [\"y\"]}\nIt specifies one rule which creates node label\nfeature.node.kubenernetes.io/my-sample-feature=true if both of the following\nconditions are true (matchFeatures implements a logical AND over the\nmatchers):
dummy network driver module has been loaded=yIn addition, the configuration only enables the custom source, disabling all\nbuilt-in labels.
Now, on X86 platforms the feature label appears after doing modprobe dummy on\na system and correspondingly the label is removed after rmmod dummy. Note a\nre-labeling delay up to the sleep-interval of nfd-worker (1 minute by default).
In addition to the rules defined in the nfd-worker configuration file, the\ncustom feature source can read more configuration files located in the\n/etc/kubernetes/node-feature-discovery/custom.d/ directory. This makes more\ndynamic and flexible configuration easier.
As an example, consider having file\n/etc/kubernetes/node-feature-discovery/custom.d/my-rule.yaml with the\nfollowing content:
- name: \"my e1000 rule\"\n labels:\n \"feature.node.kubenernetes.io/e1000.present\": \"true\"\n matchFeatures:\n - feature: kernel.loadedmodule\n matchExpressions:\n e1000: {op: Exists}\nThis simple rule will create feature.node.kubenernetes.io/e1000.present=true\nlabel if the e1000 kernel module has been loaded.
The\nsamples/custom-rules\nkustomize overlay sample contains an example for deploying a custom rule from a\nConfigMap.
Feature labels have the following format:
\n\n<namespace>/<name> = <value>\nThe namespace part (i.e. prefix) of the labels is controlled by nfd:
\n\nfeature.node.kubernetes.io.-deny-label-ns\ncommand line flag of nfd-master\n -extra-label-ns\ncommand line flag of nfd-master.\ne.g: nfd-master -deny-label-ns=\"*\" -extra-label-ns=example.comThis section describes the rule format used in\nNodeFeatureRule objects and in the\nconfiguration of the custom feature source.
It is based on a generic feature matcher that covers all features discovered by\nnfd-worker. The rules rely on a unified data model of the available features\nand a generic expression-based format. Features that can be used in the rules\nare described in detail in available features below.
\n\nTake this rule as a referential example:
\n\n - name: \"my feature rule\"\n labels:\n \"feature.node.kubernetes.io/my-special-feature\": \"my-value\"\n matchFeatures:\n - feature: cpu.cpuid\n matchExpressions:\n AVX512F: {op: Exists}\n - feature: kernel.version\n matchExpressions:\n major: {op: In, value: [\"5\"]}\n minor: {op: Gt, value: [\"1\"]}\n - feature: pci.device\n matchExpressions:\n vendor: {op: In, value: [\"8086\"]}\n class: {op: In, value: [\"0200\"]}\nThis will yield feature.node.kubernetes.io/my-special-feature=my-value node\nlabel if all of these are true (matchFeatures implements a logical AND over\nthe matchers):
The .name field is required and used as an identifier of the rule.
The .labels is a map of the node labels to create if the rule matches.
Take this rule as a referential example:
\n\napiVersion: nfd.k8s-sigs.io/v1alpha1\nkind: NodeFeatureRule\nmetadata:\n name: my-sample-rule-object\nspec:\n rules:\n - name: \"my dynamic label value rule\"\n labels:\n feature.node.kubernetes.io/linux-lsm-enabled: \"@kernel.config.LSM\"\n feature.node.kubernetes.io/custom-label: \"customlabel\"\nLabel linux-lsm-enabled uses the @ notation for dynamic values.\nThe value of the label will be the value of the attribute LSM\nof the feature kernel.config.
The @<feature-name>.<element-name> format can be used to inject values of\ndetected features to the label. See\navailable features for possible values to use.
This will yield into the following node label:
\n\n labels:\n ...\n feature.node.kubernetes.io/linux-lsm-enabled: apparmor\n feature.node.kubernetes.io/custom-label: \"customlabel\"\nThe .labelsTemplate field specifies a text template for dynamically creating\nlabels based on the matched features. See templating for\ndetails.
\n\n\nNOTE: The
\nlabelsfield has priority overlabelsTemplate, i.e.\nlabels specified in thelabelsfield will override anything\noriginating fromlabelsTemplate.
The .annotations field is a list of features to be advertised as node\nannotations.
Take this rule as a referential example:
\n\napiVersion: nfd.k8s-sigs.io/v1alpha1\nkind: NodeFeatureRule\nmetadata:\n name: feature-annotations-example\nspec:\n rules:\n - name: \"annotation-example\"\n annotations:\n feature.node.kubernetes.io/defaul-ns-annotation: \"foo\"\n custom.vendor.io/feature: \"baz\"\n matchFeatures:\n - feature: kernel.version\n matchExpressions:\n major: {op: Exists}\nThis will yield into the following node annotations:
\n\n annotations:\n ...\n feature.node.kubernetes.io/defaul-ns-annotation: \"foo\"\n custom.vendor.io/feature: \"baz\"\n ...\nNFD enforces some limitations to the namespace (or prefix)/ of the annotations:
\n\nkubernetes.io/ and its sub-namespaces (like sub.ns.kubernetes.io/) cannot\ngenerally be usedfeature.node.kubernetes.io/ and its sub-namespaces\n(like sub.ns.feature.node.kubernetes.io)my-annotation) should not be used. In NFD master unprefixed names will be automatically prefixed with\nfeature.node.kubernetes.io/ but this will change in a future version (see\nautoDefaultNs config option.\n\n\nNOTE: The
\nannotationsfield has will only advertise features via node\nannotations the features won’t be advertised as node labels unless they are\nspecified in thelabelsfield.
taints is a list of taint entries and each entry can have key, value and effect,\nwhere the value is optional. Effect could be NoSchedule, PreferNoSchedule\nor NoExecute. To learn more about the meaning of these effects, check out k8s documentation.
Example NodeFeatureRule with taints:
\n\napiVersion: nfd.k8s-sigs.io/v1alpha1\nkind: NodeFeatureRule\nmetadata:\n name: my-sample-rule-object\nspec:\n rules:\n - name: \"my sample taint rule\"\n taints:\n - effect: PreferNoSchedule\n key: \"feature.node.kubernetes.io/special-node\"\n value: \"true\"\n - effect: NoExecute\n key: \"feature.node.kubernetes.io/dedicated-node\"\n matchFeatures:\n - feature: kernel.loadedmodule\n matchExpressions:\n dummy: {op: Exists}\n - feature: kernel.config\n matchExpressions:\n X86: {op: In, value: [\"y\"]}\nIn this example, if the my sample taint rule rule is matched,\nfeature.node.kubernetes.io/pci-0300_1d0f.present=true:NoExecute\nand feature.node.kubernetes.io/cpu-cpuid.ADX:NoExecute taints are set on the node.
There are some limitations to the namespace part (i.e. prefix/) of the taint\nkey:
\n\nkubernetes.io/ and its sub-namespaces (like sub.ns.kubernetes.io/) cannot\ngenerally be usedfeature.node.kubernetes.io/ and its sub-namespaces\n(like sub.ns.feature.node.kubernetes.io)foo) keys are disallowed\n\n\nNOTE: taints field is not available for the custom rules of nfd-worker\nand only for NodeFeatureRule objects.
\n
The .vars field is a map of values (key-value pairs) to store for subsequent\nrules to use. In other words, these are variables that are not advertised as\nnode labels. See backreferences for more details on the\nusage of vars.
The .extendedResources field is a list of extended resources to advertise.\nSee extended resources for more details.
Take this rule as a referential example:
\n\napiVersion: nfd.k8s-sigs.io/v1alpha1\nkind: NodeFeatureRule\nmetadata:\n name: my-extended-resource-rule\nspec:\n rules:\n - name: \"my extended resource rule\"\n extendedResources:\n vendor.io/dynamic: \"@kernel.version.major\"\n vendor.io/static: \"123\"\n matchFeatures:\n - feature: kernel.version\n matchExpressions:\n major: {op: Exists}\nThe extended resource vendor.io/dynamic is defined in the form @feature.attribute.\nThe value of the extended resource will be the value of the attribute major\nof the feature kernel.version.
The @<feature-name>.<element-name> format can be used to inject values of\ndetected features to the extended resource. See\navailable features for possible values to use. Note that\nthe value must be eligible as a\nKubernetes resource quantity.
This will yield into the following node status:
\n\n allocatable:\n ...\n vendor.io/dynamic: \"5\"\n vendor.io/static: \"123\"\n ...\n capacity:\n ...\n vendor.io/dynamic: \"5\"\n vendor.io/static: \"123\"\n ...\nThere are some limitations to the namespace part (i.e. prefix)/ of the Extended\nResources names:
\n\nkubernetes.io/ and its sub-namespaces (like sub.ns.kubernetes.io/) cannot\ngenerally be usedfeature.node.kubernetes.io/ and its sub-namespaces\n(like sub.ns.feature.node.kubernetes.io)my-er) site.version }} unprefixed names will be\nautomatically prefixed with feature.node.kubernetes.io/ but this will\nchange in a future version (see\nautoDefaultNs config option.\n\n\nNOTE:
\n.extendedResourcesis not supported by the\ncustom feature source – it can only be used in\nNodeFeatureRule objects.
The .varsTemplate field specifies a text template for dynamically creating\nvars based on the matched features. See templating for details\non using templates and backreferences for more details on\nthe usage of vars.
\n\n\nNOTE: The
\nvarsfield has priority overvarsTemplate, i.e.\nvars specified in thevarsfield will override anything originating from\nvarsTemplate.
The .matchFeatures field specifies a feature matcher, consisting of a list of\nfeature matcher terms. It implements a logical AND over the terms i.e. all\nof them must match for the rule to trigger.
matchFeatures:\n - feature: <feature-name>\n matchExpressions:\n <key>:\n op: <op>\n value:\n - <value-1>\n - ...\n matchName:\n op: <op>\n value:\n - <value-1>\n - ...\nThe .matchFeatures[].feature field specifies the feature which to evaluate.
\n\n\nNOTE:If both
\nmatchExpressionsand\nmatchNameare specified, they both must match.
The .matchFeatures[].matchExpressions field is used to match against the\nvalue(s) of a feature. The matchExpressions field consists of a set of\nexpressions, each of which is evaluated against all elements of the specified\nfeature.
matchExpressions:\n <key>:\n op: <op>\n value:\n - <value-1>\n - ...\nIn each MatchExpression the key specifies the name of of the feature element\n(flag and attribute features) or name of the attribute (instance\nfeatures) which to look for. The behavior of MatchExpression depends on the\nfeature type:
<key>The op field specifies the operator to apply. Valid values are described\nbelow.
| Operator | \nNumber of values | \nMatches when | \n
|---|---|---|
In | \n 1 or greater | \nInput is equal to one of the values | \n
NotIn | \n 1 or greater | \nInput is not equal to any of the values | \n
InRegexp | \n 1 or greater | \nValues of the MatchExpression are treated as regexps and input matches one or more of them | \n
Exists | \n 0 | \nThe key exists | \n
DoesNotExist | \n 0 | \nThe key does not exists | \n
Gt | \n 1 | \nInput is greater than the value. Both the input and value must be integer numbers. | \n
Lt | \n 1 | \nInput is less than the value. Both the input and value must be integer numbers. | \n
GtLt | \n 2 | \nInput is between two values. Both the input and value must be integer numbers. | \n
IsTrue | \n 0 | \nInput is equal to “true” | \n
IsFalse | \n 0 | \nInput is equal “false” | \n
The value field of MatchExpression is a list of string arguments to the\noperator.
The .matchFeatures[].matchName field is used to match against the\nname(s) of a feature (whereas the matchExpressions field\nmatches against the value(s). The matchName field consists of a single\nexpression which is evaulated against the name of each element of the specified\nfeature.
matchName:\n op: <op>\n value:\n - <value-1>\n - ...\nThe behavior of matchName depends on the feature type:
The op field specifies the operator to apply. Same operators as for\nmatchExpressions above are available.
| Operator | \nNumber of values | \nMatches | \n
|---|---|---|
In | \n 1 or greater | \nAll name is equal to one of the values | \n
NotIn | \n 1 or greater | \nAll name that is not equal to any of the values | \n
InRegexp | \n 1 or greater | \nAll name that matches any of the values (treated as regexps) | \n
Exists | \n 0 | \nAll elements | \n
Other operators are not practical with matchName (DoesNotExist never\nmatches; Gt,Lt and GtLt are only usable if feature names are integers;\nIsTrue and IsFalse are only usable if the feature name is true or\nfalse).
The value field is a list of string arguments to the operator.
An example:
\n\n matchFeatures:\n - feature: cpu.cpuid\n matchName: {op: InRegexp, value: [\"^AVX\"]}\nThe snippet above would match if any CPUID feature starting with AVX is present\n(e.g. AVX1 or AVX2 or AVX512F etc).
\n\nThe .matchAny field is a list of of matchFeatures\nmatchers. A logical OR is applied over the matchers, i.e. at least one of them\nmust match for the rule to trigger.
Consider the following example:
\n\n matchAny:\n - matchFeatures:\n - feature: kernel.loadedmodule\n matchExpressions:\n kmod-1: {op: Exists}\n - feature: pci.device\n matchExpressions:\n vendor: {op: In, value: [\"0eee\"]}\n class: {op: In, value: [\"0200\"]}\n - matchFeatures:\n - feature: kernel.loadedmodule\n matchExpressions:\n kmod-2: {op: Exists}\n - feature: pci.device\n matchExpressions:\n vendor: {op: In, value: [\"0fff\"]}\n class: {op: In, value: [\"0200\"]}\nThis matches if kernel module kmod-1 is loaded and a network controller from\nvendor 0eee is present, OR, if kernel module kmod-2 has been loaded and a\nnetwork controller from vendor 0fff is present (OR both of these conditions are\ntrue).
\n\nThe following features are available for matching:
\n\n| Feature | \nFeature types | \nElements | \nValue type | \nDescription | \n
|---|---|---|---|---|
cpu.cpuid | \n flag | \n\n | \n | Supported CPU capabilities | \n
| \n | \n | <cpuid-flag> | \n \n | CPUID flag is present | \n
| \n | attribute | \n\n | \n | CPU capability attributes | \n
| \n | \n | AVX10_VERSION | \nint | \nAVX10 vector ISA version (if supported) | \n
cpu.cstate | \n attribute | \n\n | \n | Status of cstates in the intel_idle cpuidle driver | \n
| \n | \n | enabled | \n bool | \n‘true’ if cstates are set, otherwise ‘false’. Does not exist of intel_idle driver is not active. | \n
cpu.model | \n attribute | \n\n | \n | CPU model related attributes | \n
| \n | \n | family | \n int | \nCPU family | \n
| \n | \n | vendor_id | \n string | \nCPU vendor ID | \n
| \n | \n | id | \n int | \nCPU model ID | \n
cpu.pstate | \n attribute | \n\n | \n | State of the Intel pstate driver. Does not exist if the driver is not enabled. | \n
| \n | \n | status | \n string | \nStatus of the driver, possible values are ‘active’ and ‘passive’ | \n
| \n | \n | turbo | \n bool | \n‘true’ if turbo frequencies are enabled, otherwise ‘false’ | \n
| \n | \n | scaling | \n string | \nActive scaling_governor, possible values are ‘powersave’ or ‘performance’. | \n
cpu.rdt | \n attribute | \n\n | \n | Intel RDT capabilities supported by the system | \n
| \n | \n | <rdt-flag> | \n \n | RDT capability is supported, see RDT flags for details | \n
| \n | \n | RDTL3CA_NUM_CLOSID | \n int | \nThe number or available CLOSID (Class of service ID) for Intel L3 Cache Allocation Technology | \n
cpu.security | \n attribute | \n\n | \n | Features related to security and trusted execution environments | \n
| \n | \n | sgx.enabled | \n bool | \ntrue if Intel SGX (Software Guard Extensions) has been enabled, otherwise does not exist | \n
| \n | \n | sgx.epc | \n int | \nThe total amount Intel SGX Encrypted Page Cache memory in bytes. It’s only present if sgx.enabled is true. | \n
| \n | \n | se.enabled | \n bool | \ntrue if IBM Secure Execution for Linux is available and has been enabled, otherwise does not exist | \n
| \n | \n | tdx.enabled | \n bool | \ntrue if Intel TDX (Trusted Domain Extensions) is available on the host and has been enabled, otherwise does not exist | \n
| \n | \n | tdx.total_keys | \n int | \nThe total amount of keys an Intel TDX (Trusted Domain Extensions) host can provide. It’s only present if tdx.enabled is true. | \n
| \n | \n | tdx.protected | \n bool | \ntrue if a guest VM was started using Intel TDX (Trusted Domain Extensions), otherwise does not exist. | \n
| \n | \n | sev.enabled | \n bool | \ntrue if AMD SEV (Secure Encrypted Virtualization) is available on the host and has been enabled, otherwise does not exist | \n
| \n | \n | sev.es.enabled | \n bool | \ntrue if AMD SEV-ES (Encrypted State supported) is available on the host and has been enabled, otherwise does not exist | \n
| \n | \n | sev.snp.enabled | \n bool | \ntrue if AMD SEV-SNP (Secure Nested Paging supported) is available on the host and has been enabled, otherwise does not exist | \n
| \n | \n | sev.asids | \n int | \nThe total amount of AMD SEV address-space identifiers (ASIDs), based on the /sys/fs/cgroup/misc.capacity information. | \n
| \n | \n | sev.encrypted_state_ids | \n int | \nThe total amount of AMD SEV-ES and SEV-SNP supported, based on the /sys/fs/cgroup/misc.capacity information. | \n
cpu.sst | \n attribute | \n\n | \n | Intel SST (Speed Select Technology) capabilities | \n
| \n | \n | bf.enabled | \n bool | \ntrue if Intel SST-BF (Intel Speed Select Technology - Base frequency) has been enabled, otherwise does not exist | \n
cpu.topology | \n attribute | \n\n | \n | CPU topology related features | \n
| \n | \n | hardware_multithreading | \n bool | \nHardware multithreading, such as Intel HTT, is enabled | \n
| \n | \n | socket_count | \n int | \nNumber of CPU Sockets | \n
cpu.coprocessor | \n attribute | \n\n | \n | CPU Coprocessor related features | \n
| \n | \n | nx_gzip | \n bool | \nNest Accelerator GZIP support is enabled | \n
kernel.config | \n attribute | \n\n | \n | Kernel configuration options | \n
| \n | \n | <config-flag> | \n string | \nValue of the kconfig option | \n
kernel.loadedmodule | \n flag | \n\n | \n | Kernel modules loaded on the node as reported by /proc/modules | \n
kernel.enabledmodule | \n flag | \n\n | \n | Kernel modules loaded on the node and available as built-ins as reported by modules.builtin | \n
| \n | \n | mod-name | \n \n | Kernel module <mod-name> is loaded | \n
kernel.selinux | \n attribute | \n\n | \n | Kernel SELinux related features | \n
| \n | \n | enabled | \n bool | \ntrue if SELinux has been enabled and is in enforcing mode, otherwise false | \n
kernel.version | \n attribute | \n\n | \n | Kernel version information | \n
| \n | \n | full | \n string | \nFull kernel version (e.g. ‘4.5.6-7-g123abcde’) | \n
| \n | \n | major | \n int | \nFirst component of the kernel version (e.g. ‘4’) | \n
| \n | \n | minor | \n int | \nSecond component of the kernel version (e.g. ‘5’) | \n
| \n | \n | revision | \n int | \nThird component of the kernel version (e.g. ‘6’) | \n
local.label | \n attribute | \n\n | \n | Labels from feature files, i.e. labels from the local feature source | \n
local.feature | \n attribute | \n\n | \n | Features from feature files, i.e. features from the local feature source | \n
| \n | \n | <label-name> | \n string | \nLabel <label-name> created by the local feature source, value equals the value of the label | \n
memory.nv | \n instance | \n\n | \n | NVDIMM devices present in the system | \n
| \n | \n | <sysfs-attribute> | \n string | \nValue of the sysfs device attribute, available attributes: devtype, mode | \n
memory.numa | \n attribute | \n\n | \n | NUMA nodes | \n
| \n | \n | is_numa | \n bool | \ntrue if NUMA architecture, false otherwise | \n
| \n | \n | node_count | \n int | \nNumber of NUMA nodes | \n
memory.swap | \n attribute | \n\n | \n | Swap enabled on node | \n
| \n | \n | enabled | \n bool | \ntrue if swap partition detected, false otherwise | \n
network.device | \n instance | \n\n | \n | Physical (non-virtual) network interfaces present in the system | \n
| \n | \n | name | \n string | \nName of the network interface | \n
| \n | \n | <sysfs-attribute> | \n string | \nSysfs network interface attribute, available attributes: operstate, speed, sriov_numvfs, sriov_totalvfs | \n
network.virtual | \n instance | \n\n | \n | Virtual network interfaces present in the system | \n
| \n | \n | name | \n string | \nName of the network interface | \n
| \n | \n | <sysfs-attribute> | \n string | \nSysfs network interface attribute, available attributes: operstate, speed | \n
pci.device | \n instance | \n\n | \n | PCI devices present in the system | \n
| \n | \n | <sysfs-attribute> | \n string | \nValue of the sysfs device attribute, available attributes: class, vendor, device, subsystem_vendor, subsystem_device, sriov_totalvfs, iommu_group/type, iommu/intel-iommu/version | \n
storage.block | \n instance | \n\n | \n | Block storage devices present in the system | \n
| \n | \n | name | \n string | \nName of the block device | \n
| \n | \n | <sysfs-attribute> | \n string | \nSysfs network interface attribute, available attributes: dax, rotational, nr_zones, zoned | \n
system.osrelease | \n attribute | \n\n | \n | System identification data from /etc/os-release | \n
| \n | \n | <parameter> | \n string | \nOne parameter from /etc/os-release | \n
system.dmiid | \n attribute | \n\n | \n | DMI identification data from /sys/devices/virtual/dmi/id/ | \n
| \n | \n | sys_vendor | \n string | \nVendor name from /sys/devices/virtual/dmi/id/sys_vendor | \n
| \n | \n | product_name | \n string | \nProduct name from /sys/devices/virtual/dmi/id/product_name | \n
system.name | \n attribute | \n\n | \n | System name information | \n
| \n | \n | nodename | \n string | \nName of the kubernetes node object | \n
usb.device | \n instance | \n\n | \n | USB devices present in the system | \n
| \n | \n | <sysfs-attribute> | \n string | \nValue of the sysfs device attribute, available attributes: class, vendor, device, serial | \n
rule.matched | \n attribute | \n\n | \n | Previously matched rules | \n
| \n | \n | <label-or-var> | \n string | \nLabel or var from a preceding rule that matched | \n
| Flag | \nDescription | \n
|---|---|
| RDTMON | \nIntel RDT Monitoring Technology | \n
| RDTCMT | \nIntel Cache Monitoring (CMT) | \n
| RDTMBM | \nIntel Memory Bandwidth Monitoring (MBM) | \n
| RDTL3CA | \nIntel L3 Cache Allocation Technology | \n
| RDTl2CA | \nIntel L2 Cache Allocation Technology | \n
| RDTMBA | \nIntel Memory Bandwidth Allocation (MBA) Technology | \n
Rules support template-based creation of labels and vars with the\n.labelsTemplate and .varsTemplate fields. These makes it possible to\ndynamically generate labels and vars based on the features that matched.
The template must expand into a simple format with <key>=<value> pairs\nseparated by newline.
Consider the following example:\n
\n\n labelsTemplate: |\n {{ range .pci.device }}vendor-{{ .class }}-{{ .device }}.present=true\n {{ end }}\n matchFeatures:\n - feature: pci.device\n matchExpressions:\n class: {op: InRegexp, value: [\"^02\"]}\n vendor: [\"0fff\"]\nThe rule above will create individual labels\nfeature.node.kubernetes.io/vendor-<class-id>-<device-id>.present=true for\neach network controller device (device class starting with 02) from vendor\n0fff.
All the matched features of each feature matcher term under matchFeatures\nfields are available for the template engine. Matched features can be\nreferenced with {{ .<feature-name> }} in the template, and\nthe available data could be described in yaml as follows:
.\n <key-feature>:\n - Name: <matched-key>\n - ...\n\n <value-feature>:\n - Name: <matched-key>\n Value: <matched-value>\n - ...\n\n <instance-feature>:\n - <attribute-1-name>: <attribute-1-value>\n <attribute-2-name>: <attribute-2-value>\n ...\n - ...\nThat is, the per-feature data is a list of objects whose data fields depend on\nthe type of the feature:
\n\nA simple example of a template utilizing name and value from an attribute\nfeature:\n
\n\n labelsTemplate: |\n {{ range .system.osrelease }}system-{{ .Name }}={{ .Value }}\n {{ end }}\n matchFeatures:\n - feature: system.osRelease\n matchExpressions:\n ID: {op: Exists}\n VERSION_ID.major: {op: Exists}\n\n\n\nNOTE:If both
\nmatchExpressionsandmatchNamefor a feature matcher\nterm (seematchFeatures) is specified, the list of\nmatched features (for the template engine) is the union from both of these.\n\nNOTE: In case of matchAny is specified, the template is executed\nseparately against each individualmatchFeaturesfield and the final set of\nlabels will be superset of all these separate template expansions. E.g.\nconsider the following:
- name: <name>\n labelsTemplate: <template>\n matchFeatures: <matcher#1>\n matchAny:\n - matchFeatures: <matcher#2>\n - matchFeatures: <matcher#3>\nIn the example above (assuming the overall result is a match) the template\nwould be executed on matcher#1 as well as on matcher#2 and/or matcher#3\n(depending on whether both or only one of them match). All the labels from\nthese separate expansions would be created, i.e. the end result would be a\nunion of all the individual expansions.
\n\nRule templates use the Golang text/template\npackage and all its built-in functionality (e.g. pipelines and functions) can\nbe used. An example template taking use of the built-in len function,\nadvertising the number of PCI network controllers from a specific vendor:\n
labelsTemplate: |\n num-intel-network-controllers={{ .pci.device | len }}\n matchFeatures:\n - feature: pci.device\n matchExpressions:\n vendor: {op: In, value: [\"8086\"]}\n class: {op: In, value: [\"0200\"]}\n\nImaginative template pipelines are possible, but care must be taken to\nproduce understandable and maintainable rule sets.
\n\nRules support referencing the output of preceding rules. This enables\nsophisticated scenarios where multiple rules are combined together\nto for more complex heuristics than a single rule can provide. The labels and\nvars created by the execution of preceding rules are available as a special\nrule.matched feature.
Consider the following configuration:
\n\n - name: \"my kernel label rule\"\n labels:\n kernel-feature: \"true\"\n matchFeatures:\n - feature: kernel.version\n matchExpressions:\n major: {op: Gt, value: [\"4\"]}\n\n - name: \"my var rule\"\n vars:\n nolabel-feature: \"true\"\n matchFeatures:\n - feature: cpu.cpuid\n matchExpressions:\n AVX512F: {op: Exists}\n - feature: pci.device\n matchExpressions:\n vendor: {op: In, value: [\"0fff\"]}\n device: {op: In, value: [\"1234\", \"1235\"]}\n\n - name: \"my high level feature rule\"\n labels:\n high-level-feature: \"true\"\n matchFeatures:\n - feature: rule.matched\n matchExpressions:\n kernel-feature: {op: IsTrue}\n nolabel-feature: {op: IsTrue}\nThe feature.node.kubernetes.io/high-level-feature = true label depends on the\ntwo previous rules.
Note that when referencing rules across multiple\nNodeFeatureRule objects attention must be\npaid to the ordering. NodeFeatureRule objects are processed in alphabetical\norder (based on their .metadata.name).
Some more configuration examples below.
\n\nMatch certain CPUID features:
\n\n - name: \"example cpuid rule\"\n labels:\n my-special-cpu-feature: \"true\"\n matchFeatures:\n - feature: cpu.cpuid\n matchExpressions:\n AESNI: {op: Exists}\n AVX: {op: Exists}\nRequire a certain loaded kernel module and OS version:
\n\n - name: \"my multi-feature rule\"\n labels:\n my-special-multi-feature: \"true\"\n matchFeatures:\n - feature: kernel.loadedmodule\n matchExpressions:\n e1000: {op: Exists}\n - feature: system.osrelease\n matchExpressions:\n NAME: {op: InRegexp, values: [\"^openSUSE\"]}\n VERSION_ID.major: {op: Gt, values: [\"14\"]}\nRequire a loaded kernel module and two specific PCI devices (both of which\nmust be present):
\n\n - name: \"my multi-device rule\"\n labels:\n my-multi-device-feature: \"true\"\n matchFeatures:\n - feature: kernel.loadedmodule\n matchExpressions:\n my-driver-module: {op: Exists}\n - pci.device:\n vendor: \"0fff\"\n device: \"1234\"\n - pci.device:\n vendor: \"0fff\"\n device: \"abcd\"\nNode Feature Discovery follows semantic versioning where\nthe version number consists of three components, i.e. MAJOR.MINOR.PATCH.
\n\nThe most recent two minor releases (or release branches) of Node Feature\nDiscovery are supported. That is, with X being the latest release, X and X-1\nare supported and X-1 reaches end-of-life when X+1 is released.
\n\nBuilt-in feature labels and\nfeatures are supported\nfor 2 releases after being deprecated, at minimum. That is, if a feature label\nis deprecated in version X, it will be supported in X+1 and X+2 and\nmay be dropped in X+3.
\n\nCommand-line flags and configuration file options are supported for 1 more\nrelease after being deprecated, at minimum. That is, if option/flag is\ndeprecated in version X, it will be supported in X+1 and may be removed\nin X+2.
\n\nThe same policy (support for 1 release after deprecation) also applies to Helm\nchart parameters.
\n\nNode Feature Discovery is compatible with Kubernetes v1.24 and later.
\n","dir":"/reference/","name":"versions.md","path":"reference/versions.md","url":"/reference/versions.html"},{"title":"Examples and demos","layout":"default","sort":9,"content":"This page contains usage examples and demos.
\n\n
A demo on the benefits of using node feature discovery can be found in the\nsource code repository under\ndemo/.
\n","dir":"/usage/","name":"examples-and-demos.md","path":"usage/examples-and-demos.md","url":"/usage/examples-and-demos.html"},{"title":"Feature Gates","layout":"default","sort":10,"content":"Feature gates are a set of key-value pairs that control the behavior of NFD.\nThey are used to enable or disable certain features of NFD.\nThe feature gates are set using the -feature-gates command line flag or\nfeatureGates value in the Helm chart. The following feature gates are available:
| Name | \nDefault | \nStage | \nSince | \nUntil | \n
|---|---|---|---|---|
NodeFeatureAPI | \n true | \nBeta | \nV0.14 | \nv0.16 | \n
NodeFeatureAPI | \n true | \nGA | \nV0.17 | \n\n |
DisableAutoPrefix | \n false | \nAlpha | \nV0.16 | \n\n |
NodeFeatureGroupAPI | \n false | \nAlpha | \nV0.16 | \n\n |
The NodeFeatureAPI feature gate enables the Node Feature API.\nWhen enabled, NFD will register the Node Feature API with the Kubernetes API\nserver. The Node Feature API is used to expose node-specific hardware and\nsoftware features to the Kubernetes scheduler. The Node Feature API is a beta\nfeature and is enabled by default.
The NodeFeatureGroupAPI feature gate enables the Node Feature Group API.\nWhen enabled, NFD will register the Node Feature Group API with the Kubernetes API\nserver. The Node Feature Group API is used to create node groups based on\nhardware and software features. The Node Feature Group API is an alpha feature\nand is disabled by default.
The DisableAutoPrefix feature gate controls the automatic prefixing of names.\nWhen enabled nfd-master does not automatically add the default\nfeature.node.kubernetes.io/ prefix to unprefixed labels, annotations and\nextended resources. Automatic prefixing is the default behavior in NFD v0.16\nand earlier.
Note that enabling the feature gate effectively causes unprefixed names to be\nfiltered out as NFD does not allow unprefixed names of labels, annotations or\nextended resources. For example, with the DisableAutoPrefix feature gate set\nto false, a NodeFeatureRule with
labels:\n foo: bar\nwill turn into feature.node.kubernetes.io/foo=bar node label. With\nDisableAutoPrefix set to true, no prefix is added and the label will be\nfiltered out.
Note that taint keys are not affected by this feature gate.
\n","dir":"/reference/","name":"feature-gates.md","path":"reference/feature-gates.md","url":"/reference/feature-gates.html"},{"title":"Kubectl plugin","layout":"default","sort":10,"content":"\n\n\nDeveloper Preview This feature is currently in developer preview and\nsubject to change. It is not recommended to use it in production\nenvironments.
\n
The kubectl plugin kubectl nfd can be used to validate/dryrun and test\nNodeFeatureRule objects. It can be installed with the following command:
git clone https://github.com/kubernetes-sigs/node-feature-discovery\ncd node-feature-discovery\nmake build-kubectl-nfd\nKUBECTL_PATH=/usr/local/bin/\nmv ./bin/kubectl-nfd ${KUBECTL_PATH}\nThe plugin can be used to validate a NodeFeatureRule object:
\n\nkubectl nfd validate -f <nodefeaturerule.yaml>\nThe plugin can be used to test a NodeFeatureRule object against a node:
\n\nkubectl nfd test -f <nodefeaturerule.yaml> -n <node-name>\nThe plugin can be used to DryRun a NodeFeatureRule object against a NodeFeature\nfile:
\n\nkubectl get -n node-feature-discovery nodefeature <nodename> -o yaml > <nodefeature.yaml>\nkubectl nfd dryrun -f <nodefeaturerule.yaml> -n <nodefeature.yaml>\nOr you can use the example NodeFeature file(it is a minimal NodeFeature file):
\n\n$ kubectl nfd dryrun -f examples/nodefeaturerule.yaml -n examples/nodefeature.yaml\nEvaluating NodeFeatureRule \"examples/nodefeaturerule.yaml\" against NodeFeature \"examples/nodefeature.yaml\"\nProcessing rule: my sample rule\n*** Labels ***\nvendor.io/my-sample-feature=true\nNodeFeatureRule \"examples/nodefeaturerule.yaml\" is valid for NodeFeature \"examples/nodefeature.yaml\"\nNode Feature Discovery provides a Helm chart to manage its deployment.
NOTE: NFD is not ideal for other Helm charts to depend on as that may result in multiple parallel NFD deployments in the same cluster which is not fully supported by the NFD Helm chart.
Helm package manager should be installed.
To install the latest stable version:
export NFD_NS=node-feature-discovery
+ Helm · Node Feature Discovery
Deployment with Helm
Table of contents
Node Feature Discovery provides a Helm chart to manage its deployment.
NOTE: NFD is not ideal for other Helm charts to depend on as that may result in multiple parallel NFD deployments in the same cluster which is not fully supported by the NFD Helm chart.
Prerequisites
Helm package manager should be installed.
Deployment
To install the latest stable version:
export NFD_NS=node-feature-discovery
helm repo add nfd https://kubernetes-sigs.github.io/node-feature-discovery/charts
helm repo update
helm install nfd/node-feature-discovery --namespace $NFD_NS --create-namespace --generate-name
@@ -41,4 +41,4 @@ kubectl apply -f https://raw.githubusercontent.com/kuber
helm upgrade node-feature-discovery nfd/node-feature-discovery --namespace $NFD_NS
Uninstalling the chart
To uninstall the node-feature-discovery deployment:
export NFD_NS=node-feature-discovery
helm uninstall node-feature-discovery --namespace $NFD_NS
-
The command removes all the Kubernetes components associated with the chart and deletes the release. It also runs a post-delete hook that cleans up the nodes of all labels, annotations, taints and extended resources that were created by NFD.
Chart parameters
To tailor the deployment of the Node Feature Discovery to your needs following Chart parameters are available.
General parameters
Name Type Default Description image.repositorystring gcr.io/k8s-staging-nfd/node-feature-discoveryNFD image repository image.tagstring masterNFD image tag image.pullPolicystring AlwaysImage pull policy imagePullSecretsarray [] ImagePullSecrets is an optional list of references to secrets in the same namespace to use for pulling any of the images used by this PodSpec. More info. nameOverridestring Override the name of the chart fullnameOverridestring Override a default fully qualified app name featureGates.NodeFeatureAPIbool true Enable the NodeFeature CRD API for communicating node features. This will automatically disable the gRPC communication. featureGates.NodeFeatureGroupAPIbool false Enable the NodeFeatureGroup CRD API. featureGates.DisableAutoPrefixbool false Enable DisableAutoPrefix feature gate. Disables automatic prefixing of unprefixed labels, annotations and extended resources. prometheus.enablebool false Specifies whether to expose metrics using prometheus operator prometheus.labelsdict {} Specifies labels for use with the prometheus operator to control how it is selected prometheus.scrapeIntervalstring 10s Specifies the interval by which metrics are scraped priorityClassNamestring The name of the PriorityClass to be used for the NFD pods.
Metrics are configured to be exposed using prometheus operator API's by default. If you want to expose metrics using the prometheus operator API's you need to install the prometheus operator in your cluster.
Master pod parameters
Name Type Default Description master.*dict NFD master deployment configuration master.enablebool true Specifies whether nfd-master should be deployed master.hostNetworkbool false Specifies whether to enable or disable running the container in the host's network namespace master.metricsPortinteger 8081 Port on which to expose metrics from components to prometheus operator. DEPRECATED: will be replaced by master.port in NFD v0.18. master.healthPortinteger 8082 Port on which to expose the grpc health endpoint, will be also used for the probes. DEPRECATED: will be replaced by master.port in NFD v0.18. master.instancestring Instance name. Used to separate annotation namespaces for multiple parallel deployments master.resyncPeriodstring NFD API controller resync period. master.extraLabelNsarray [] List of allowed extra label namespaces master.enableTaintsbool false Specifies whether to enable or disable node tainting master.replicaCountinteger 1 Number of desired pods. This is a pointer to distinguish between explicit zero and not specified master.podSecurityContextdict {} PodSecurityContext holds pod-level security attributes and common container settings master.securityContextdict {} Container security settings master.serviceAccount.createbool true Specifies whether a service account should be created master.serviceAccount.annotationsdict {} Annotations to add to the service account master.serviceAccount.namestring The name of the service account to use. If not set and create is true, a name is generated using the fullname template master.rbac.createbool true Specifies whether to create RBAC configuration for nfd-master master.resources.limitsdict {memory: 4Gi} NFD master pod resources limits master.resources.requestsdict {cpu: 100m, memory: 128Mi} NFD master pod resources requests. See [0] for more info master.tolerationsdict Schedule to control-plane node NFD master pod tolerations master.annotationsdict {} NFD master pod annotations master.affinitydict NFD master pod required node affinity master.deploymentAnnotationsdict {} NFD master deployment annotations master.nfdApiParallelisminteger 10 Specifies the maximum number of concurrent node updates. master.configdict NFD master configuration master.extraArgsarray [] Additional command line arguments to pass to nfd-master master.extraEnvsarray [] Additional environment variables to pass to nfd-master master.revisionHistoryLimitinteger Specify how many old ReplicaSets for this Deployment you want to retain. revisionHistoryLimit master.startupProbe.initialDelaySecond sinteger 0 (by Kubernetes) Specifies the number of seconds after the container has started before startup probes are initiated. master.startupProbe.failureThresholdinteger 30 Specifies the number of consecutive failures of startup probes before considering the pod as not ready. master.startupProbe.periodSecondsinteger 10 (by Kubernetes) Specifies how often (in seconds) to perform the startup probe. master.startupProbe.timeoutSecondsinteger 1 (by Kubernetes) Specifies the number of seconds after which the probe times out. master.livenessProbe.initialDelaySecondsinteger 0 (by Kubernetes) Specifies the number of seconds after the container has started before liveness probes are initiated. master.livenessProbe.failureThresholdinteger 3 (by Kubernetes) Specifies the number of consecutive failures of liveness probes before considering the pod as not ready. master.livenessProbe.periodSecondsinteger 10 (by Kubernetes) Specifies how often (in seconds) to perform the liveness probe. master.livenessProbe.timeoutSecondsinteger 1 (by Kubernetes) Specifies the number of seconds after which the probe times out. master.readinessProbe.initialDelaySecondsinteger 0 (by Kubernetes) Specifies the number of seconds after the container has started before readiness probes are initiated. master.readinessProbe.failureThresholdinteger 10 Specifies the number of consecutive failures of readiness probes before considering the pod as not ready. master.readinessProbe.periodSecondsinteger 10 (by Kubernetes) Specifies how often (in seconds) to perform the readiness probe. master.readinessProbe.timeoutSecondsinteger 1 (by Kubernetes) Specifies the number of seconds after which the probe times out. master.readinessProbe.successThresholdinteger 1 (by Kubernetes) Specifies the number of consecutive successes of readiness probes before considering the pod as ready.
[0] Additional info for master.resources.requests:
You may want to use the same value for requests.memory and limits.memory. The "requests" value affects scheduling to accommodate pods on nodes. If there is a large difference between "requests" and "limits" and nodes experience memory pressure, the kernel may invoke the OOM Killer, even if the memory does not exceed the "limits" threshold. This can cause unexpected pod evictions. Memory cannot be compressed and once allocated to a pod, it can only be reclaimed by killing the pod. Natan Yellin 22/09/2022 that discusses this issue.
Worker pod parameters
Name Type Default Description worker.*dict NFD worker daemonset configuration worker.enablebool true Specifies whether nfd-worker should be deployed worker.hostNetworkbool false Specifies whether to enable or disable running the container in the host's network namespace worker.metricsPortint 8081 Port on which to expose metrics from components to prometheus operator. DEPRECATED: will be replaced by worker.port in NFD v0.18. worker.healthPortint 8082 Port on which to expose the grpc health endpoint, will be also used for the probes. DEPRECATED: will be replaced by worker.port in NFD v0.18. worker.configdict NFD worker configuration worker.podSecurityContextdict {} PodSecurityContext holds pod-level security attributes and common container settins worker.securityContextdict {} Container security settings worker.serviceAccount.createbool true Specifies whether a service account for nfd-worker should be created worker.serviceAccount.annotationsdict {} Annotations to add to the service account for nfd-worker worker.serviceAccount.namestring The name of the service account to use for nfd-worker. If not set and create is true, a name is generated using the fullname template (suffixed with -worker) worker.rbac.createbool true Specifies whether to create RBAC configuration for nfd-worker worker.mountUsrSrcbool false Specifies whether to allow users to mount the hostpath /user/src. Does not work on systems without /usr/src AND a read-only /usr worker.resources.limitsdict {memory: 512Mi} NFD worker pod resources limits worker.resources.requestsdict {cpu: 5m, memory: 64Mi} NFD worker pod resources requests worker.nodeSelectordict {} NFD worker pod node selector worker.tolerationsdict {} NFD worker pod node tolerations worker.priorityClassNamestring NFD worker pod priority class worker.annotationsdict {} NFD worker pod annotations worker.daemonsetAnnotationsdict {} NFD worker daemonset annotations worker.extraArgsarray [] Additional command line arguments to pass to nfd-worker worker.extraEnvsarray [] Additional environment variables to pass to nfd-worker worker.revisionHistoryLimitinteger Specify how many old ControllerRevisions for this DaemonSet you want to retain. revisionHistoryLimit worker.livenessProbe.initialDelaySecondsinteger 10 Specifies the number of seconds after the container has started before liveness probes are initiated. worker.livenessProbe.failureThresholdinteger 3 (by Kubernetes) Specifies the number of consecutive failures of liveness probes before considering the pod as not ready. worker.livenessProbe.periodSecondsinteger 10 (by Kubernetes) Specifies how often (in seconds) to perform the liveness probe. worker.livenessProbe.timeoutSecondsinteger 1 (by Kubernetes) Specifies the number of seconds after which the probe times out. worker.readinessProbe.initialDelaySecondsinteger 5 Specifies the number of seconds after the container has started before readiness probes are initiated. worker.readinessProbe.failureThresholdinteger 10 Specifies the number of consecutive failures of readiness probes before considering the pod as not ready. worker.readinessProbe.periodSecondsinteger 10 (by Kubernetes) Specifies how often (in seconds) to perform the readiness probe. worker.readinessProbe.timeoutSecondsinteger 1 (by Kubernetes) Specifies the number of seconds after which the probe times out. worker.readinessProbe.successThresholdinteger 1 (by Kubernetes) Specifies the number of consecutive successes of readiness probes before considering the pod as ready.
Topology updater parameters
Name Type Default Description topologyUpdater.*dict NFD Topology Updater configuration topologyUpdater.enablebool false Specifies whether the NFD Topology Updater should be created topologyUpdater.hostNetworkbool false Specifies whether to enable or disable running the container in the host's network namespace topologyUpdater.createCRDsbool false Specifies whether the NFD Topology Updater CRDs should be created topologyUpdater.serviceAccount.createbool true Specifies whether the service account for topology updater should be created topologyUpdater.serviceAccount.annotationsdict {} Annotations to add to the service account for topology updater topologyUpdater.serviceAccount.namestring The name of the service account for topology updater to use. If not set and create is true, a name is generated using the fullname template and -topology-updater suffix topologyUpdater.rbac.createbool true Specifies whether to create RBAC configuration for topology updater topologyUpdater.metricsPortinteger 8081 Port on which to expose prometheus metrics. DEPRECATED: will be replaced by topologyUpdater.port in NFD v0.18. topologyUpdater.healthPortinteger 8082 Port on which to expose the grpc health endpoint, will be also used for the probes. DEPRECATED: will be replaced by topologyUpdater.port in NFD v0.18. topologyUpdater.kubeletConfigPathstring "" Specifies the kubelet config host path topologyUpdater.kubeletPodResourcesSockPathstring "" Specifies the kubelet sock path to read pod resources topologyUpdater.updateIntervalstring 60s Time to sleep between CR updates. Non-positive value implies no CR update. topologyUpdater.watchNamespacestring *Namespace to watch pods, * for all namespaces topologyUpdater.podSecurityContextdict {} PodSecurityContext holds pod-level security attributes and common container sett topologyUpdater.securityContextdict {} Container security settings topologyUpdater.resources.limitsdict {memory: 60Mi} NFD Topology Updater pod resources limits topologyUpdater.resources.requestsdict {cpu: 50m, memory: 40Mi} NFD Topology Updater pod resources requests topologyUpdater.nodeSelectordict {} Topology updater pod node selector topologyUpdater.tolerationsdict {} Topology updater pod node tolerations topologyUpdater.annotationsdict {} Topology updater pod annotations topologyUpdater.daemonsetAnnotationsdict {} Topology updater daemonset annotations topologyUpdater.affinitydict {} Topology updater pod affinity topologyUpdater.configdict configuration topologyUpdater.podSetFingerprintbool true Enables compute and report of pod fingerprint in NRT objects. topologyUpdater.kubeletStateDirstring /var/lib/kubelet Specifies kubelet state directory path for watching state and checkpoint files. Empty value disables kubelet state tracking. topologyUpdater.extraArgsarray [] Additional command line arguments to pass to nfd-topology-updater topologyUpdater.extraEnvsarray [] Additional environment variables to pass to nfd-topology-updater topologyUpdater.revisionHistoryLimitinteger Specify how many old ControllerRevisions for this DaemonSet you want to retain. revisionHistoryLimit topologyUpdater.livenessProbe.initialDelaySecondsinteger 10 Specifies the number of seconds after the container has started before liveness probes are initiated. topologyUpdater.livenessProbe.failureThresholdinteger 3 (by Kubernetes) Specifies the number of consecutive failures of liveness probes before considering the pod as not ready. topologyUpdater.livenessProbe.periodSecondsinteger 10 (by Kubernetes) Specifies how often (in seconds) to perform the liveness probe. topologyUpdater.livenessProbe.timeoutSecondsinteger 1 (by Kubernetes) Specifies the number of seconds after which the probe times out. topologyUpdater.readinessProbe.initialDelaySecondsinteger 5 Specifies the number of seconds after the container has started before readiness probes are initiated. topologyUpdater.readinessProbe.failureThresholdinteger 10 Specifies the number of consecutive failures of readiness probes before considering the pod as not ready. topologyUpdater.readinessProbe.periodSecondsinteger 10 (by Kubernetes) Specifies how often (in seconds) to perform the readiness probe. topologyUpdater.readinessProbe.timeoutSecondsinteger 1 (by Kubernetes) Specifies the number of seconds after which the probe times out. topologyUpdater.readinessProbe.successThresholdinteger 1 (by Kubernetes) Specifies the number of consecutive successes of readiness probes before considering the pod as ready.
Garbage collector parameters
Name Type Default Description gc.*dict NFD Garbage Collector configuration gc.enablebool true Specifies whether the NFD Garbage Collector should be created gc.hostNetworkbool false Specifies whether to enable or disable running the container in the host's network namespace gc.serviceAccount.createbool true Specifies whether the service account for garbage collector should be created gc.serviceAccount.annotationsdict {} Annotations to add to the service account for garbage collector gc.serviceAccount.namestring The name of the service account for garbage collector to use. If not set and create is true, a name is generated using the fullname template and -gc suffix gc.rbac.createbool true Specifies whether to create RBAC configuration for garbage collector gc.intervalstring 1h Time between periodic garbage collector runs gc.podSecurityContextdict {} PodSecurityContext holds pod-level security attributes and common container settings gc.resources.limitsdict {memory: 1Gi} NFD Garbage Collector pod resources limits gc.resources.requestsdict {cpu: 10m, memory: 128Mi} NFD Garbage Collector pod resources requests gc.metricsPortinteger 8081 Port on which to serve Prometheus metrics. DEPRECATED: will be replaced by gc.port in NFD v0.18. gc.nodeSelectordict {} Garbage collector pod node selector gc.tolerationsdict {} Garbage collector pod node tolerations gc.annotationsdict {} Garbage collector pod annotations gc.deploymentAnnotationsdict {} Garbage collector deployment annotations gc.affinitydict {} Garbage collector pod affinity gc.extraArgsarray [] Additional command line arguments to pass to nfd-gc gc.extraEnvsarray [] Additional environment variables to pass to nfd-gc gc.revisionHistoryLimitinteger Specify how many old ReplicaSets for this Deployment you want to retain. revisionHistoryLimit
Node Feature Discovery master
\ No newline at end of file
+The command removes all the Kubernetes components associated with the chart and deletes the release. It also runs a post-delete hook that cleans up the nodes of all labels, annotations, taints and extended resources that were created by NFD.
To tailor the deployment of the Node Feature Discovery to your needs following Chart parameters are available.
| Name | Type | Default | Description |
|---|---|---|---|
image.repository | string | gcr.io/k8s-staging-nfd/node-feature-discovery | NFD image repository |
image.tag | string | master | NFD image tag |
image.pullPolicy | string | Always | Image pull policy |
imagePullSecrets | array | [] | ImagePullSecrets is an optional list of references to secrets in the same namespace to use for pulling any of the images used by this PodSpec. More info. |
nameOverride | string | Override the name of the chart | |
fullnameOverride | string | Override a default fully qualified app name | |
featureGates.NodeFeatureGroupAPI | bool | false | Enable the NodeFeatureGroup CRD API. |
featureGates.DisableAutoPrefix | bool | false | Enable DisableAutoPrefix feature gate. Disables automatic prefixing of unprefixed labels, annotations and extended resources. |
prometheus.enable | bool | false | Specifies whether to expose metrics using prometheus operator |
prometheus.labels | dict | {} | Specifies labels for use with the prometheus operator to control how it is selected |
prometheus.scrapeInterval | string | 10s | Specifies the interval by which metrics are scraped |
priorityClassName | string | The name of the PriorityClass to be used for the NFD pods. |
Metrics are configured to be exposed using prometheus operator API's by default. If you want to expose metrics using the prometheus operator API's you need to install the prometheus operator in your cluster.
| Name | Type | Default | Description |
|---|---|---|---|
master.* | dict | NFD master deployment configuration | |
master.enable | bool | true | Specifies whether nfd-master should be deployed |
master.hostNetwork | bool | false | Specifies whether to enable or disable running the container in the host's network namespace |
master.metricsPort | integer | 8081 | Port on which to expose metrics from components to prometheus operator. DEPRECATED: will be replaced by master.port in NFD v0.18. |
master.healthPort | integer | 8082 | Port on which to expose the grpc health endpoint, will be also used for the probes. DEPRECATED: will be replaced by master.port in NFD v0.18. |
master.instance | string | Instance name. Used to separate annotation namespaces for multiple parallel deployments | |
master.resyncPeriod | string | NFD API controller resync period. | |
master.extraLabelNs | array | [] | List of allowed extra label namespaces |
master.enableTaints | bool | false | Specifies whether to enable or disable node tainting |
master.replicaCount | integer | 1 | Number of desired pods. This is a pointer to distinguish between explicit zero and not specified |
master.podSecurityContext | dict | {} | PodSecurityContext holds pod-level security attributes and common container settings |
master.securityContext | dict | {} | Container security settings |
master.serviceAccount.create | bool | true | Specifies whether a service account should be created |
master.serviceAccount.annotations | dict | {} | Annotations to add to the service account |
master.serviceAccount.name | string | The name of the service account to use. If not set and create is true, a name is generated using the fullname template | |
master.rbac.create | bool | true | Specifies whether to create RBAC configuration for nfd-master |
master.resources.limits | dict | {memory: 4Gi} | NFD master pod resources limits |
master.resources.requests | dict | {cpu: 100m, memory: 128Mi} | NFD master pod resources requests. See [0] for more info |
master.tolerations | dict | Schedule to control-plane node | NFD master pod tolerations |
master.annotations | dict | {} | NFD master pod annotations |
master.affinity | dict | NFD master pod required node affinity | |
master.deploymentAnnotations | dict | {} | NFD master deployment annotations |
master.nfdApiParallelism | integer | 10 | Specifies the maximum number of concurrent node updates. |
master.config | dict | NFD master configuration | |
master.extraArgs | array | [] | Additional command line arguments to pass to nfd-master |
master.extraEnvs | array | [] | Additional environment variables to pass to nfd-master |
master.revisionHistoryLimit | integer | Specify how many old ReplicaSets for this Deployment you want to retain. revisionHistoryLimit | |
master.startupProbe.initialDelaySecond s | integer | 0 (by Kubernetes) | Specifies the number of seconds after the container has started before startup probes are initiated. |
master.startupProbe.failureThreshold | integer | 30 | Specifies the number of consecutive failures of startup probes before considering the pod as not ready. |
master.startupProbe.periodSeconds | integer | 10 (by Kubernetes) | Specifies how often (in seconds) to perform the startup probe. |
master.startupProbe.timeoutSeconds | integer | 1 (by Kubernetes) | Specifies the number of seconds after which the probe times out. |
master.livenessProbe.initialDelaySeconds | integer | 0 (by Kubernetes) | Specifies the number of seconds after the container has started before liveness probes are initiated. |
master.livenessProbe.failureThreshold | integer | 3 (by Kubernetes) | Specifies the number of consecutive failures of liveness probes before considering the pod as not ready. |
master.livenessProbe.periodSeconds | integer | 10 (by Kubernetes) | Specifies how often (in seconds) to perform the liveness probe. |
master.livenessProbe.timeoutSeconds | integer | 1 (by Kubernetes) | Specifies the number of seconds after which the probe times out. |
master.readinessProbe.initialDelaySeconds | integer | 0 (by Kubernetes) | Specifies the number of seconds after the container has started before readiness probes are initiated. |
master.readinessProbe.failureThreshold | integer | 10 | Specifies the number of consecutive failures of readiness probes before considering the pod as not ready. |
master.readinessProbe.periodSeconds | integer | 10 (by Kubernetes) | Specifies how often (in seconds) to perform the readiness probe. |
master.readinessProbe.timeoutSeconds | integer | 1 (by Kubernetes) | Specifies the number of seconds after which the probe times out. |
master.readinessProbe.successThreshold | integer | 1 (by Kubernetes) | Specifies the number of consecutive successes of readiness probes before considering the pod as ready. |
[0]Additional info formaster.resources.requests:
You may want to use the same value forrequests.memoryandlimits.memory. The "requests" value affects scheduling to accommodate pods on nodes. If there is a large difference between "requests" and "limits" and nodes experience memory pressure, the kernel may invoke the OOM Killer, even if the memory does not exceed the "limits" threshold. This can cause unexpected pod evictions. Memory cannot be compressed and once allocated to a pod, it can only be reclaimed by killing the pod. Natan Yellin 22/09/2022 that discusses this issue.
| Name | Type | Default | Description |
|---|---|---|---|
worker.* | dict | NFD worker daemonset configuration | |
worker.enable | bool | true | Specifies whether nfd-worker should be deployed |
worker.hostNetwork | bool | false | Specifies whether to enable or disable running the container in the host's network namespace |
worker.metricsPort | int | 8081 | Port on which to expose metrics from components to prometheus operator. DEPRECATED: will be replaced by worker.port in NFD v0.18. |
worker.healthPort | int | 8082 | Port on which to expose the grpc health endpoint, will be also used for the probes. DEPRECATED: will be replaced by worker.port in NFD v0.18. |
worker.config | dict | NFD worker configuration | |
worker.podSecurityContext | dict | {} | PodSecurityContext holds pod-level security attributes and common container settins |
worker.securityContext | dict | {} | Container security settings |
worker.serviceAccount.create | bool | true | Specifies whether a service account for nfd-worker should be created |
worker.serviceAccount.annotations | dict | {} | Annotations to add to the service account for nfd-worker |
worker.serviceAccount.name | string | The name of the service account to use for nfd-worker. If not set and create is true, a name is generated using the fullname template (suffixed with -worker) | |
worker.rbac.create | bool | true | Specifies whether to create RBAC configuration for nfd-worker |
worker.mountUsrSrc | bool | false | Specifies whether to allow users to mount the hostpath /user/src. Does not work on systems without /usr/src AND a read-only /usr |
worker.resources.limits | dict | {memory: 512Mi} | NFD worker pod resources limits |
worker.resources.requests | dict | {cpu: 5m, memory: 64Mi} | NFD worker pod resources requests |
worker.nodeSelector | dict | {} | NFD worker pod node selector |
worker.tolerations | dict | {} | NFD worker pod node tolerations |
worker.priorityClassName | string | NFD worker pod priority class | |
worker.annotations | dict | {} | NFD worker pod annotations |
worker.daemonsetAnnotations | dict | {} | NFD worker daemonset annotations |
worker.extraArgs | array | [] | Additional command line arguments to pass to nfd-worker |
worker.extraEnvs | array | [] | Additional environment variables to pass to nfd-worker |
worker.revisionHistoryLimit | integer | Specify how many old ControllerRevisions for this DaemonSet you want to retain. revisionHistoryLimit | |
worker.livenessProbe.initialDelaySeconds | integer | 10 | Specifies the number of seconds after the container has started before liveness probes are initiated. |
worker.livenessProbe.failureThreshold | integer | 3 (by Kubernetes) | Specifies the number of consecutive failures of liveness probes before considering the pod as not ready. |
worker.livenessProbe.periodSeconds | integer | 10 (by Kubernetes) | Specifies how often (in seconds) to perform the liveness probe. |
worker.livenessProbe.timeoutSeconds | integer | 1 (by Kubernetes) | Specifies the number of seconds after which the probe times out. |
worker.readinessProbe.initialDelaySeconds | integer | 5 | Specifies the number of seconds after the container has started before readiness probes are initiated. |
worker.readinessProbe.failureThreshold | integer | 10 | Specifies the number of consecutive failures of readiness probes before considering the pod as not ready. |
worker.readinessProbe.periodSeconds | integer | 10 (by Kubernetes) | Specifies how often (in seconds) to perform the readiness probe. |
worker.readinessProbe.timeoutSeconds | integer | 1 (by Kubernetes) | Specifies the number of seconds after which the probe times out. |
worker.readinessProbe.successThreshold | integer | 1 (by Kubernetes) | Specifies the number of consecutive successes of readiness probes before considering the pod as ready. |
| Name | Type | Default | Description |
|---|---|---|---|
topologyUpdater.* | dict | NFD Topology Updater configuration | |
topologyUpdater.enable | bool | false | Specifies whether the NFD Topology Updater should be created |
topologyUpdater.hostNetwork | bool | false | Specifies whether to enable or disable running the container in the host's network namespace |
topologyUpdater.createCRDs | bool | false | Specifies whether the NFD Topology Updater CRDs should be created |
topologyUpdater.serviceAccount.create | bool | true | Specifies whether the service account for topology updater should be created |
topologyUpdater.serviceAccount.annotations | dict | {} | Annotations to add to the service account for topology updater |
topologyUpdater.serviceAccount.name | string | The name of the service account for topology updater to use. If not set and create is true, a name is generated using the fullname template and -topology-updater suffix | |
topologyUpdater.rbac.create | bool | true | Specifies whether to create RBAC configuration for topology updater |
topologyUpdater.metricsPort | integer | 8081 | Port on which to expose prometheus metrics. DEPRECATED: will be replaced by topologyUpdater.port in NFD v0.18. |
topologyUpdater.healthPort | integer | 8082 | Port on which to expose the grpc health endpoint, will be also used for the probes. DEPRECATED: will be replaced by topologyUpdater.port in NFD v0.18. |
topologyUpdater.kubeletConfigPath | string | "" | Specifies the kubelet config host path |
topologyUpdater.kubeletPodResourcesSockPath | string | "" | Specifies the kubelet sock path to read pod resources |
topologyUpdater.updateInterval | string | 60s | Time to sleep between CR updates. Non-positive value implies no CR update. |
topologyUpdater.watchNamespace | string | * | Namespace to watch pods, * for all namespaces |
topologyUpdater.podSecurityContext | dict | {} | PodSecurityContext holds pod-level security attributes and common container sett |
topologyUpdater.securityContext | dict | {} | Container security settings |
topologyUpdater.resources.limits | dict | {memory: 60Mi} | NFD Topology Updater pod resources limits |
topologyUpdater.resources.requests | dict | {cpu: 50m, memory: 40Mi} | NFD Topology Updater pod resources requests |
topologyUpdater.nodeSelector | dict | {} | Topology updater pod node selector |
topologyUpdater.tolerations | dict | {} | Topology updater pod node tolerations |
topologyUpdater.annotations | dict | {} | Topology updater pod annotations |
topologyUpdater.daemonsetAnnotations | dict | {} | Topology updater daemonset annotations |
topologyUpdater.affinity | dict | {} | Topology updater pod affinity |
topologyUpdater.config | dict | configuration | |
topologyUpdater.podSetFingerprint | bool | true | Enables compute and report of pod fingerprint in NRT objects. |
topologyUpdater.kubeletStateDir | string | /var/lib/kubelet | Specifies kubelet state directory path for watching state and checkpoint files. Empty value disables kubelet state tracking. |
topologyUpdater.extraArgs | array | [] | Additional command line arguments to pass to nfd-topology-updater |
topologyUpdater.extraEnvs | array | [] | Additional environment variables to pass to nfd-topology-updater |
topologyUpdater.revisionHistoryLimit | integer | Specify how many old ControllerRevisions for this DaemonSet you want to retain. revisionHistoryLimit | |
topologyUpdater.livenessProbe.initialDelaySeconds | integer | 10 | Specifies the number of seconds after the container has started before liveness probes are initiated. |
topologyUpdater.livenessProbe.failureThreshold | integer | 3 (by Kubernetes) | Specifies the number of consecutive failures of liveness probes before considering the pod as not ready. |
topologyUpdater.livenessProbe.periodSeconds | integer | 10 (by Kubernetes) | Specifies how often (in seconds) to perform the liveness probe. |
topologyUpdater.livenessProbe.timeoutSeconds | integer | 1 (by Kubernetes) | Specifies the number of seconds after which the probe times out. |
topologyUpdater.readinessProbe.initialDelaySeconds | integer | 5 | Specifies the number of seconds after the container has started before readiness probes are initiated. |
topologyUpdater.readinessProbe.failureThreshold | integer | 10 | Specifies the number of consecutive failures of readiness probes before considering the pod as not ready. |
topologyUpdater.readinessProbe.periodSeconds | integer | 10 (by Kubernetes) | Specifies how often (in seconds) to perform the readiness probe. |
topologyUpdater.readinessProbe.timeoutSeconds | integer | 1 (by Kubernetes) | Specifies the number of seconds after which the probe times out. |
topologyUpdater.readinessProbe.successThreshold | integer | 1 (by Kubernetes) | Specifies the number of consecutive successes of readiness probes before considering the pod as ready. |
| Name | Type | Default | Description |
|---|---|---|---|
gc.* | dict | NFD Garbage Collector configuration | |
gc.enable | bool | true | Specifies whether the NFD Garbage Collector should be created |
gc.hostNetwork | bool | false | Specifies whether to enable or disable running the container in the host's network namespace |
gc.serviceAccount.create | bool | true | Specifies whether the service account for garbage collector should be created |
gc.serviceAccount.annotations | dict | {} | Annotations to add to the service account for garbage collector |
gc.serviceAccount.name | string | The name of the service account for garbage collector to use. If not set and create is true, a name is generated using the fullname template and -gc suffix | |
gc.rbac.create | bool | true | Specifies whether to create RBAC configuration for garbage collector |
gc.interval | string | 1h | Time between periodic garbage collector runs |
gc.podSecurityContext | dict | {} | PodSecurityContext holds pod-level security attributes and common container settings |
gc.resources.limits | dict | {memory: 1Gi} | NFD Garbage Collector pod resources limits |
gc.resources.requests | dict | {cpu: 10m, memory: 128Mi} | NFD Garbage Collector pod resources requests |
gc.metricsPort | integer | 8081 | Port on which to serve Prometheus metrics. DEPRECATED: will be replaced by gc.port in NFD v0.18. |
gc.nodeSelector | dict | {} | Garbage collector pod node selector |
gc.tolerations | dict | {} | Garbage collector pod node tolerations |
gc.annotations | dict | {} | Garbage collector pod annotations |
gc.deploymentAnnotations | dict | {} | Garbage collector deployment annotations |
gc.affinity | dict | {} | Garbage collector pod affinity |
gc.extraArgs | array | [] | Additional command line arguments to pass to nfd-gc |
gc.extraEnvs | array | [] | Additional environment variables to pass to nfd-gc |
gc.revisionHistoryLimit | integer | Specify how many old ReplicaSets for this Deployment you want to retain. revisionHistoryLimit |
NFD offers two variants of the container image. Released container images are available for x86_64 and Arm64 architectures.
The default is a minimal image based on scratch and only supports running statically linked binaries.
For backwards compatibility a container image tag with suffix -minimal (e.g. gcr.io/k8s-staging-nfd/node-feature-discovery:master-minimal) is provided.
This image is based on debian:bookworm-slim and contains a full Linux system for doing live debugging and diagnosis of the NFD images.
The container image tag has suffix -full (e.g. gcr.io/k8s-staging-nfd/node-feature-discovery:master-full).
NFD offers two variants of the container image. Released container images are available for x86_64 and Arm64 architectures.
The default is a minimal image based on scratch and only supports running statically linked binaries.
For backwards compatibility a container image tag with suffix -minimal (e.g. gcr.io/k8s-staging-nfd/node-feature-discovery:master-minimal) is provided.
This image is based on debian:bookworm-slim and contains a full Linux system for doing live debugging and diagnosis of the NFD images.
The container image tag has suffix -full (e.g. gcr.io/k8s-staging-nfd/node-feature-discovery:master-full).
Node Feature Discovery can be deployed on any recent version of Kubernetes (v1.24+).
See Image variants for description of the different NFD container images available.
Using Kustomize provides straightforward deployment with kubectl integration and declarative customization.
Using Helm provides easy management of NFD deployments with nice configuration management and easy upgrades.
Using Operator provides deployment and configuration management via CRDs.
Node Feature Discovery can be deployed on any recent version of Kubernetes (v1.24+).
See Image variants for description of the different NFD container images available.
Using Kustomize provides straightforward deployment with kubectl integration and declarative customization.
Using Helm provides easy management of NFD deployments with nice configuration management and easy upgrades.
Using Operator provides deployment and configuration management via CRDs.
Kustomize can be used to deploy NFD. Customization of the deployment is done by maintaining declarative overlays on top of the base overlays in NFD.
To follow the deployment instructions here, kubectl v1.24 or later is required.
The kustomize overlays provided in the repo can be used directly:
kubectl apply -k "https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default?ref=master"
+ Kustomize · Node Feature Discovery
Deployment with Kustomize
Table of contents
Kustomize can be used to deploy NFD. Customization of the deployment is done by maintaining declarative overlays on top of the base overlays in NFD.
To follow the deployment instructions here, kubectl v1.24 or later is required.
The kustomize overlays provided in the repo can be used directly:
kubectl apply -k "https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default?ref=master"
This will required RBAC rules and deploy nfd-master (as a deployment) and nfd-worker (as daemonset) in the node-feature-discovery namespace.
NOTE: nfd-topology-updater is not deployed as part of the default overlay. Refer to the Master Worker Topologyupdater and Topologyupdater below.
Alternatively you can clone the repository and customize the deployment by creating your own overlays. See kustomize for more information about managing deployment configurations.
Overlays
The NFD repository hosts a set of overlays for different usages and deployment scenarios under deployment/overlays
default: default deployment of nfd-worker as a daemonset, described abovedefault-job: see Worker one-shot belowmaster-worker-topologyupdater: see Master Worker Topologyupdater belowtopologyupdater: see Topology Updater belowprometheus: see Metrics belowprune: clean up the cluster after uninstallation, see Removing feature labelssamples/custom-rules: an example for spicing up the default deployment with a separately managed configmap of custom labeling rules, see Custom feature source for more information about custom node labels
Worker one-shot
Feature discovery can alternatively be configured as a one-shot job. The default-job overlay may be used to achieve this:
NUM_NODES=$(kubectl get no -o jsonpath='{.items[*].metadata.name}' | wc -w)
kubectl kustomize "https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default-job?ref=master" | \
sed s"/NUM_NODES/$NUM_NODES/" | \
@@ -21,4 +21,4 @@ kubectl -n $NFD_NS delete svc nf
kubectl -n $NFD_NS delete sa nfd-master
kubectl delete clusterrole nfd-master
kubectl delete clusterrolebinding nfd-master
-
Node Feature Discovery master
\ No newline at end of file
+Metrics are configured to be exposed using prometheus operator API's by default. If you want to expose metrics using the prometheus operator API's you need to install the prometheus operator in your cluster. By default NFD Master and Worker expose metrics on port 8081.
The exposed metrics are
| Metric | Type | Description |
|---|---|---|
nfd_master_build_info | Gauge | Version from which nfd-master was built |
nfd_worker_build_info | Gauge | Version from which nfd-worker was built |
nfd_gc_build_info | Gauge | Version from which nfd-gc was built |
nfd_topology_updater_build_info | Gauge | Version from which nfd-topology-updater was built |
nfd_master_node_update_requests_total | Counter | Number of node update requests received by the master over gRPC |
nfd_master_node_updates_total | Counter | Number of nodes updated |
nfd_master_node_feature_group_update_requests_total | Counter | Number of cluster feature update requests processed by the master |
nfd_master_node_update_failures_total | Counter | Number of nodes update failures |
nfd_master_node_labels_rejected_total | Counter | Number of nodes labels rejected by nfd-master |
nfd_master_node_extendedresources_rejected_total | Counter | Number of nodes extended resources rejected by nfd-master |
nfd_master_node_taints_rejected_total | Counter | Number of nodes taints rejected by nfd-master |
nfd_master_nodefeaturerule_processing_duration_seconds | Histogram | Time taken to process NodeFeatureRule objects |
nfd_master_nodefeaturerule_processing_errors_total | Counter | Number or errors encountered while processing NodeFeatureRule objects |
nfd_worker_feature_discovery_duration_seconds | Histogram | Time taken to discover features on a node |
nfd_topology_updater_scan_errors_total | Counter | Number of errors in scanning resource allocation of pods. |
nfd_gc_objects_deleted_total | Counter | Number of NodeFeature and NodeResourceTopology objects garbage collected. |
nfd_gc_object_delete_failures_total | Counter | Number of errors in deleting NodeFeature and NodeResourceTopology objects. |
To deploy NFD with metrics enabled using kustomize, you can use the prometheus overlay.
By default metrics are enabled when deploying NFD via Helm. To enable Prometheus to scrape metrics from NFD, you need to pass the following values to Helm:
--set prometheus.enable=true
-For more info on Helm deployment, see Helm.
It is recommended to specify --set prometheus.prometheusSpec.podMonitorSelectorNilUsesHelmValues=false when deploying prometheus-operator via Helm to enable the prometheus-operator to scrape metrics from any PodMonitor.
or setting labels on the PodMonitor via the helm parameter prometheus.labels to control which Prometheus instances will scrape this PodMonitor.
NFD contains an example Grafana dashboard. You can import examples/grafana-dashboard.json to your Grafana instance to visualize the NFD metrics.
Metrics are configured to be exposed using prometheus operator API's by default. If you want to expose metrics using the prometheus operator API's you need to install the prometheus operator in your cluster. By default NFD Master and Worker expose metrics on port 8081.
The exposed metrics are
| Metric | Type | Description |
|---|---|---|
nfd_master_build_info | Gauge | Version from which nfd-master was built |
nfd_worker_build_info | Gauge | Version from which nfd-worker was built |
nfd_gc_build_info | Gauge | Version from which nfd-gc was built |
nfd_topology_updater_build_info | Gauge | Version from which nfd-topology-updater was built |
nfd_master_node_update_requests_total | Counter | Number of node update requests received by the master over gRPC |
nfd_master_node_updates_total | Counter | Number of nodes updated |
nfd_master_node_feature_group_update_requests_total | Counter | Number of cluster feature update requests processed by the master |
nfd_master_node_update_failures_total | Counter | Number of nodes update failures |
nfd_master_node_labels_rejected_total | Counter | Number of nodes labels rejected by nfd-master |
nfd_master_node_extendedresources_rejected_total | Counter | Number of nodes extended resources rejected by nfd-master |
nfd_master_node_taints_rejected_total | Counter | Number of nodes taints rejected by nfd-master |
nfd_master_nodefeaturerule_processing_duration_seconds | Histogram | Time taken to process NodeFeatureRule objects |
nfd_master_nodefeaturerule_processing_errors_total | Counter | Number or errors encountered while processing NodeFeatureRule objects |
nfd_worker_feature_discovery_duration_seconds | Histogram | Time taken to discover features on a node |
nfd_topology_updater_scan_errors_total | Counter | Number of errors in scanning resource allocation of pods. |
nfd_gc_objects_deleted_total | Counter | Number of NodeFeature and NodeResourceTopology objects garbage collected. |
nfd_gc_object_delete_failures_total | Counter | Number of errors in deleting NodeFeature and NodeResourceTopology objects. |
To deploy NFD with metrics enabled using kustomize, you can use the prometheus overlay.
By default metrics are enabled when deploying NFD via Helm. To enable Prometheus to scrape metrics from NFD, you need to pass the following values to Helm:
--set prometheus.enable=true
+For more info on Helm deployment, see Helm.
It is recommended to specify --set prometheus.prometheusSpec.podMonitorSelectorNilUsesHelmValues=false when deploying prometheus-operator via Helm to enable the prometheus-operator to scrape metrics from any PodMonitor.
or setting labels on the PodMonitor via the helm parameter prometheus.labels to control which Prometheus instances will scrape this PodMonitor.
NFD contains an example Grafana dashboard. You can import examples/grafana-dashboard.json to your Grafana instance to visualize the NFD metrics.
The Node Feature Discovery Operator automates installation, configuration and updates of NFD using a specific NodeFeatureDiscovery custom resource. This also provides good support for managing NFD as a dependency of other operators.
Deployment using the Node Feature Discovery Operator is recommended to be done via operatorhub.io.
Install the operator:
kubectl create -f https://operatorhub.io/install/nfd-operator.yaml
+ NFD Operator · Node Feature Discovery
Deployment with NFD Operator
Table of contents
The Node Feature Discovery Operator automates installation, configuration and updates of NFD using a specific NodeFeatureDiscovery custom resource. This also provides good support for managing NFD as a dependency of other operators.
Deployment
Deployment using the Node Feature Discovery Operator is recommended to be done via operatorhub.io.
- You need to have OLM installed. If you don't, take a look at the latest release for detailed instructions.
-
Install the operator:
kubectl create -f https://operatorhub.io/install/nfd-operator.yaml
-
Create NodeFeatureDiscovery object (in nfd namespace here):
cat << EOF | kubectl apply -f -
apiVersion: v1
kind: Namespace
@@ -17,4 +17,4 @@ spec:
EOF
Uninstallation
If you followed the deployment instructions above you can uninstall NFD with:
kubectl -n nfd delete NodeFeatureDiscovery my-nfd-deployment
Optionally, you can also remove the namespace:
kubectl delete ns nfd
-
See the node-feature-discovery-operator and OLM project documentation for instructions for uninstalling the operator and operator lifecycle manager, respectively.
Node Feature Discovery master
\ No newline at end of file
+See the node-feature-discovery-operator and OLM project documentation for instructions for uninstalling the operator and operator lifecycle manager, respectively.
Follow the uninstallation instructions of the deployment method used (kustomize, helm or operator).
NOTE: This is unnecessary when using the Helm chart for deployment as it will clean up the nodes when NFD is uninstalled.
NFD-Master has a special -prune command line flag for removing all nfd-related node labels, annotations, extended resources and taints from the cluster.
kubectl apply -k "https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/prune?ref=master"
+ Uninstallation · Node Feature Discovery
Uninstallation
Follow the uninstallation instructions of the deployment method used (kustomize, helm or operator).
Removing feature labels
NOTE: This is unnecessary when using the Helm chart for deployment as it will clean up the nodes when NFD is uninstalled.
NFD-Master has a special -prune command line flag for removing all nfd-related node labels, annotations, extended resources and taints from the cluster.
kubectl apply -k "https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/prune?ref=master"
kubectl -n node-feature-discovery wait job.batch/nfd-master --for=condition=complete && \
kubectl delete -k "https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/prune?ref=master"
-
NOTE: You must run prune before removing the RBAC rules (serviceaccount, clusterrole and clusterrolebinding).
Node Feature Discovery master
\ No newline at end of file
+NOTE: You must run prune before removing the RBAC rules (serviceaccount, clusterrole and clusterrolebinding).
git clone https://github.com/kubernetes-sigs/node-feature-discovery
+ Developer guide · Node Feature Discovery
Developer guide
Table of contents
Building from source
Download the source code
git clone https://github.com/kubernetes-sigs/node-feature-discovery
cd node-feature-discovery
Docker build
Build the container image
See customizing the build below for altering the container image registry, for example.
make
Push the container image
Optional, this example with Docker.
docker push <IMAGE_TAG>
@@ -12,18 +12,15 @@ kubectl apply -k .
by overriding the variable value
make IMAGE_BUILD_CMD="buildah bud"
Testing
Unit tests are automatically run as part of the container image build. You can also run them manually in the source code tree by running:
make test
End-to-end tests are built on top of the e2e test framework of Kubernetes, and, they required a cluster to run them on. For running the tests on your test cluster you need to specify the kubeconfig to be used:
make e2e-test KUBECONFIG=$HOME/.kube/config
-
There are several environment variables that can be used to customize the e2e-tests:
Variable Description Default value KUBECONFIG Kubeconfig for running e2e-tests empty E2E_TEST_CONFIG Parameterization file of e2e-tests (see example) empty E2E_PULL_IF_NOT_PRESENT True-ish value makes the image pull policy IfNotPresent (to be used only in e2e tests) false E2E_TEST_FULL_IMAGE Run e2e-test also against the Full Image tag false E2E_GINKGO_LABEL_FILTER Ginkgo label filter to use for running e2e tests empty OPENSHIFT Non-empty value enables OpenShift specific support (only affects e2e tests) empty
NFD-Master
When running as a standalone container labeling is expected to fail because Kubernetes API is not available. Thus, it is recommended to use -no-publish.
$ export NFD_CONTAINER_IMAGE=gcr.io/k8s-staging-nfd/node-feature-discovery:master
-$ docker run --rm --name=nfd-test ${NFD_CONTAINER_IMAGE} nfd-master -no-publish -crd-controller=false -feature-gates NodeFeatureAPI=false
-2019/02/01 14:48:21 Node Feature Discovery Master <NFD_VERSION>
-
NFD-Worker
To run nfd-worker as a "stand-alone" container you need to run it in the same network namespace as the nfd-master container:
$ docker run --rm --network=container:nfd-test ${NFD_CONTAINER_IMAGE} nfd-worker -feature-gates NodeFeatureAPI=false
-2019/02/01 14:48:56 Node Feature Discovery Worker <NFD_VERSION>
-...
-
If you just want to try out feature discovery without connecting to nfd-master, pass the -no-publish flag to nfd-worker.
NOTE: Some feature sources need certain directories and/or files from the host mounted inside the NFD container. Thus, you need to provide Docker with the correct --volume options for them to work correctly when run stand-alone directly with docker run. See the default deployment for up-to-date information about the required volume mounts.
NFD-Topology-Updater
To run nfd-topology-updater as a "stand-alone" container you need to run it in with the -no-publish flag to disable communication to the Kubernetes apiserver.
$ docker run --rm ${NFD_CONTAINER_IMAGE} nfd-topology-updater -no-publish
-2019/02/01 14:48:56 Node Feature Discovery Topology Updater <NFD_VERSION>
-...
-
If you just want to try out resource topology discovery without connecting to the Kubernetes API, pass the -no-publish flag to nfd-topology-updater.
NOTE: NFD topology updater needs certain directories and/or files from the host mounted inside the NFD container. Thus, you need to provide Docker with the correct --volume options for them to work correctly when run stand-alone directly with docker run. See the template spec for up-to-date information about the required volume mounts.
Running with Tilt
Another option for building NFD locally is via Tilt tool, which can build container images, push them to a local registry and reload your Kubernetes pods automatically. When using Tilt, you don't have to build container images and re-deploy your pods manually but instead let the Tilt take care of it. Tiltfile is a configuration file for the Tilt and is located at the root directory. To develop NFD with Tilt, follow the steps below.
Prerequisites
- Install Docker
- Setup Docker as a non-root user.
- Install kubectl
- Install kustomize
- Install tilt
- Create a local Kubernetes cluster
- Create image registry first
- Create a Kubernetes cluster. Please note that docker containers will be served as controller node and worker nodes, and NFD-worker will run as a DaemonSet in nested container. Therefore, to make sure the NFD-worker can discover the host features, the host folders "/boot" and "/lib" should be mounted into worker node docker containers when creating the Kubernetes cluster.
-
Start up node feature discovery development environment To start up your Tilt development environment, run at the root of your local NFD codebase.
tilt up
+
There are several environment variables that can be used to customize the e2e-tests:
Variable Description Default value KUBECONFIG Kubeconfig for running e2e-tests empty E2E_TEST_CONFIG Parameterization file of e2e-tests (see example) empty E2E_PULL_IF_NOT_PRESENT True-ish value makes the image pull policy IfNotPresent (to be used only in e2e tests) false E2E_TEST_FULL_IMAGE Run e2e-test also against the Full Image tag false E2E_GINKGO_LABEL_FILTER Ginkgo label filter to use for running e2e tests empty OPENSHIFT Non-empty value enables OpenShift specific support (only affects e2e tests) empty
NFD-Master
For development and debugging it is possible to run nfd-master as a stand-alone binary outside the cluster. The -no-publish flag can be used to prevent nfd-master making changes to the nodes. If -no-publish is not set, nfd-master also requires the NODE_NAME environment variable to be set for cleaning up stale annotations.
make build
+NODE_NAME=<EXISTING_NODE> ./nfd-master -no-publish -kubeconfig ~/.kube/config
+
NFD-Worker
For development and debugging it is possible to run nfd-worker as a stand-alone binary outside the cluster. The -no-publish flag can be used to prevent nfd-worker from creating NodeFeature objects in the target cluster. If the -no-publish is not set, nfd-worker also requires the NODE_NAME and KUBERNETES_NAMESPACE environment variables to be defined to create the NodeFeature object in the target cluster.
make build
+KUBERNETES_NAMESPACE=default NODE_NAME=nonexistent-node ./bin/nfd-worker -kubeconfig ~/.kube/config
+
NOTE: Running nfd-worker locally this way discovers and publishes features of the local development system you're running nfd-worker on.
NFD-Topology-Updater
For development and debugging it is possible to run nfd-topology-updater as a stand-alone binary outside the cluster. However, it requires access to the kubelet's local pod-resources socket and the kubelet http api so in practice it needs to be run on a host acting as a Kubernetes node and thus running kubelet. Running kubelet with --read-only-port=10255 (or readOnlyPort: 10255 in config) makes it possible to connect to kubelet without auth-token (never do this in a production cluster). Also, the -no-publish flag can be used to prevent nfd-topology-updater from creating NodeResourceTopology objects in the target cluster. If the -no-publish is not set, nfd-topology-updater also requires the NODE_NAME and KUBERNETES_NAMESPACE environment variables to be defined.
make build
+KUBERNETES_NAMESPACE=default NODE_NAME=nonexistent-node ./bin/nfd-topology-updater -kubeconfig ~/.kube/config -kubelet-config-uri http://127.0.0.1:10255
+
Running with Tilt
Another option for building NFD locally is via Tilt tool, which can build container images, push them to a local registry and reload your Kubernetes pods automatically. When using Tilt, you don't have to build container images and re-deploy your pods manually but instead let the Tilt take care of it. Tiltfile is a configuration file for the Tilt and is located at the root directory. To develop NFD with Tilt, follow the steps below.
Prerequisites
- Install Docker
- Setup Docker as a non-root user.
- Install kubectl
- Install kustomize
- Install tilt
- Create a local Kubernetes cluster
- Create image registry first
- Create a Kubernetes cluster. Please note that docker containers will be served as controller node and worker nodes, and NFD-worker will run as a DaemonSet in nested container. Therefore, to make sure the NFD-worker can discover the host features, the host folders "/boot" and "/lib" should be mounted into worker node docker containers when creating the Kubernetes cluster.
-
Start up node feature discovery development environment To start up your Tilt development environment, run at the root of your local NFD codebase.
tilt up
Tilt will start a web interface in the localhost and port 10350. From the web interface, you are able to see how NFD worker and master are progressing, watch their build and runtime logs. Once your code changes are saved locally, Tilt will notice it and re-build the container image from the current code, push the image to the registry and re-deploy NFD pods with the latest container image.
Environment variables
To override environment variables used in the Tiltfile during image build, export them in your current terminal before starting Tilt.
export IMAGE_TAG_NAME="v1"
tilt up
This will override the default value(master) of IMAGE_TAG_NAME variable defined in the Tiltfile.
Documentation
All documentation resides under the docs directory in the source tree. It is designed to be served as a html site by GitHub Pages.
Building the documentation is containerized to fix the build environment. The recommended way for developing documentation is to run:
make site-serve
This will build the documentation in a container and serve it under localhost:4000/ making it easy to verify the results. Any changes made to the docs/ will automatically re-trigger a rebuild and are reflected in the served content and can be inspected with a browser refresh.
To just build the html documentation run:
make site-build
-
This will generate html documentation under docs/_site/.
Node Feature Discovery master
\ No newline at end of file
+This will generate html documentation under docs/_site/.
Welcome to Node Feature Discovery – a Kubernetes add-on for detecting hardware features and system configuration!
Continue to:
Introduction for more details on the project.
Quick start for quick step-by-step instructions on how to get NFD running on your cluster.
$ kubectl apply -k https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default?ref=master
+ Get started · Node Feature Discovery
Node Feature Discovery
Welcome to Node Feature Discovery – a Kubernetes add-on for detecting hardware features and system configuration!
Continue to:
-
Introduction for more details on the project.
-
Quick start for quick step-by-step instructions on how to get NFD running on your cluster.
Quick-start – the short-short version
$ kubectl apply -k https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default?ref=master
namespace/node-feature-discovery created
serviceaccount/nfd-master created
clusterrole.rbac.authorization.k8s.io/nfd-master created
clusterrolebinding.rbac.authorization.k8s.io/nfd-master created
configmap/nfd-worker-conf created
- service/nfd-master created
deployment.apps/nfd-master created
daemonset.apps/nfd-worker created
@@ -22,4 +21,4 @@
"feature.node.kubernetes.io/cpu-cpuid.AESNI": "true",
...
-
Node Feature Discovery master
\ No newline at end of file
+This software enables node feature discovery for Kubernetes. It detects hardware features available on each node in a Kubernetes cluster, and advertises those features using node labels and optionally node extended resources, annotations and node taints. Node Feature Discovery is compatible with any recent version of Kubernetes (v1.24+).
NFD consists of four software components:
NFD-Master is the daemon responsible for communication towards the Kubernetes API. That is, it receives labeling requests from the worker and modifies node objects accordingly.
NFD-Worker is a daemon responsible for feature detection. It then communicates the information to nfd-master which does the actual node labeling. One instance of nfd-worker is supposed to be running on each node of the cluster,
NFD-Topology-Updater is a daemon responsible for examining allocated resources on a worker node to account for resources available to be allocated to new pod on a per-zone basis (where a zone can be a NUMA node). It then creates or updates a NodeResourceTopology custom resource object specific to this node. One instance of nfd-topology-updater is supposed to be running on each node of the cluster.
NFD-GC is a daemon responsible for cleaning obsolete NodeFeature and NodeResourceTopology objects.
One instance of nfd-gc is supposed to be running in the cluster.
Feature discovery is divided into domain-specific feature sources:
Each feature source is responsible for detecting a set of features which. in turn, are turned into node feature labels. Feature labels are prefixed with feature.node.kubernetes.io/ and also contain the name of the feature source. Non-standard user-specific feature labels can be created with the local and custom feature sources.
An overview of the default feature labels:
{
+ Introduction · Node Feature Discovery
Introduction
Table of contents
- NFD-Master
- NFD-Worker
- NFD-Topology-Updater
- NFD-GC
- Feature Discovery
- Node annotations
- Custom resources
This software enables node feature discovery for Kubernetes. It detects hardware features available on each node in a Kubernetes cluster, and advertises those features using node labels and optionally node extended resources, annotations and node taints. Node Feature Discovery is compatible with any recent version of Kubernetes (v1.24+).
NFD consists of four software components:
- nfd-master
- nfd-worker
- nfd-topology-updater
- nfd-gc
NFD-Master
NFD-Master is the daemon responsible for communication towards the Kubernetes API. That is, it receives labeling requests from the worker and modifies node objects accordingly.
NFD-Worker
NFD-Worker is a daemon responsible for feature detection. It then communicates the information to nfd-master which does the actual node labeling. One instance of nfd-worker is supposed to be running on each node of the cluster,
NFD-Topology-Updater
NFD-Topology-Updater is a daemon responsible for examining allocated resources on a worker node to account for resources available to be allocated to new pod on a per-zone basis (where a zone can be a NUMA node). It then creates or updates a NodeResourceTopology custom resource object specific to this node. One instance of nfd-topology-updater is supposed to be running on each node of the cluster.
NFD-GC
NFD-GC is a daemon responsible for cleaning obsolete NodeFeature and NodeResourceTopology objects.
One instance of nfd-gc is supposed to be running in the cluster.
Feature Discovery
Feature discovery is divided into domain-specific feature sources:
- CPU
- Kernel
- Memory
- Network
- PCI
- Storage
- System
- USB
- Custom (rule-based custom features)
- Local (features files)
Each feature source is responsible for detecting a set of features which. in turn, are turned into node feature labels. Feature labels are prefixed with feature.node.kubernetes.io/ and also contain the name of the feature source. Non-standard user-specific feature labels can be created with the local and custom feature sources.
An overview of the default feature labels:
{
"feature.node.kubernetes.io/cpu-<feature-name>": "true",
"feature.node.kubernetes.io/custom-<feature-name>": "true",
"feature.node.kubernetes.io/kernel-<feature name>": "<feature value>",
@@ -10,4 +10,4 @@
"feature.node.kubernetes.io/usb-<device label>.present": "<feature value>",
"feature.node.kubernetes.io/<file name>-<feature name>": "<feature value>"
}
-
Node annotations
NFD also annotates nodes it is running on:
Annotation Description [<instance>.]nfd.node.kubernetes.io/feature-labels Comma-separated list of node labels managed by NFD. NFD uses this internally so must not be edited by users. [<instance>.]nfd.node.kubernetes.io/feature-annotations Comma-separated list of node annotations managed by NFD. NFD uses this internally so must not be edited by users. [<instance>.]nfd.node.kubernetes.io/extended-resources Comma-separated list of node extended resources managed by NFD. NFD uses this internally so must not be edited by users. [<instance>.]nfd.node.kubernetes.io/taints Comma-separated list of node taints managed by NFD. NFD uses this internally so must not be edited by users.
NOTE: the -instance command line flag affects the annotation names
Unapplicable annotations are not created, i.e. for example nfd.node.kubernetes.io/extended-resources is only placed if some extended resources were created by NFD.
Custom resources
NFD takes use of some Kubernetes Custom Resources.
NodeFeatures is be used for representing node features and requesting node labels to be generated.
NFD-Master uses NodeFeatureRules for custom labeling of nodes.
NFD-Topology-Updater creates NodeResourceTopology objects that describe the hardware topology of node resources.
Node Feature Discovery master
\ No newline at end of file
+NFD also annotates nodes it is running on:
| Annotation | Description |
|---|---|
| [<instance>.]nfd.node.kubernetes.io/feature-labels | Comma-separated list of node labels managed by NFD. NFD uses this internally so must not be edited by users. |
| [<instance>.]nfd.node.kubernetes.io/feature-annotations | Comma-separated list of node annotations managed by NFD. NFD uses this internally so must not be edited by users. |
| [<instance>.]nfd.node.kubernetes.io/extended-resources | Comma-separated list of node extended resources managed by NFD. NFD uses this internally so must not be edited by users. |
| [<instance>.]nfd.node.kubernetes.io/taints | Comma-separated list of node taints managed by NFD. NFD uses this internally so must not be edited by users. |
NOTE: the
-instancecommand line flag affects the annotation names
Unapplicable annotations are not created, i.e. for example nfd.node.kubernetes.io/extended-resources is only placed if some extended resources were created by NFD.
NFD takes use of some Kubernetes Custom Resources.
NodeFeatures is be used for representing node features and requesting node labels to be generated.
NFD-Master uses NodeFeatureRules for custom labeling of nodes.
NFD-Topology-Updater creates NodeResourceTopology objects that describe the hardware topology of node resources.
Minimal steps to deploy latest released version of NFD in your cluster.
Deploy with kustomize – creates a new namespace, service and required RBAC rules and deploys nfd-master and nfd-worker daemons.
kubectl apply -k "https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default?ref=master"
+ Quick start · Node Feature Discovery
Quick start
Minimal steps to deploy latest released version of NFD in your cluster.
Installation
Deploy with kustomize – creates a new namespace, service and required RBAC rules and deploys nfd-master and nfd-worker daemons.
kubectl apply -k "https://github.com/kubernetes-sigs/node-feature-discovery/deployment/overlays/default?ref=master"
Verify
Wait until NFD master and NFD worker are running.
$ kubectl -n node-feature-discovery get ds,deploy
NAME DESIRED CURRENT READY UP-TO-DATE AVAILABLE NODE SELECTOR AGE
daemonset.apps/nfd-worker 2 2 2 2 2 <none> 10s
@@ -40,4 +40,4 @@ daemonset.apps/nfd-topology-updater 2 2 2 2 2
NAME AGE
kind-control-plane 23s
kind-worker 23s
-
Node Feature Discovery master
\ No newline at end of file
+Feature gates are a set of key-value pairs that control the behavior of NFD. They are used to enable or disable certain features of NFD. The feature gates are set using the -feature-gates command line flag or featureGates value in the Helm chart. The following feature gates are available:
| Name | Default | Stage | Since | Until |
|---|---|---|---|---|
NodeFeatureAPI | true | Beta | V0.14 | v0.16 |
NodeFeatureAPI | true | GA | V0.17 | |
DisableAutoPrefix | false | Alpha | V0.16 | |
NodeFeatureGroupAPI | false | Alpha | V0.16 |
The NodeFeatureAPI feature gate enables the Node Feature API. When enabled, NFD will register the Node Feature API with the Kubernetes API server. The Node Feature API is used to expose node-specific hardware and software features to the Kubernetes scheduler. The Node Feature API is a beta feature and is enabled by default.
The NodeFeatureGroupAPI feature gate enables the Node Feature Group API. When enabled, NFD will register the Node Feature Group API with the Kubernetes API server. The Node Feature Group API is used to create node groups based on hardware and software features. The Node Feature Group API is an alpha feature and is disabled by default.
The DisableAutoPrefix feature gate controls the automatic prefixing of names. When enabled nfd-master does not automatically add the default feature.node.kubernetes.io/ prefix to unprefixed labels, annotations and extended resources. Automatic prefixing is the default behavior in NFD v0.16 and earlier.
Note that enabling the feature gate effectively causes unprefixed names to be filtered out as NFD does not allow unprefixed names of labels, annotations or extended resources. For example, with the DisableAutoPrefix feature gate set to false, a NodeFeatureRule with
labels:
+ Feature Gates · Node Feature Discovery
Feature Gates
Feature gates are a set of key-value pairs that control the behavior of NFD. They are used to enable or disable certain features of NFD. The feature gates are set using the -feature-gates command line flag or featureGates value in the Helm chart. The following feature gates are available:
Name Default Stage Since Until NodeFeatureAPItrue Beta V0.14 v0.16 NodeFeatureAPItrue GA V0.17 DisableAutoPrefixfalse Alpha V0.16 NodeFeatureGroupAPIfalse Alpha V0.16
NodeFeatureAPI
The NodeFeatureAPI feature gate enables the Node Feature API. When enabled, NFD will register the Node Feature API with the Kubernetes API server. The Node Feature API is used to expose node-specific hardware and software features to the Kubernetes scheduler. The Node Feature API is a beta feature and is enabled by default.
NodeFeatureGroupAPI
The NodeFeatureGroupAPI feature gate enables the Node Feature Group API. When enabled, NFD will register the Node Feature Group API with the Kubernetes API server. The Node Feature Group API is used to create node groups based on hardware and software features. The Node Feature Group API is an alpha feature and is disabled by default.
DisableAutoPrefix
The DisableAutoPrefix feature gate controls the automatic prefixing of names. When enabled nfd-master does not automatically add the default feature.node.kubernetes.io/ prefix to unprefixed labels, annotations and extended resources. Automatic prefixing is the default behavior in NFD v0.16 and earlier.
Note that enabling the feature gate effectively causes unprefixed names to be filtered out as NFD does not allow unprefixed names of labels, annotations or extended resources. For example, with the DisableAutoPrefix feature gate set to false, a NodeFeatureRule with
labels:
foo: bar
-
will turn into feature.node.kubernetes.io/foo=bar node label. With DisableAutoPrefix set to true, no prefix is added and the label will be filtered out.
Note that taint keys are not affected by this feature gate.
Node Feature Discovery master
\ No newline at end of file
+will turn into feature.node.kubernetes.io/foo=bar node label. With DisableAutoPrefix set to true, no prefix is added and the label will be filtered out.
Note that taint keys are not affected by this feature gate.
To quickly view available command line flags execute nfd-gc -help. In a docker container:
docker run gcr.io/k8s-staging-nfd/node-feature-discovery:master \
+ Garbage Collector Cmdline Reference · Node Feature Discovery
NFD-GC Commandline Flags
Table of Contents
To quickly view available command line flags execute nfd-gc -help. In a docker container:
docker run gcr.io/k8s-staging-nfd/node-feature-discovery:master \
nfd-gc -help
-h, -help
Print usage and exit.
-version
Print version and exit.
-gc-interval
The -gc-interval specifies the interval between periodic garbage collector runs.
Default: 1h
Example:
nfd-gc -gc-interval=1h
-
Node Feature Discovery master
\ No newline at end of file
+Command line and configuration reference.
Command line and configuration reference.
To quickly view available command line flags execute nfd-master -help. In a docker container:
docker run gcr.io/k8s-staging-nfd/node-feature-discovery:master nfd-master -help
+ Master cmdline reference · Node Feature Discovery
Commandline flags of nfd-master
Table of contents
- -h, -help
- -version
- -feature-gates
- -prune
- -metrics
- -instance
- -enable-leader-election
- -enable-taints
- -no-publish
- -label-whitelist
- -extra-label-ns
- -deny-label-ns
- -config
- -options
- -nfd-api-parallelism
- Logging
- -resync-period
To quickly view available command line flags execute nfd-master -help. In a docker container:
docker run gcr.io/k8s-staging-nfd/node-feature-discovery:master nfd-master -help
-h, -help
Print usage and exit.
-version
Print version and exit.
-feature-gates
The -feature-gates flag is used to enable or disable non GA features. The list of available feature gates can be found in the feature gates documentation.
Example:
nfd-master -feature-gates NodeFeatureGroupAPI=true
-prune
The -prune flag is a sub-command like option for cleaning up the cluster. It causes nfd-master to remove all NFD related labels, annotations and extended resources from all Node objects of the cluster and exit.
-metrics
DEPRECATED: Will be removed in NFD v0.17 and replaced by -port.
The -metrics flag specifies the port on which to expose Prometheus metrics. Setting this to 0 disables the metrics server on nfd-master.
Default: 8081
Example:
nfd-master -metrics=12345
-instance
The -instance flag makes it possible to run multiple NFD deployments in parallel. In practice, it separates the node annotations between deployments so that each of them can store metadata independently. The instance name must start and end with an alphanumeric character and may only contain alphanumeric characters, -, _ or ..
Default: empty
Example:
nfd-master -instance=network
@@ -12,4 +12,4 @@
-options
The -options flag may be used to specify and override configuration file options directly from the command line. The required format is the same as in the config file i.e. JSON or YAML. Configuration options specified via this flag will override those from the configuration file:
Default: empty
Example:
nfd-master -options='{"noPublish": true}'
-nfd-api-parallelism
The -nfd-api-parallelism flag can be used to specify the maximum number of concurrent node updates.
Default: 10
Example:
nfd-master -nfd-api-parallelism=1
Logging
The following logging-related flags are inherited from the klog package.
-add_dir_header
If true, adds the file directory to the header of the log messages.
Default: false
-alsologtostderr
Log to standard error as well as files.
Default: false
-log_backtrace_at
When logging hits line file:N, emit a stack trace.
Default: empty
-log_dir
If non-empty, write log files in this directory.
Default: empty
-log_file
If non-empty, use this log file.
Default: empty
-log_file_max_size
Defines the maximum size a log file can grow to. Unit is megabytes. If the value is 0, the maximum file size is unlimited.
Default: 1800
-logtostderr
Log to standard error instead of files
Default: true
-skip_headers
If true, avoid header prefixes in the log messages.
Default: false
-skip_log_headers
If true, avoid headers when opening log files.
Default: false
-stderrthreshold
Logs at or above this threshold go to stderr.
Default: 2
-v
Number for the log level verbosity.
Default: 0
-vmodule
Comma-separated list of pattern=N settings for file-filtered logging.
Default: empty
-resync-period
The -resync-period flag specifies the NFD API controller resync period. The resync means nfd-master replaying all NodeFeature and NodeFeatureRule objects, thus effectively re-syncing all nodes in the cluster (i.e. ensuring labels, annotations, extended resources and taints are in place).
Default: 1 hour.
Example:
nfd-master -resync-period=2h
-
Node Feature Discovery master
\ No newline at end of file
+See the sample configuration file for a full example configuration.
noPublish option disables updates to the Node objects in the Kubernetes API server, making a "dry-run" flag for nfd-master. No Labels, Annotations, Taints or ExtendedResources of nodes are updated.
Default: false
Example:
noPublish: true
+ Master config reference · Node Feature Discovery
Configuration file reference of nfd-master
Table of contents
- noPublish
- extraLabelNs
- denyLabelNs
- autoDefaultNs
- enableTaints
- labelWhiteList
- resyncPeriod
- leaderElection
- nfdApiParallelism
- klog
- restrictions (EXPERIMENTAL)
See the sample configuration file for a full example configuration.
noPublish
noPublish option disables updates to the Node objects in the Kubernetes API server, making a "dry-run" flag for nfd-master. No Labels, Annotations, Taints or ExtendedResources of nodes are updated.
Default: false
Example:
noPublish: true
extraLabelNs
extraLabelNs specifies a list of allowed feature label namespaces. This option can be used to allow other vendor or application specific namespaces for custom labels from the local and custom feature sources, even though these labels were denied using the denyLabelNs parameter.
Default: empty
Example:
extraLabelNs: ["added.ns.io","added.kubernets.io"]
denyLabelNs
denyLabelNs specifies a list of excluded label namespaces. By default, nfd-master allows creating labels in all namespaces, excluding kubernetes.io namespace and its sub-namespaces (i.e. *.kubernetes.io). However, you should note that kubernetes.io and its sub-namespaces are always denied. This option can be used to exclude some vendors or application specific namespaces.
Default: empty
Example:
denyLabelNs: ["denied.ns.io","denied.kubernetes.io"]
autoDefaultNs
DEPRECATED: Will be removed in NFD v0.17. Use the DisableAutoPrefix feature gate instead.
The autoDefaultNs option controls the automatic prefixing of names. When set to true (the default in NFD version master) nfd-master automatically adds the default feature.node.kubernetes.io/ prefix to unprefixed labels, annotations and extended resources - this is also the default behavior in NFD v0.15 and earlier. When the option is set to false, no prefix will be prepended to unprefixed names, effectively causing them to be filtered out (as NFD does not allow unprefixed names of labels, annotations or extended resources). The default will be changed to false in a future release.
For example, with the autoDefaultNs set to true, a NodeFeatureRule with
labels:
@@ -33,4 +33,4 @@
allowOverwrite: false
restrictions.denyNodeFeatureLabels
The denyNodeFeatureLabels option specifies whether to deny labels from 3rd party NodeFeature objects or not. NodeFeature objects created by nfd-worker are not affected.
Default: false
Example:
restrictions:
denyNodeFeatureLabels: true
-
Node Feature Discovery master
\ No newline at end of file
+To quickly view available command line flags execute kubectl nfd -help.
Print usage and exit.
Validate a NodeFeatureRule file.
The --nodefeature-file flag specifies the path to the NodeFeatureRule file to validate.
Test a NodeFeatureRule file against a node without applying it.
The --kubeconfig flag specifies the path to the kubeconfig file to use for CLI requests.
The --namespace flag specifies the namespace to use for CLI requests. Default: default.
The --nodename flag specifies the name of the node to test the NodeFeatureRule against.
The --nodefeaturerule-file flag specifies the path to the NodeFeatureRule file to test.
Process a NodeFeatureRule file against a NodeFeature file.
The --nodefeaturerule-file flag specifies the path to the NodeFeatureRule file to test.
The --nodefeature-file flag specifies the path to the NodeFeature file to test.
To quickly view available command line flags execute kubectl nfd -help.
Print usage and exit.
Validate a NodeFeatureRule file.
The --nodefeature-file flag specifies the path to the NodeFeatureRule file to validate.
Test a NodeFeatureRule file against a node without applying it.
The --kubeconfig flag specifies the path to the kubeconfig file to use for CLI requests.
The --namespace flag specifies the namespace to use for CLI requests. Default: default.
The --nodename flag specifies the name of the node to test the NodeFeatureRule against.
The --nodefeaturerule-file flag specifies the path to the NodeFeatureRule file to test.
Process a NodeFeatureRule file against a NodeFeature file.
The --nodefeaturerule-file flag specifies the path to the NodeFeatureRule file to test.
The --nodefeature-file flag specifies the path to the NodeFeature file to test.
To quickly view available command line flags execute nfd-topology-updater -help. In a docker container:
docker run gcr.io/k8s-staging-nfd/node-feature-discovery:master \
+ Topology Updater Cmdline Reference · Node Feature Discovery
NFD-Topology-Updater Commandline Flags
Table of Contents
- -h, -help
- -version
- -config
- -no-publish
- -oneshot
- -metrics
- -sleep-interval
- -watch-namespace
- -kubelet-config-uri
- -api-auth-token-file
- -podresources-socket
- -pods-fingerprint
- -kubelet-state-dir
To quickly view available command line flags execute nfd-topology-updater -help. In a docker container:
docker run gcr.io/k8s-staging-nfd/node-feature-discovery:master \
nfd-topology-updater -help
-h, -help
Print usage and exit.
-version
Print version and exit.
-config
The -config flag specifies the path of the nfd-topology-updater configuration file to use.
Default: /etc/kubernetes/node-feature-discovery/nfd-topology-updater.conf
Example:
nfd-topology-updater -config=/opt/nfd/nfd-topology-updater.conf
-no-publish
The -no-publish flag makes for a "dry-run" flag for nfd-topology-updater. NFD-Topology-Updater runs resource hardware topology detection normally, but NodeResourceTopology objects are not created or updated.
Default: false
Example:
nfd-topology-updater -no-publish
@@ -11,4 +11,4 @@ nfd-topology-updater -help
-podresources-socket
The -podresources-socket specifies the path to the Unix socket where kubelet exports a gRPC service to enable discovery of in-use CPUs and devices, and to provide metadata for them.
Default: /host-var/lib/kubelet/pod-resources/kubelet.sock
Example:
nfd-topology-updater -podresources-socket=/var/lib/kubelet/pod-resources/kubelet.sock
-pods-fingerprint
Enables compute and report the pod set fingerprint in the NRT. A pod fingerprint is a compact representation of the "node state" regarding resources.
Default: true
Example:
nfd-topology-updater -pods-fingerprint=false
-kubelet-state-dir
The -kubelet-state-dir specifies the path to the Kubelet state directory, where state and checkpoint files are stored. The files are mount as read-only and cannot be change by the updater. Enabled by default. Passing an empty string will disable the watching.
Default: /host-var/lib/kubelet
Example:
nfd-topology-updater -kubelet-state-dir=/var/lib/kubelet
-
Node Feature Discovery master
\ No newline at end of file
+See the sample configuration file for a full example configuration.
The excludeList specifies a key-value map of allocated resources that should not be examined by the topology-updater agent per node. Each key is a node name with a value as a list of resources that should not be examined by the agent for that specific node.
Default: empty
Example:
excludeList:
+ Topology-Updater config reference · Node Feature Discovery
Configuration file reference of nfd-topology-updater
Table of contents
See the sample configuration file for a full example configuration.
excludeList
The excludeList specifies a key-value map of allocated resources that should not be examined by the topology-updater agent per node. Each key is a node name with a value as a list of resources that should not be examined by the agent for that specific node.
Default: empty
Example:
excludeList:
nodeA: [hugepages-2Mi]
nodeB: [memory]
nodeC: [cpu, hugepages-2Mi]
excludeList.*
excludeList.* is a special value that use to specify all nodes. A resource that would be listed under this key, would be excluded from all nodes.
Default: empty
Example:
excludeList:
'*': [hugepages-2Mi]
-
Node Feature Discovery master
\ No newline at end of file
+Node Feature Discovery follows semantic versioning where the version number consists of three components, i.e. MAJOR.MINOR.PATCH.
The most recent two minor releases (or release branches) of Node Feature Discovery are supported. That is, with X being the latest release, X and X-1 are supported and X-1 reaches end-of-life when X+1 is released.
Built-in feature labels and features are supported for 2 releases after being deprecated, at minimum. That is, if a feature label is deprecated in version X, it will be supported in X+1 and X+2 and may be dropped in X+3.
Command-line flags and configuration file options are supported for 1 more release after being deprecated, at minimum. That is, if option/flag is deprecated in version X, it will be supported in X+1 and may be removed in X+2.
The same policy (support for 1 release after deprecation) also applies to Helm chart parameters.
Node Feature Discovery is compatible with Kubernetes v1.24 and later.
Node Feature Discovery follows semantic versioning where the version number consists of three components, i.e. MAJOR.MINOR.PATCH.
The most recent two minor releases (or release branches) of Node Feature Discovery are supported. That is, with X being the latest release, X and X-1 are supported and X-1 reaches end-of-life when X+1 is released.
Built-in feature labels and features are supported for 2 releases after being deprecated, at minimum. That is, if a feature label is deprecated in version X, it will be supported in X+1 and X+2 and may be dropped in X+3.
Command-line flags and configuration file options are supported for 1 more release after being deprecated, at minimum. That is, if option/flag is deprecated in version X, it will be supported in X+1 and may be removed in X+2.
The same policy (support for 1 release after deprecation) also applies to Helm chart parameters.
Node Feature Discovery is compatible with Kubernetes v1.24 and later.
To quickly view available command line flags execute nfd-worker -help. In a docker container:
docker run gcr.io/k8s-staging-nfd/node-feature-discovery:master nfd-worker -help
+ Worker cmdline reference · Node Feature Discovery
Commandline flags of nfd-worker
Table of contents
- -h, -help
- -version
- -feature-gates
- -config
- -options
- -kubeconfig
- -feature-sources
- -label-sources
- -metrics
- -no-publish
- -no-owner-refs
- -oneshot
- Logging
To quickly view available command line flags execute nfd-worker -help. In a docker container:
docker run gcr.io/k8s-staging-nfd/node-feature-discovery:master nfd-worker -help
-h, -help
Print usage and exit.
-version
Print version and exit.
-feature-gates
The -feature-gates flag is used to enable or disable non GA features. The list of available feature gates can be found in the feature gates documentation.
Example:
nfd-master -feature-gates NodeFeatureGroupAPI=true
-config
The -config flag specifies the path of the nfd-worker configuration file to use.
Default: /etc/kubernetes/node-feature-discovery/nfd-worker.conf
Example:
nfd-worker -config=/opt/nfd/worker.conf
-options
The -options flag may be used to specify and override configuration file options directly from the command line. The required format is the same as in the config file i.e. JSON or YAML. Configuration options specified via this flag will override those from the configuration file:
Default: empty
Example:
nfd-worker -options='{"sources":{"cpu":{"cpuid":{"attributeWhitelist":["AVX","AVX2"]}}}}'
@@ -9,4 +9,4 @@
-no-publish
The -no-publish flag disables all communication with the nfd-master and the Kubernetes API server. It is effectively a "dry-run" flag for nfd-worker. NFD-Worker runs feature detection normally, but no labeling requests are sent to nfd-master and no NodeFeature objects are created or updated in the API server.
NOTE: This flag takes precedence over the core.noPublish configuration file option.
Default: false
Example:
nfd-worker -no-publish
-no-owner-refs
The -no-owner-refs flag disables setting the owner references to Pod of the NodeFeature object.
NOTE: This flag takes precedence over the core.noOwnerRefs configuration file option.
Default: false
Example:
nfd-worker -no-owner-refs
-oneshot
The -oneshot flag causes nfd-worker to exit after one pass of feature detection.
Default: false
Example:
nfd-worker -oneshot -no-publish
-
Logging
The following logging-related flags are inherited from the klog package.
NOTE: The logger setup can also be specified via the core.klog configuration file options. However, the command line flags take precedence over any corresponding config file options specified.
-add_dir_header
If true, adds the file directory to the header of the log messages.
Default: false
-alsologtostderr
Log to standard error as well as files.
Default: false
-log_backtrace_at
When logging hits line file:N, emit a stack trace.
Default: empty
-log_dir
If non-empty, write log files in this directory.
Default: empty
-log_file
If non-empty, use this log file.
Default: empty
-log_file_max_size
Defines the maximum size a log file can grow to. Unit is megabytes. If the value is 0, the maximum file size is unlimited.
Default: 1800
-logtostderr
Log to standard error instead of files
Default: true
-skip_headers
If true, avoid header prefixes in the log messages.
Default: false
-skip_log_headers
If true, avoid headers when opening log files.
Default: false
-stderrthreshold
Logs at or above this threshold go to stderr.
Default: 2
-v
Number for the log level verbosity.
Default: 0
-vmodule
Comma-separated list of pattern=N settings for file-filtered logging.
Default: empty
Node Feature Discovery master
\ No newline at end of file
+The following logging-related flags are inherited from the klog package.
NOTE: The logger setup can also be specified via the
core.klogconfiguration file options. However, the command line flags take precedence over any corresponding config file options specified.
If true, adds the file directory to the header of the log messages.
Default: false
Log to standard error as well as files.
Default: false
When logging hits line file:N, emit a stack trace.
Default: empty
If non-empty, write log files in this directory.
Default: empty
If non-empty, use this log file.
Default: empty
Defines the maximum size a log file can grow to. Unit is megabytes. If the value is 0, the maximum file size is unlimited.
Default: 1800
Log to standard error instead of files
Default: true
If true, avoid header prefixes in the log messages.
Default: false
If true, avoid headers when opening log files.
Default: false
Logs at or above this threshold go to stderr.
Default: 2
Number for the log level verbosity.
Default: 0
Comma-separated list of pattern=N settings for file-filtered logging.
Default: empty
See the sample configuration file for a full example configuration.
The core section contains common configuration settings that are not specific to any particular feature source.
core.sleepInterval specifies the interval between consecutive passes of feature (re-)detection, and thus also the interval between node re-labeling. A non-positive value implies infinite sleep interval, i.e. no re-detection or re-labeling is done.
Default: 60s
Example:
core:
+ Worker config reference · Node Feature Discovery
Configuration file reference of nfd-worker
Table of contents
See the sample configuration file for a full example configuration.
core
The core section contains common configuration settings that are not specific to any particular feature source.
core.sleepInterval
core.sleepInterval specifies the interval between consecutive passes of feature (re-)detection, and thus also the interval between node re-labeling. A non-positive value implies infinite sleep interval, i.e. no re-detection or re-labeling is done.
Default: 60s
Example:
core:
sleepInterval: 60s
core.featureSources
core.featureSources specifies the list of enabled feature sources. A special value all enables all sources. Prefixing a source name with - indicates that the source will be disabled instead - this is only meaningful when used in conjunction with all. This option allows completely disabling the feature detection so that neither standard feature labels are generated nor the raw feature data is available for custom rule processing.
Default: [all]
Example:
core:
# Enable all but cpu and local sources
@@ -67,4 +67,4 @@
matchExpressions:
class: {op: In, value: ["0200"]}
vendor: {op: In, value: ["8086"]}
-
Node Feature Discovery master
\ No newline at end of file
+NFD uses some Kubernetes custom resources.
NodeFeature is an NFD-specific custom resource for communicating node features and node labeling requests. The nfd-master pod watches for NodeFeature objects, labels nodes as specified and uses the listed features as input when evaluating NodeFeatureRules. NodeFeature objects can be used for implementing 3rd party extensions (see customization guide for more details).
apiVersion: nfd.k8s-sigs.io/v1alpha1
+ CRDs · Node Feature Discovery
Custom Resources
Table of contents
NFD uses some Kubernetes custom resources.
NodeFeature
NodeFeature is an NFD-specific custom resource for communicating node features and node labeling requests. The nfd-master pod watches for NodeFeature objects, labels nodes as specified and uses the listed features as input when evaluating NodeFeatureRules. NodeFeature objects can be used for implementing 3rd party extensions (see customization guide for more details).
apiVersion: nfd.k8s-sigs.io/v1alpha1
kind: NodeFeature
metadata:
labels:
@@ -88,4 +88,4 @@
capacity: 3
allocatable: 3
available: 3
-
The NodeResourceTopology objects created by NFD can be used to gain insight into the allocatable resources along with the granularity of those resources at a per-zone level (represented by node-0 and node-1 in the above example) or can be used by an external entity (e.g. topology-aware scheduler plugin) to take an action based on the gathered information.
Node Feature Discovery master
\ No newline at end of file
+The NodeResourceTopology objects created by NFD can be used to gain insight into the allocatable resources along with the granularity of those resources at a per-zone level (represented by node-0 and node-1 in the above example) or can be used by an external entity (e.g. topology-aware scheduler plugin) to take an action based on the gathered information.
NFD provides multiple extension points for vendor and application specific labeling:
NodeFeature objects can be used to communicate "raw" node features and node labeling requests to nfd-master.NodeFeatureRule objects provide a way to deploy custom labeling rules via the Kubernetes API.local feature source of nfd-worker creates labels by reading text files.custom feature source of nfd-worker creates labels based on user-specified rules.NodeFeature objects provide a way for 3rd party extensions to advertise custom features, both as "raw" features that serve as input to NodeFeatureRule objects and as feature labels directly.
Note that RBAC rules must be created for each extension for them to be able to create and manipulate NodeFeature objects in their namespace.
Consider the following referential example:
apiVersion: nfd.k8s-sigs.io/v1alpha1
+ Customization guide · Node Feature Discovery
Customization guide
Table of contents
- Overview
- NodeFeature custom resource
- NodeFeatureRule custom resource
- NodeFeatureGroup custom resource
- Local feature source
- Custom feature source
- Node labels
- Feature rule format
Overview
NFD provides multiple extension points for vendor and application specific labeling:
NodeFeature objects can be used to communicate "raw" node features and node labeling requests to nfd-master.NodeFeatureRule objects provide a way to deploy custom labeling rules via the Kubernetes API.local feature source of nfd-worker creates labels by reading text files.custom feature source of nfd-worker creates labels based on user-specified rules.
NodeFeature custom resource
NodeFeature objects provide a way for 3rd party extensions to advertise custom features, both as "raw" features that serve as input to NodeFeatureRule objects and as feature labels directly.
Note that RBAC rules must be created for each extension for them to be able to create and manipulate NodeFeature objects in their namespace.
A NodeFeature example
Consider the following referential example:
apiVersion: nfd.k8s-sigs.io/v1alpha1
kind: NodeFeature
metadata:
labels:
@@ -363,4 +363,4 @@ vendor.io/my-feature=value
- pci.device:
vendor: "0fff"
device: "abcd"
-
Node Feature Discovery master
\ No newline at end of file
+This page contains usage examples and demos.
A demo on the benefits of using node feature discovery can be found in the source code repository under demo/.
This page contains usage examples and demos.
A demo on the benefits of using node feature discovery can be found in the source code repository under demo/.
Features are advertised as labels in the Kubernetes Node object.
Label creation in nfd-worker is performed by a set of separate modules called label sources. The core.labelSources configuration option (or -label-sources flag) of nfd-worker controls which sources to enable for label generation.
All built-in labels use the feature.node.kubernetes.io label namespace and have the following format.
feature.node.kubernetes.io/<feature> = <value>
-NOTE: Consecutive runs of nfd-worker will update the labels on a given node. If features are not discovered on a consecutive run, the corresponding label will be removed. This includes any restrictions placed on the consecutive run, such as restricting discovered features with the
-label-whitelistflag of nfd-master orcore.labelWhiteListoption of nfd-worker.
| Feature name | Value | Description |
|---|---|---|
cpu-cpuid.<cpuid-flag> | true | CPU capability is supported. NOTE: the capability might be supported but not enabled. |
cpu-cpuid.<cpuid-attribute> | string | CPU attribute value |
cpu-hardware_multithreading | true | Hardware multithreading, such as Intel HTT, enabled (number of logical CPUs is greater than physical CPUs) |
cpu-coprocessor.nx_gzip | true | Nest Accelerator for GZIP is supported(Power). |
cpu-power.sst_bf.enabled | true | Intel SST-BF (Intel Speed Select Technology - Base frequency) enabled |
cpu-pstate.status | string | The status of the Intel pstate driver when in use and enabled, either ‘active' or ‘passive'. |
cpu-pstate.turbo | bool | Set to ‘true' if turbo frequencies are enabled in Intel pstate driver, set to ‘false' if they have been disabled. |
cpu-pstate.scaling_governor | string | The value of the Intel pstate scaling_governor when in use, either ‘powersave' or ‘performance'. |
cpu-cstate.enabled | bool | Set to ‘true' if cstates are set in the intel_idle driver, otherwise set to ‘false'. Unset if intel_idle cpuidle driver is not active. |
cpu-security.sgx.enabled | true | Set to ‘true' if Intel SGX is enabled in BIOS (based on a non-zero sum value of SGX EPC section sizes). |
cpu-security.se.enabled | true | Set to ‘true' if IBM Secure Execution for Linux (IBM Z & LinuxONE) is available and enabled (requires /sys/firmware/uv/prot_virt_host facility) |
cpu-security.tdx.enabled | true | Set to ‘true' if Intel TDX is available on the host and has been enabled (requires /sys/module/kvm_intel/parameters/tdx). |
cpu-security.tdx.protected | true | Set to ‘true' if Intel TDX was used to start the guest node, based on the existence of the "TDX_GUEST" information as part of cpuid features. |
cpu-security.sev.enabled | true | Set to ‘true' if ADM SEV is available on the host and has been enabled (requires /sys/module/kvm_amd/parameters/sev). |
cpu-security.sev.es.enabled | true | Set to ‘true' if ADM SEV-ES is available on the host and has been enabled (requires /sys/module/kvm_amd/parameters/sev_es). |
cpu-security.sev.snp.enabled | true | Set to ‘true' if ADM SEV-SNP is available on the host and has been enabled (requires /sys/module/kvm_amd/parameters/sev_snp). |
cpu-model.vendor_id | string | Comparable CPU vendor ID. |
cpu-model.family | int | CPU family. |
cpu-model.id | int | CPU model number. |
The CPU label source is configurable, see worker configuration and sources.cpu configuration options for details.
| Flag | Description |
|---|---|
| ADX | Multi-Precision Add-Carry Instruction Extensions (ADX) |
| AESNI | Advanced Encryption Standard (AES) New Instructions (AES-NI) |
| APX_F | Intel Advanced Performance Extensions (APX) |
| AVX10 | Intel Advanced Vector Extensions 10 (AVX10) |
| AVX10_256, AVX10_512 | Intel AVX10 256-bit and 512-bit vector support |
| AVX | Advanced Vector Extensions (AVX) |
| AVX2 | Advanced Vector Extensions 2 (AVX2) |
| AVXIFMA | AVX-IFMA instructions |
| AVXVNNI | AVX (VEX encoded) VNNI neural network instructions |
| AMXBF16 | Advanced Matrix Extension, tile multiplication operations on BFLOAT16 numbers |
| AMXINT8 | Advanced Matrix Extension, tile multiplication operations on 8-bit integers |
| AMXFP16 | Advanced Matrix Extension, tile multiplication operations on FP16 numbers |
| AMXFP8 | Advanced Matrix Extension, tile multiplication operations on FP8 numbers |
| AMXTILE | Advanced Matrix Extension, base tile architecture support |
| AVX512BF16 | AVX-512 BFLOAT16 instructions |
| AVX512BITALG | AVX-512 bit Algorithms |
| AVX512BW | AVX-512 byte and word Instructions |
| AVX512CD | AVX-512 conflict detection instructions |
| AVX512DQ | AVX-512 doubleword and quadword instructions |
| AVX512ER | AVX-512 exponential and reciprocal instructions |
| AVX512F | AVX-512 foundation |
| AVX512FP16 | AVX-512 FP16 instructions |
| AVX512IFMA | AVX-512 integer fused multiply-add instructions |
| AVX512PF | AVX-512 prefetch instructions |
| AVX512VBMI | AVX-512 vector bit manipulation instructions |
| AVX512VBMI2 | AVX-512 vector bit manipulation instructions, version 2 |
| AVX512VL | AVX-512 vector length extensions |
| AVX512VNNI | AVX-512 vector neural network instructions |
| AVX512VP2INTERSECT | AVX-512 intersect for D/Q |
| AVX512VPOPCNTDQ | AVX-512 vector population count doubleword and quadword |
| AVXNECONVERT | AVX-NE-CONVERT instructions |
| AVXVNNIINT8 | AVX-VNNI-INT8 instructions |
| AVXVNNIINT16 | AVX-VNNI-INT16 instructions |
| CMPCCXADD | CMPCCXADD instructions |
| ENQCMD | Enqueue Command |
| GFNI | Galois Field New Instructions |
| HYPERVISOR | Running under hypervisor |
| MSRLIST | Read/Write List of Model Specific Registers |
| PREFETCHI | PREFETCHIT0/1 instructions |
| VAES | AVX-512 vector AES instructions |
| VPCLMULQDQ | Carry-less multiplication quadword |
| WRMSRNS | Non-Serializing Write to Model Specific Register |
By default, the following CPUID flags have been blacklisted: AVX10 (use AVX10_VERSION instead), BMI1, BMI2, CLMUL, CMOV, CX16, ERMS, F16C, HTT, LZCNT, MMX, MMXEXT, NX, POPCNT, RDRAND, RDSEED, RDTSCP, SGX, SSE, SSE2, SSE3, SSE4, SSE42, SSSE3 and TDX_GUEST. See sources.cpu configuration options to change the behavior.
See the full list in github.com/klauspost/cpuid.
| Attribute | Description |
|---|---|
| AVX10_VERSION | AVX10 vector ISA version (if supported) |
| Flag | Description |
|---|---|
| IDIVA | Integer divide instructions available in ARM mode |
| IDIVT | Integer divide instructions available in Thumb mode |
| THUMB | Thumb instructions |
| FASTMUL | Fast multiplication |
| VFP | Vector floating point instruction extension (VFP) |
| VFPv3 | Vector floating point extension v3 |
| VFPv4 | Vector floating point extension v4 |
| VFPD32 | VFP with 32 D-registers |
| HALF | Half-word loads and stores |
| EDSP | DSP extensions |
| NEON | NEON SIMD instructions |
| LPAE | Large Physical Address Extensions |
| Flag | Description |
|---|---|
| AES | Announcing the Advanced Encryption Standard |
| EVSTRM | Event Stream Frequency Features |
| FPHP | Half Precision(16bit) Floating Point Data Processing Instructions |
| ASIMDHP | Half Precision(16bit) Asimd Data Processing Instructions |
| ATOMICS | Atomic Instructions to the A64 |
| ASIMRDM | Support for Rounding Double Multiply Add/Subtract |
| PMULL | Optional Cryptographic and CRC32 Instructions |
| JSCVT | Perform Conversion to Match Javascript |
| DCPOP | Persistent Memory Support |
| Feature | Value | Description |
|---|---|---|
kernel-config.<option> | true | Kernel config option is enabled (set ‘y' or ‘m'). Default options are NO_HZ, NO_HZ_IDLE, NO_HZ_FULL and PREEMPT |
kernel-selinux.enabled | true | Selinux is enabled on the node |
kernel-version.full | string | Full kernel version as reported by /proc/sys/kernel/osrelease (e.g. ‘4.5.6-7-g123abcde') |
kernel-version.major | string | First component of the kernel version (e.g. ‘4') |
kernel-version.minor | string | Second component of the kernel version (e.g. ‘5') |
kernel-version.revision | string | Third component of the kernel version (e.g. ‘6') |
The kernel label source is configurable, see worker configuration and sources.kernel configuration options for details.
| Feature | Value | Description |
|---|---|---|
memory-numa | true | Multiple memory nodes i.e. NUMA architecture detected |
memory-nv.present | true | NVDIMM device(s) are present |
memory-nv.dax | true | NVDIMM region(s) configured in DAX mode are present |
memory-swap.enabled | true | Swap is enabled on the node |
| Feature | Value | Description |
|---|---|---|
network-sriov.capable | true | Single Root Input/Output Virtualization (SR-IOV) enabled Network Interface Card(s) present |
network-sriov.configured | true | SR-IOV virtual functions have been configured |
| Feature | Value | Description |
|---|---|---|
pci-<device label>.present | true | PCI device is detected |
pci-<device label>.sriov.capable | true | Single Root Input/Output Virtualization (SR-IOV) enabled PCI device present |
<device label> is format is configurable and set to <class>_<vendor> by default. For more more details about configuration of the pci labels, see sources.pci options and worker configuration instructions.
| Feature | Value | Description |
|---|---|---|
usb-<device label>.present | true | USB device is detected |
<device label> is format is configurable and set to <class>_<vendor>_<device> by default. For more more details about configuration of the usb labels, see sources.usb options and worker configuration instructions.
| Feature | Value | Description |
|---|---|---|
storage-nonrotationaldisk | true | Non-rotational disk, like SSD, is present in the node |
| Feature | Value | Description |
|---|---|---|
system-os_release.ID | string | Operating system identifier |
system-os_release.VERSION_ID | string | Operating system version identifier (e.g. ‘6.7') |
system-os_release.VERSION_ID.major | string | First component of the OS version id (e.g. ‘6') |
system-os_release.VERSION_ID.minor | string | Second component of the OS version id (e.g. ‘7') |
The custom label source is designed for creating user defined labels. However, it has a few statically defined built-in labels:
| Feature | Value | Description |
|---|---|---|
custom-rdma.capable | true | The node has an RDMA capable Network adapter |
custom-rdma.enabled | true | The node has the needed RDMA modules loaded to run RDMA traffic |
NFD has many extension points for creating vendor and application specific labels. See the customization guide for detailed documentation.
NFD is able to create extended resources, see the NodeFeatureRule CRD and its extendedResources field for more details.
Note that NFD is not a replacement for the usage of device plugins.
An example use-case for extended resources could be based on custom feature (created e.g. with feature files that exposes the node SGX EPC memory section size. This value will then be turned into an extended resource of the node, allowing PODs to request that resource and the Kubernetes scheduler to schedule such PODs to only those nodes which have a sufficient capacity of said resource left.
Features are advertised as labels in the Kubernetes Node object.
Label creation in nfd-worker is performed by a set of separate modules called label sources. The core.labelSources configuration option (or -label-sources flag) of nfd-worker controls which sources to enable for label generation.
All built-in labels use the feature.node.kubernetes.io label namespace and have the following format.
feature.node.kubernetes.io/<feature> = <value>
+NOTE: Consecutive runs of nfd-worker will update the labels on a given node. If features are not discovered on a consecutive run, the corresponding label will be removed. This includes any restrictions placed on the consecutive run, such as restricting discovered features with the
-label-whitelistflag of nfd-master orcore.labelWhiteListoption of nfd-worker.
| Feature name | Value | Description |
|---|---|---|
cpu-cpuid.<cpuid-flag> | true | CPU capability is supported. NOTE: the capability might be supported but not enabled. |
cpu-cpuid.<cpuid-attribute> | string | CPU attribute value |
cpu-hardware_multithreading | true | Hardware multithreading, such as Intel HTT, enabled (number of logical CPUs is greater than physical CPUs) |
cpu-coprocessor.nx_gzip | true | Nest Accelerator for GZIP is supported(Power). |
cpu-power.sst_bf.enabled | true | Intel SST-BF (Intel Speed Select Technology - Base frequency) enabled |
cpu-pstate.status | string | The status of the Intel pstate driver when in use and enabled, either ‘active' or ‘passive'. |
cpu-pstate.turbo | bool | Set to ‘true' if turbo frequencies are enabled in Intel pstate driver, set to ‘false' if they have been disabled. |
cpu-pstate.scaling_governor | string | The value of the Intel pstate scaling_governor when in use, either ‘powersave' or ‘performance'. |
cpu-cstate.enabled | bool | Set to ‘true' if cstates are set in the intel_idle driver, otherwise set to ‘false'. Unset if intel_idle cpuidle driver is not active. |
cpu-security.sgx.enabled | true | Set to ‘true' if Intel SGX is enabled in BIOS (based on a non-zero sum value of SGX EPC section sizes). |
cpu-security.se.enabled | true | Set to ‘true' if IBM Secure Execution for Linux (IBM Z & LinuxONE) is available and enabled (requires /sys/firmware/uv/prot_virt_host facility) |
cpu-security.tdx.enabled | true | Set to ‘true' if Intel TDX is available on the host and has been enabled (requires /sys/module/kvm_intel/parameters/tdx). |
cpu-security.tdx.protected | true | Set to ‘true' if Intel TDX was used to start the guest node, based on the existence of the "TDX_GUEST" information as part of cpuid features. |
cpu-security.sev.enabled | true | Set to ‘true' if ADM SEV is available on the host and has been enabled (requires /sys/module/kvm_amd/parameters/sev). |
cpu-security.sev.es.enabled | true | Set to ‘true' if ADM SEV-ES is available on the host and has been enabled (requires /sys/module/kvm_amd/parameters/sev_es). |
cpu-security.sev.snp.enabled | true | Set to ‘true' if ADM SEV-SNP is available on the host and has been enabled (requires /sys/module/kvm_amd/parameters/sev_snp). |
cpu-model.vendor_id | string | Comparable CPU vendor ID. |
cpu-model.family | int | CPU family. |
cpu-model.id | int | CPU model number. |
The CPU label source is configurable, see worker configuration and sources.cpu configuration options for details.
| Flag | Description |
|---|---|
| ADX | Multi-Precision Add-Carry Instruction Extensions (ADX) |
| AESNI | Advanced Encryption Standard (AES) New Instructions (AES-NI) |
| APX_F | Intel Advanced Performance Extensions (APX) |
| AVX10 | Intel Advanced Vector Extensions 10 (AVX10) |
| AVX10_256, AVX10_512 | Intel AVX10 256-bit and 512-bit vector support |
| AVX | Advanced Vector Extensions (AVX) |
| AVX2 | Advanced Vector Extensions 2 (AVX2) |
| AVXIFMA | AVX-IFMA instructions |
| AVXVNNI | AVX (VEX encoded) VNNI neural network instructions |
| AMXBF16 | Advanced Matrix Extension, tile multiplication operations on BFLOAT16 numbers |
| AMXINT8 | Advanced Matrix Extension, tile multiplication operations on 8-bit integers |
| AMXFP16 | Advanced Matrix Extension, tile multiplication operations on FP16 numbers |
| AMXFP8 | Advanced Matrix Extension, tile multiplication operations on FP8 numbers |
| AMXTILE | Advanced Matrix Extension, base tile architecture support |
| AVX512BF16 | AVX-512 BFLOAT16 instructions |
| AVX512BITALG | AVX-512 bit Algorithms |
| AVX512BW | AVX-512 byte and word Instructions |
| AVX512CD | AVX-512 conflict detection instructions |
| AVX512DQ | AVX-512 doubleword and quadword instructions |
| AVX512ER | AVX-512 exponential and reciprocal instructions |
| AVX512F | AVX-512 foundation |
| AVX512FP16 | AVX-512 FP16 instructions |
| AVX512IFMA | AVX-512 integer fused multiply-add instructions |
| AVX512PF | AVX-512 prefetch instructions |
| AVX512VBMI | AVX-512 vector bit manipulation instructions |
| AVX512VBMI2 | AVX-512 vector bit manipulation instructions, version 2 |
| AVX512VL | AVX-512 vector length extensions |
| AVX512VNNI | AVX-512 vector neural network instructions |
| AVX512VP2INTERSECT | AVX-512 intersect for D/Q |
| AVX512VPOPCNTDQ | AVX-512 vector population count doubleword and quadword |
| AVXNECONVERT | AVX-NE-CONVERT instructions |
| AVXVNNIINT8 | AVX-VNNI-INT8 instructions |
| AVXVNNIINT16 | AVX-VNNI-INT16 instructions |
| CMPCCXADD | CMPCCXADD instructions |
| ENQCMD | Enqueue Command |
| GFNI | Galois Field New Instructions |
| HYPERVISOR | Running under hypervisor |
| MSRLIST | Read/Write List of Model Specific Registers |
| PREFETCHI | PREFETCHIT0/1 instructions |
| VAES | AVX-512 vector AES instructions |
| VPCLMULQDQ | Carry-less multiplication quadword |
| WRMSRNS | Non-Serializing Write to Model Specific Register |
By default, the following CPUID flags have been blacklisted: AVX10 (use AVX10_VERSION instead), BMI1, BMI2, CLMUL, CMOV, CX16, ERMS, F16C, HTT, LZCNT, MMX, MMXEXT, NX, POPCNT, RDRAND, RDSEED, RDTSCP, SGX, SSE, SSE2, SSE3, SSE4, SSE42, SSSE3 and TDX_GUEST. See sources.cpu configuration options to change the behavior.
See the full list in github.com/klauspost/cpuid.
| Attribute | Description |
|---|---|
| AVX10_VERSION | AVX10 vector ISA version (if supported) |
| Flag | Description |
|---|---|
| IDIVA | Integer divide instructions available in ARM mode |
| IDIVT | Integer divide instructions available in Thumb mode |
| THUMB | Thumb instructions |
| FASTMUL | Fast multiplication |
| VFP | Vector floating point instruction extension (VFP) |
| VFPv3 | Vector floating point extension v3 |
| VFPv4 | Vector floating point extension v4 |
| VFPD32 | VFP with 32 D-registers |
| HALF | Half-word loads and stores |
| EDSP | DSP extensions |
| NEON | NEON SIMD instructions |
| LPAE | Large Physical Address Extensions |
| Flag | Description |
|---|---|
| AES | Announcing the Advanced Encryption Standard |
| EVSTRM | Event Stream Frequency Features |
| FPHP | Half Precision(16bit) Floating Point Data Processing Instructions |
| ASIMDHP | Half Precision(16bit) Asimd Data Processing Instructions |
| ATOMICS | Atomic Instructions to the A64 |
| ASIMRDM | Support for Rounding Double Multiply Add/Subtract |
| PMULL | Optional Cryptographic and CRC32 Instructions |
| JSCVT | Perform Conversion to Match Javascript |
| DCPOP | Persistent Memory Support |
| Feature | Value | Description |
|---|---|---|
kernel-config.<option> | true | Kernel config option is enabled (set ‘y' or ‘m'). Default options are NO_HZ, NO_HZ_IDLE, NO_HZ_FULL and PREEMPT |
kernel-selinux.enabled | true | Selinux is enabled on the node |
kernel-version.full | string | Full kernel version as reported by /proc/sys/kernel/osrelease (e.g. ‘4.5.6-7-g123abcde') |
kernel-version.major | string | First component of the kernel version (e.g. ‘4') |
kernel-version.minor | string | Second component of the kernel version (e.g. ‘5') |
kernel-version.revision | string | Third component of the kernel version (e.g. ‘6') |
The kernel label source is configurable, see worker configuration and sources.kernel configuration options for details.
| Feature | Value | Description |
|---|---|---|
memory-numa | true | Multiple memory nodes i.e. NUMA architecture detected |
memory-nv.present | true | NVDIMM device(s) are present |
memory-nv.dax | true | NVDIMM region(s) configured in DAX mode are present |
memory-swap.enabled | true | Swap is enabled on the node |
| Feature | Value | Description |
|---|---|---|
network-sriov.capable | true | Single Root Input/Output Virtualization (SR-IOV) enabled Network Interface Card(s) present |
network-sriov.configured | true | SR-IOV virtual functions have been configured |
| Feature | Value | Description |
|---|---|---|
pci-<device label>.present | true | PCI device is detected |
pci-<device label>.sriov.capable | true | Single Root Input/Output Virtualization (SR-IOV) enabled PCI device present |
<device label> is format is configurable and set to <class>_<vendor> by default. For more more details about configuration of the pci labels, see sources.pci options and worker configuration instructions.
| Feature | Value | Description |
|---|---|---|
usb-<device label>.present | true | USB device is detected |
<device label> is format is configurable and set to <class>_<vendor>_<device> by default. For more more details about configuration of the usb labels, see sources.usb options and worker configuration instructions.
| Feature | Value | Description |
|---|---|---|
storage-nonrotationaldisk | true | Non-rotational disk, like SSD, is present in the node |
| Feature | Value | Description |
|---|---|---|
system-os_release.ID | string | Operating system identifier |
system-os_release.VERSION_ID | string | Operating system version identifier (e.g. ‘6.7') |
system-os_release.VERSION_ID.major | string | First component of the OS version id (e.g. ‘6') |
system-os_release.VERSION_ID.minor | string | Second component of the OS version id (e.g. ‘7') |
The custom label source is designed for creating user defined labels. However, it has a few statically defined built-in labels:
| Feature | Value | Description |
|---|---|---|
custom-rdma.capable | true | The node has an RDMA capable Network adapter |
custom-rdma.enabled | true | The node has the needed RDMA modules loaded to run RDMA traffic |
NFD has many extension points for creating vendor and application specific labels. See the customization guide for detailed documentation.
NFD is able to create extended resources, see the NodeFeatureRule CRD and its extendedResources field for more details.
Note that NFD is not a replacement for the usage of device plugins.
An example use-case for extended resources could be based on custom feature (created e.g. with feature files that exposes the node SGX EPC memory section size. This value will then be turned into an extended resource of the node, allowing PODs to request that resource and the Kubernetes scheduler to schedule such PODs to only those nodes which have a sufficient capacity of said resource left.
Usage instructions.
Usage instructions.
Developer Preview This feature is currently in developer preview and subject to change. It is not recommended to use it in production environments.
The kubectl plugin kubectl nfd can be used to validate/dryrun and test NodeFeatureRule objects. It can be installed with the following command:
git clone https://github.com/kubernetes-sigs/node-feature-discovery
+ Kubectl plugin · Node Feature Discovery
Kubectl plugin
Table of contents
Developer Preview This feature is currently in developer preview and subject to change. It is not recommended to use it in production environments.
Overview
The kubectl plugin kubectl nfd can be used to validate/dryrun and test NodeFeatureRule objects. It can be installed with the following command:
git clone https://github.com/kubernetes-sigs/node-feature-discovery
cd node-feature-discovery
make build-kubectl-nfd
KUBECTL_PATH=/usr/local/bin/
@@ -13,4 +13,4 @@ Processing rule: my sample rule
*** Labels ***
vendor.io/my-sample-feature=true
NodeFeatureRule "examples/nodefeaturerule.yaml" is valid for NodeFeature "examples/nodefeature.yaml"
-
Node Feature Discovery master
\ No newline at end of file
+NFD-GC (NFD Garbage-Collector) is preferably run as a Kubernetes deployment with one replica. It makes sure that all NodeFeature and NodeResourceTopology objects have corresponding nodes and removes stale objects for non-existent nodes.
The daemon watches for Node deletion events and removes NodeFeature and NodeResourceTopology objects upon them. It also runs periodically to make sure no node delete event was missed and to remove any NodeFeature or NodeResourceTopology objects that were created without corresponding node. The default garbage collector interval is set to 1h which is the value when no -gc-interval is specified.
In Helm deployments see garbage collector parameters for altering the nfd-gc configuration.
NFD-GC (NFD Garbage-Collector) is preferably run as a Kubernetes deployment with one replica. It makes sure that all NodeFeature and NodeResourceTopology objects have corresponding nodes and removes stale objects for non-existent nodes.
The daemon watches for Node deletion events and removes NodeFeature and NodeResourceTopology objects upon them. It also runs periodically to make sure no node delete event was missed and to remove any NodeFeature or NodeResourceTopology objects that were created without corresponding node. The default garbage collector interval is set to 1h which is the value when no -gc-interval is specified.
In Helm deployments see garbage collector parameters for altering the nfd-gc configuration.
NFD-Master is responsible for connecting to the Kubernetes API server and updating node objects. More specifically, it modifies node labels, taints and extended resources based on requests from nfd-workers and 3rd party extensions.
The NodeFeature Controller uses NodeFeature objects as the input for the NodeFeatureRule processing pipeline. In addition, any labels listed in the NodeFeature object are created on the node (note the allowed label namespaces are controlled).
NFD-Master acts as the controller for NodeFeatureRule objects. It applies the rules specified in NodeFeatureRule objects on raw feature data and creates node labels accordingly. The feature data used as the input is received from nfd-worker instances through NodeFeature objects.
NFD-Master supports configuration through a configuration file. The default location is /etc/kubernetes/node-feature-discovery/nfd-master.conf, but, this can be changed by specifying the-config command line flag.
Master configuration file is read inside the container, and thus, Volumes and VolumeMounts are needed to make your configuration available for NFD. The preferred method is to use a ConfigMap which provides easy deployment and re-configurability.
The provided deployment methods (Helm and Kustomize) create an empty configmap and mount it inside the nfd-master containers.
In Helm deployments, Master pod parameter master.config can be used to edit the respective configuration.
In Kustomize deployments, modify the nfd-master-conf ConfigMap with a custom overlay.
NOTE: dynamic run-time reconfiguration was dropped in NFD v0.17. Re-configuration is handled by pod restarts.
See nfd-master configuration file reference for more details. The (empty-by-default) example config contains all available configuration options and can be used as a reference for creating a configuration.
NFD-Master runs as a deployment, by default it prefers running on the cluster's master nodes but will run on worker nodes if no master nodes are found.
For High Availability, you should increase the replica count of the deployment object. You should also look into adding inter-pod affinity to prevent masters from running on the same node. However note that inter-pod affinity is costly and is not recommended in bigger clusters.
Note: When NFD-Master is intended to run with more than one replica, it is advised to use
-enable-leader-electionflag. This flag turns on leader election for NFD-Master and let only one replica to act on changes in NodeFeature and NodeFeatureRule objects.
If you have RBAC authorization enabled (as is the default e.g. with clusters initialized with kubeadm) you need to configure the appropriate ClusterRoles, ClusterRoleBindings and a ServiceAccount for NFD to create node labels. The provided template will configure these for you.
NFD-Master is responsible for connecting to the Kubernetes API server and updating node objects. More specifically, it modifies node labels, taints and extended resources based on requests from nfd-workers and 3rd party extensions.
The NodeFeature Controller uses NodeFeature objects as the input for the NodeFeatureRule processing pipeline. In addition, any labels listed in the NodeFeature object are created on the node (note the allowed label namespaces are controlled).
NFD-Master acts as the controller for NodeFeatureRule objects. It applies the rules specified in NodeFeatureRule objects on raw feature data and creates node labels accordingly. The feature data used as the input is received from nfd-worker instances through NodeFeature objects.
NFD-Master supports configuration through a configuration file. The default location is /etc/kubernetes/node-feature-discovery/nfd-master.conf, but, this can be changed by specifying the-config command line flag.
Master configuration file is read inside the container, and thus, Volumes and VolumeMounts are needed to make your configuration available for NFD. The preferred method is to use a ConfigMap which provides easy deployment and re-configurability.
The provided deployment methods (Helm and Kustomize) create an empty configmap and mount it inside the nfd-master containers.
In Helm deployments, Master pod parameter master.config can be used to edit the respective configuration.
In Kustomize deployments, modify the nfd-master-conf ConfigMap with a custom overlay.
NOTE: dynamic run-time reconfiguration was dropped in NFD v0.17. Re-configuration is handled by pod restarts.
See nfd-master configuration file reference for more details. The (empty-by-default) example config contains all available configuration options and can be used as a reference for creating a configuration.
NFD-Master runs as a deployment, by default it prefers running on the cluster's master nodes but will run on worker nodes if no master nodes are found.
For High Availability, you should increase the replica count of the deployment object. You should also look into adding inter-pod affinity to prevent masters from running on the same node. However note that inter-pod affinity is costly and is not recommended in bigger clusters.
Note: When NFD-Master is intended to run with more than one replica, it is advised to use
-enable-leader-electionflag. This flag turns on leader election for NFD-Master and let only one replica to act on changes in NodeFeature and NodeFeatureRule objects.
If you have RBAC authorization enabled (as is the default e.g. with clusters initialized with kubeadm) you need to configure the appropriate ClusterRoles, ClusterRoleBindings and a ServiceAccount for NFD to create node labels. The provided template will configure these for you.
NFD-Topology-Updater is preferably run as a Kubernetes DaemonSet. This assures re-examination on regular intervals and/or per pod life-cycle events, capturing changes in the allocated resources and hence the allocatable resources on a per-zone basis by updating NodeResourceTopology custom resources. It makes sure that new NodeResourceTopology instances are created for each new nodes that get added to the cluster.
Because of the design and implementation of Kubernetes, only resources exclusively allocated to Guaranteed Quality of Service pods will be accounted. This includes CPU cores, memory and devices.
When run as a daemonset, nodes are re-examined for the allocated resources (to determine the information of the allocatable resources on a per-zone basis where a zone can be a NUMA node) at an interval specified using the -sleep-interval option. The default sleep interval is set to 60s which is the value when no -sleep-interval is specified. The re-examination can be disabled by setting the sleep-interval to 0.
Another option is to configure the updater to update the allocated resources per pod life-cycle events. The updater will monitor the checkpoint file stated in -kubelet-state-dir and triggers an update for every change occurs in the files.
In addition, it can avoid examining specific allocated resources given a configuration of resources to exclude via -excludeList
Kubelet PodResource API with the GetAllocatableResources functionality enabled is a prerequisite for nfd-topology-updater to be able to run (i.e. Kubernetes v1.21 or later is required).
Preceding Kubernetes v1.23, the kubelet must be started with --feature-gates=KubeletPodResourcesGetAllocatable=true.
Starting from Kubernetes v1.23, the KubeletPodResourcesGetAllocatable feature gate. is enabled by default
NFD-Topology-Updater supports configuration through a configuration file. The default location is /etc/kubernetes/node-feature-discovery/topology-updater.conf, but, this can be changed by specifying the-config command line flag.
Topology-Updater configuration file is read inside the container, and thus, Volumes and VolumeMounts are needed to make your configuration available for NFD. The preferred method is to use a ConfigMap which provides easy deployment and re-configurability.
The provided deployment templates create an empty configmap and mount it inside the nfd-topology-updater containers.
In Helm deployments, Topology Updater parameters toplogyUpdater.config can be used to edit the respective configuration.
In Kustomize deployments, modify the nfd-worker-conf ConfigMap with a custom overlay.
See nfd-topology-updater configuration file reference for more details. The (empty-by-default) example config contains all available configuration options and can be used as a reference for creating a configuration.
NFD-Topology-Updater is preferably run as a Kubernetes DaemonSet. This assures re-examination on regular intervals and/or per pod life-cycle events, capturing changes in the allocated resources and hence the allocatable resources on a per-zone basis by updating NodeResourceTopology custom resources. It makes sure that new NodeResourceTopology instances are created for each new nodes that get added to the cluster.
Because of the design and implementation of Kubernetes, only resources exclusively allocated to Guaranteed Quality of Service pods will be accounted. This includes CPU cores, memory and devices.
When run as a daemonset, nodes are re-examined for the allocated resources (to determine the information of the allocatable resources on a per-zone basis where a zone can be a NUMA node) at an interval specified using the -sleep-interval option. The default sleep interval is set to 60s which is the value when no -sleep-interval is specified. The re-examination can be disabled by setting the sleep-interval to 0.
Another option is to configure the updater to update the allocated resources per pod life-cycle events. The updater will monitor the checkpoint file stated in -kubelet-state-dir and triggers an update for every change occurs in the files.
In addition, it can avoid examining specific allocated resources given a configuration of resources to exclude via -excludeList
Kubelet PodResource API with the GetAllocatableResources functionality enabled is a prerequisite for nfd-topology-updater to be able to run (i.e. Kubernetes v1.21 or later is required).
Preceding Kubernetes v1.23, the kubelet must be started with --feature-gates=KubeletPodResourcesGetAllocatable=true.
Starting from Kubernetes v1.23, the KubeletPodResourcesGetAllocatable feature gate. is enabled by default
NFD-Topology-Updater supports configuration through a configuration file. The default location is /etc/kubernetes/node-feature-discovery/topology-updater.conf, but, this can be changed by specifying the-config command line flag.
Topology-Updater configuration file is read inside the container, and thus, Volumes and VolumeMounts are needed to make your configuration available for NFD. The preferred method is to use a ConfigMap which provides easy deployment and re-configurability.
The provided deployment templates create an empty configmap and mount it inside the nfd-topology-updater containers.
In Helm deployments, Topology Updater parameters toplogyUpdater.config can be used to edit the respective configuration.
In Kustomize deployments, modify the nfd-worker-conf ConfigMap with a custom overlay.
See nfd-topology-updater configuration file reference for more details. The (empty-by-default) example config contains all available configuration options and can be used as a reference for creating a configuration.
NFD-Worker is preferably run as a Kubernetes DaemonSet. This assures re-labeling on regular intervals capturing changes in the system configuration and makes sure that new nodes are labeled as they are added to the cluster. Worker connects to the nfd-master service to advertise hardware features.
When run as a daemonset, nodes are re-labeled at an default interval of 60s. This can be changed by using the core.sleepInterval config option.
NFD-Worker supports configuration through a configuration file. The default location is /etc/kubernetes/node-feature-discovery/nfd-worker.conf, but, this can be changed by specifying the-config command line flag. Configuration file is re-read whenever it is modified which makes run-time re-configuration of nfd-worker straightforward.
Worker configuration file is read inside the container, and thus, Volumes and VolumeMounts are needed to make your configuration available for NFD. The preferred method is to use a ConfigMap which provides easy deployment and re-configurability.
The provided deployment methods (Helm and Kustomize) create an empty configmap and mount it inside the nfd-master containers.
In Helm deployments, Worker pod parameter worker.config can be used to edit the respective configuration.
In Kustomize deployments, modify the nfd-worker-conf ConfigMap with a custom overlay.
NOTE: dynamic run-time reconfiguration was dropped in NFD v0.17. Re-configuration is handled by pod restarts.
See nfd-worker configuration file reference for more details. The (empty-by-default) example config contains all available configuration options and can be used as a reference for creating a configuration.
Configuration options can also be specified via the -options command line flag, in which case no mounts need to be used. The same format as in the config file must be used, i.e. JSON (or YAML). For example:
-options='{"sources": { "pci": { "deviceClassWhitelist": ["12"] } } }'
-Configuration options specified from the command line will override those read from the config file.
NFD-Worker is preferably run as a Kubernetes DaemonSet. This assures re-labeling on regular intervals capturing changes in the system configuration and makes sure that new nodes are labeled as they are added to the cluster. Worker connects to the nfd-master service to advertise hardware features.
When run as a daemonset, nodes are re-labeled at an default interval of 60s. This can be changed by using the core.sleepInterval config option.
NFD-Worker supports configuration through a configuration file. The default location is /etc/kubernetes/node-feature-discovery/nfd-worker.conf, but, this can be changed by specifying the-config command line flag. Configuration file is re-read whenever it is modified which makes run-time re-configuration of nfd-worker straightforward.
Worker configuration file is read inside the container, and thus, Volumes and VolumeMounts are needed to make your configuration available for NFD. The preferred method is to use a ConfigMap which provides easy deployment and re-configurability.
The provided deployment methods (Helm and Kustomize) create an empty configmap and mount it inside the nfd-master containers.
In Helm deployments, Worker pod parameter worker.config can be used to edit the respective configuration.
In Kustomize deployments, modify the nfd-worker-conf ConfigMap with a custom overlay.
NOTE: dynamic run-time reconfiguration was dropped in NFD v0.17. Re-configuration is handled by pod restarts.
See nfd-worker configuration file reference for more details. The (empty-by-default) example config contains all available configuration options and can be used as a reference for creating a configuration.
Configuration options can also be specified via the -options command line flag, in which case no mounts need to be used. The same format as in the config file must be used, i.e. JSON (or YAML). For example:
-options='{"sources": { "pci": { "deviceClassWhitelist": ["12"] } } }'
+Configuration options specified from the command line will override those read from the config file.
Nodes with specific features can be targeted using the nodeSelector field. The following example shows how to target nodes with Intel TurboBoost enabled.
apiVersion: v1
+ Using node labels · Node Feature Discovery
Using node labels
Nodes with specific features can be targeted using the nodeSelector field. The following example shows how to target nodes with Intel TurboBoost enabled.
apiVersion: v1
kind: Pod
metadata:
labels:
@@ -10,4 +10,4 @@
name: go1
nodeSelector:
feature.node.kubernetes.io/cpu-pstate.turbo: 'true'
-
For more details on targeting nodes, see node selection.
Node Feature Discovery master
\ No newline at end of file
+For more details on targeting nodes, see node selection.