1
0
Fork 0
mirror of https://github.com/kubernetes-sigs/node-feature-discovery.git synced 2024-12-14 11:57:51 +00:00
node-feature-discovery/cmd/nfd-topology-updater/main.go

151 lines
5.5 KiB
Go
Raw Normal View History

topologyupdater: Bootstrap nfd-topology-updater in NFD - This patch allows to expose Resource Hardware Topology information through CRDs in Node Feature Discovery. - In order to do this we introduce another software component called nfd-topology-updater in addition to the already existing software components nfd-master and nfd-worker. - nfd-master was enhanced to communicate with nfd-topology-updater over gRPC followed by creation of CRs corresponding to the nodes in the cluster exposing resource hardware topology information of that node. - Pin kubernetes dependency to one that include pod resource implementation - This code is responsible for obtaining hardware information from the system as well as pod resource information from the Pod Resource API in order to determine the allocatable resource information for each NUMA zone. This information along with Costs for NUMA zones (obtained by reading NUMA distances) is gathered by nfd-topology-updater running on all the nodes of the cluster and propagate NUMA zone costs to master in order to populate that information in the CRs corresponding to the nodes. - We use GHW facilities for obtaining system information like CPUs, topology, NUMA distances etc. - This also includes updates made to Makefile and Dockerfile and Manifests for deploying nfd-topology-updater. - This patch includes unit tests - As part of the Topology Aware Scheduling work, this patch captures the configured Topology manager scope in addition to the Topology manager policy. Based on the value of both attribues a single string will be populated to the CRD. The string value will be on of the following {SingleNUMANodeContainerLevel, SingleNUMANodePodLevel, BestEffort, Restricted, None} Co-Authored-by: Artyom Lukianov <alukiano@redhat.com> Co-Authored-by: Francesco Romani <fromani@redhat.com> Co-Authored-by: Talor Itzhak <titzhak@redhat.com> Signed-off-by: Swati Sehgal <swsehgal@redhat.com>
2021-05-13 10:55:33 +00:00
/*
Copyright 2021 The Kubernetes Authors.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
*/
package main
import (
"flag"
"fmt"
"net/url"
topologyupdater: Bootstrap nfd-topology-updater in NFD - This patch allows to expose Resource Hardware Topology information through CRDs in Node Feature Discovery. - In order to do this we introduce another software component called nfd-topology-updater in addition to the already existing software components nfd-master and nfd-worker. - nfd-master was enhanced to communicate with nfd-topology-updater over gRPC followed by creation of CRs corresponding to the nodes in the cluster exposing resource hardware topology information of that node. - Pin kubernetes dependency to one that include pod resource implementation - This code is responsible for obtaining hardware information from the system as well as pod resource information from the Pod Resource API in order to determine the allocatable resource information for each NUMA zone. This information along with Costs for NUMA zones (obtained by reading NUMA distances) is gathered by nfd-topology-updater running on all the nodes of the cluster and propagate NUMA zone costs to master in order to populate that information in the CRs corresponding to the nodes. - We use GHW facilities for obtaining system information like CPUs, topology, NUMA distances etc. - This also includes updates made to Makefile and Dockerfile and Manifests for deploying nfd-topology-updater. - This patch includes unit tests - As part of the Topology Aware Scheduling work, this patch captures the configured Topology manager scope in addition to the Topology manager policy. Based on the value of both attribues a single string will be populated to the CRD. The string value will be on of the following {SingleNUMANodeContainerLevel, SingleNUMANodePodLevel, BestEffort, Restricted, None} Co-Authored-by: Artyom Lukianov <alukiano@redhat.com> Co-Authored-by: Francesco Romani <fromani@redhat.com> Co-Authored-by: Talor Itzhak <titzhak@redhat.com> Signed-off-by: Swati Sehgal <swsehgal@redhat.com>
2021-05-13 10:55:33 +00:00
"os"
"time"
"k8s.io/klog/v2"
kubeletconfigv1beta1 "k8s.io/kubelet/config/v1beta1"
topologyupdater: Bootstrap nfd-topology-updater in NFD - This patch allows to expose Resource Hardware Topology information through CRDs in Node Feature Discovery. - In order to do this we introduce another software component called nfd-topology-updater in addition to the already existing software components nfd-master and nfd-worker. - nfd-master was enhanced to communicate with nfd-topology-updater over gRPC followed by creation of CRs corresponding to the nodes in the cluster exposing resource hardware topology information of that node. - Pin kubernetes dependency to one that include pod resource implementation - This code is responsible for obtaining hardware information from the system as well as pod resource information from the Pod Resource API in order to determine the allocatable resource information for each NUMA zone. This information along with Costs for NUMA zones (obtained by reading NUMA distances) is gathered by nfd-topology-updater running on all the nodes of the cluster and propagate NUMA zone costs to master in order to populate that information in the CRs corresponding to the nodes. - We use GHW facilities for obtaining system information like CPUs, topology, NUMA distances etc. - This also includes updates made to Makefile and Dockerfile and Manifests for deploying nfd-topology-updater. - This patch includes unit tests - As part of the Topology Aware Scheduling work, this patch captures the configured Topology manager scope in addition to the Topology manager policy. Based on the value of both attribues a single string will be populated to the CRD. The string value will be on of the following {SingleNUMANodeContainerLevel, SingleNUMANodePodLevel, BestEffort, Restricted, None} Co-Authored-by: Artyom Lukianov <alukiano@redhat.com> Co-Authored-by: Francesco Romani <fromani@redhat.com> Co-Authored-by: Talor Itzhak <titzhak@redhat.com> Signed-off-by: Swati Sehgal <swsehgal@redhat.com>
2021-05-13 10:55:33 +00:00
topology "sigs.k8s.io/node-feature-discovery/pkg/nfd-client/topology-updater"
"sigs.k8s.io/node-feature-discovery/pkg/resourcemonitor"
"sigs.k8s.io/node-feature-discovery/pkg/topologypolicy"
"sigs.k8s.io/node-feature-discovery/pkg/utils"
"sigs.k8s.io/node-feature-discovery/pkg/utils/hostpath"
"sigs.k8s.io/node-feature-discovery/pkg/utils/kubeconf"
topologyupdater: Bootstrap nfd-topology-updater in NFD - This patch allows to expose Resource Hardware Topology information through CRDs in Node Feature Discovery. - In order to do this we introduce another software component called nfd-topology-updater in addition to the already existing software components nfd-master and nfd-worker. - nfd-master was enhanced to communicate with nfd-topology-updater over gRPC followed by creation of CRs corresponding to the nodes in the cluster exposing resource hardware topology information of that node. - Pin kubernetes dependency to one that include pod resource implementation - This code is responsible for obtaining hardware information from the system as well as pod resource information from the Pod Resource API in order to determine the allocatable resource information for each NUMA zone. This information along with Costs for NUMA zones (obtained by reading NUMA distances) is gathered by nfd-topology-updater running on all the nodes of the cluster and propagate NUMA zone costs to master in order to populate that information in the CRs corresponding to the nodes. - We use GHW facilities for obtaining system information like CPUs, topology, NUMA distances etc. - This also includes updates made to Makefile and Dockerfile and Manifests for deploying nfd-topology-updater. - This patch includes unit tests - As part of the Topology Aware Scheduling work, this patch captures the configured Topology manager scope in addition to the Topology manager policy. Based on the value of both attribues a single string will be populated to the CRD. The string value will be on of the following {SingleNUMANodeContainerLevel, SingleNUMANodePodLevel, BestEffort, Restricted, None} Co-Authored-by: Artyom Lukianov <alukiano@redhat.com> Co-Authored-by: Francesco Romani <fromani@redhat.com> Co-Authored-by: Talor Itzhak <titzhak@redhat.com> Signed-off-by: Swati Sehgal <swsehgal@redhat.com>
2021-05-13 10:55:33 +00:00
"sigs.k8s.io/node-feature-discovery/pkg/version"
)
const (
// ProgramName is the canonical name of this program
ProgramName = "nfd-topology-updater"
kubeletSecurePort = 10250
topologyupdater: Bootstrap nfd-topology-updater in NFD - This patch allows to expose Resource Hardware Topology information through CRDs in Node Feature Discovery. - In order to do this we introduce another software component called nfd-topology-updater in addition to the already existing software components nfd-master and nfd-worker. - nfd-master was enhanced to communicate with nfd-topology-updater over gRPC followed by creation of CRs corresponding to the nodes in the cluster exposing resource hardware topology information of that node. - Pin kubernetes dependency to one that include pod resource implementation - This code is responsible for obtaining hardware information from the system as well as pod resource information from the Pod Resource API in order to determine the allocatable resource information for each NUMA zone. This information along with Costs for NUMA zones (obtained by reading NUMA distances) is gathered by nfd-topology-updater running on all the nodes of the cluster and propagate NUMA zone costs to master in order to populate that information in the CRs corresponding to the nodes. - We use GHW facilities for obtaining system information like CPUs, topology, NUMA distances etc. - This also includes updates made to Makefile and Dockerfile and Manifests for deploying nfd-topology-updater. - This patch includes unit tests - As part of the Topology Aware Scheduling work, this patch captures the configured Topology manager scope in addition to the Topology manager policy. Based on the value of both attribues a single string will be populated to the CRD. The string value will be on of the following {SingleNUMANodeContainerLevel, SingleNUMANodePodLevel, BestEffort, Restricted, None} Co-Authored-by: Artyom Lukianov <alukiano@redhat.com> Co-Authored-by: Francesco Romani <fromani@redhat.com> Co-Authored-by: Talor Itzhak <titzhak@redhat.com> Signed-off-by: Swati Sehgal <swsehgal@redhat.com>
2021-05-13 10:55:33 +00:00
)
func main() {
flags := flag.NewFlagSet(ProgramName, flag.ExitOnError)
printVersion := flags.Bool("version", false, "Print version and exit.")
args, resourcemonitorArgs := parseArgs(flags, os.Args[1:]...)
if *printVersion {
fmt.Println(ProgramName, version.Get())
os.Exit(0)
}
// Assert that the version is known
if version.Undefined() {
klog.Warningf("version not set! Set -ldflags \"-X sigs.k8s.io/node-feature-discovery/pkg/version.version=`git describe --tags --dirty --always`\" during build or run.")
}
// Plug klog into grpc logging infrastructure
utils.ConfigureGrpcKlog()
u, err := url.ParseRequestURI(resourcemonitorArgs.KubeletConfigURI)
topologyupdater: Bootstrap nfd-topology-updater in NFD - This patch allows to expose Resource Hardware Topology information through CRDs in Node Feature Discovery. - In order to do this we introduce another software component called nfd-topology-updater in addition to the already existing software components nfd-master and nfd-worker. - nfd-master was enhanced to communicate with nfd-topology-updater over gRPC followed by creation of CRs corresponding to the nodes in the cluster exposing resource hardware topology information of that node. - Pin kubernetes dependency to one that include pod resource implementation - This code is responsible for obtaining hardware information from the system as well as pod resource information from the Pod Resource API in order to determine the allocatable resource information for each NUMA zone. This information along with Costs for NUMA zones (obtained by reading NUMA distances) is gathered by nfd-topology-updater running on all the nodes of the cluster and propagate NUMA zone costs to master in order to populate that information in the CRs corresponding to the nodes. - We use GHW facilities for obtaining system information like CPUs, topology, NUMA distances etc. - This also includes updates made to Makefile and Dockerfile and Manifests for deploying nfd-topology-updater. - This patch includes unit tests - As part of the Topology Aware Scheduling work, this patch captures the configured Topology manager scope in addition to the Topology manager policy. Based on the value of both attribues a single string will be populated to the CRD. The string value will be on of the following {SingleNUMANodeContainerLevel, SingleNUMANodePodLevel, BestEffort, Restricted, None} Co-Authored-by: Artyom Lukianov <alukiano@redhat.com> Co-Authored-by: Francesco Romani <fromani@redhat.com> Co-Authored-by: Talor Itzhak <titzhak@redhat.com> Signed-off-by: Swati Sehgal <swsehgal@redhat.com>
2021-05-13 10:55:33 +00:00
if err != nil {
klog.Exitf("failed to parse args for kubelet-config-uri: %v", err)
topologyupdater: Bootstrap nfd-topology-updater in NFD - This patch allows to expose Resource Hardware Topology information through CRDs in Node Feature Discovery. - In order to do this we introduce another software component called nfd-topology-updater in addition to the already existing software components nfd-master and nfd-worker. - nfd-master was enhanced to communicate with nfd-topology-updater over gRPC followed by creation of CRs corresponding to the nodes in the cluster exposing resource hardware topology information of that node. - Pin kubernetes dependency to one that include pod resource implementation - This code is responsible for obtaining hardware information from the system as well as pod resource information from the Pod Resource API in order to determine the allocatable resource information for each NUMA zone. This information along with Costs for NUMA zones (obtained by reading NUMA distances) is gathered by nfd-topology-updater running on all the nodes of the cluster and propagate NUMA zone costs to master in order to populate that information in the CRs corresponding to the nodes. - We use GHW facilities for obtaining system information like CPUs, topology, NUMA distances etc. - This also includes updates made to Makefile and Dockerfile and Manifests for deploying nfd-topology-updater. - This patch includes unit tests - As part of the Topology Aware Scheduling work, this patch captures the configured Topology manager scope in addition to the Topology manager policy. Based on the value of both attribues a single string will be populated to the CRD. The string value will be on of the following {SingleNUMANodeContainerLevel, SingleNUMANodePodLevel, BestEffort, Restricted, None} Co-Authored-by: Artyom Lukianov <alukiano@redhat.com> Co-Authored-by: Francesco Romani <fromani@redhat.com> Co-Authored-by: Talor Itzhak <titzhak@redhat.com> Signed-off-by: Swati Sehgal <swsehgal@redhat.com>
2021-05-13 10:55:33 +00:00
}
// init kubelet API client
var klConfig *kubeletconfigv1beta1.KubeletConfiguration
switch u.Scheme {
case "file":
klConfig, err = kubeconf.GetKubeletConfigFromLocalFile(u.Path)
if err != nil {
klog.Exitf("failed to read kubelet config: %v", err)
}
case "https":
restConfig, err := kubeconf.InsecureConfig(u.String(), resourcemonitorArgs.APIAuthTokenFile)
if err != nil {
klog.Exitf("failed to initialize rest config for kubelet config uri: %v", err)
}
klConfig, err = kubeconf.GetKubeletConfiguration(restConfig)
if err != nil {
klog.Exitf("failed to get kubelet config from configz endpoint: %v", err)
}
default:
klog.Exitf("unsupported URI scheme: %v", u.Scheme)
}
topologyupdater: Bootstrap nfd-topology-updater in NFD - This patch allows to expose Resource Hardware Topology information through CRDs in Node Feature Discovery. - In order to do this we introduce another software component called nfd-topology-updater in addition to the already existing software components nfd-master and nfd-worker. - nfd-master was enhanced to communicate with nfd-topology-updater over gRPC followed by creation of CRs corresponding to the nodes in the cluster exposing resource hardware topology information of that node. - Pin kubernetes dependency to one that include pod resource implementation - This code is responsible for obtaining hardware information from the system as well as pod resource information from the Pod Resource API in order to determine the allocatable resource information for each NUMA zone. This information along with Costs for NUMA zones (obtained by reading NUMA distances) is gathered by nfd-topology-updater running on all the nodes of the cluster and propagate NUMA zone costs to master in order to populate that information in the CRs corresponding to the nodes. - We use GHW facilities for obtaining system information like CPUs, topology, NUMA distances etc. - This also includes updates made to Makefile and Dockerfile and Manifests for deploying nfd-topology-updater. - This patch includes unit tests - As part of the Topology Aware Scheduling work, this patch captures the configured Topology manager scope in addition to the Topology manager policy. Based on the value of both attribues a single string will be populated to the CRD. The string value will be on of the following {SingleNUMANodeContainerLevel, SingleNUMANodePodLevel, BestEffort, Restricted, None} Co-Authored-by: Artyom Lukianov <alukiano@redhat.com> Co-Authored-by: Francesco Romani <fromani@redhat.com> Co-Authored-by: Talor Itzhak <titzhak@redhat.com> Signed-off-by: Swati Sehgal <swsehgal@redhat.com>
2021-05-13 10:55:33 +00:00
tmPolicy := string(topologypolicy.DetectTopologyPolicy(klConfig.TopologyManagerPolicy, klConfig.TopologyManagerScope))
klog.Infof("detected kubelet Topology Manager policy %q", tmPolicy)
// Get new TopologyUpdater instance
instance := topology.NewTopologyUpdater(*args, *resourcemonitorArgs, tmPolicy)
topologyupdater: Bootstrap nfd-topology-updater in NFD - This patch allows to expose Resource Hardware Topology information through CRDs in Node Feature Discovery. - In order to do this we introduce another software component called nfd-topology-updater in addition to the already existing software components nfd-master and nfd-worker. - nfd-master was enhanced to communicate with nfd-topology-updater over gRPC followed by creation of CRs corresponding to the nodes in the cluster exposing resource hardware topology information of that node. - Pin kubernetes dependency to one that include pod resource implementation - This code is responsible for obtaining hardware information from the system as well as pod resource information from the Pod Resource API in order to determine the allocatable resource information for each NUMA zone. This information along with Costs for NUMA zones (obtained by reading NUMA distances) is gathered by nfd-topology-updater running on all the nodes of the cluster and propagate NUMA zone costs to master in order to populate that information in the CRs corresponding to the nodes. - We use GHW facilities for obtaining system information like CPUs, topology, NUMA distances etc. - This also includes updates made to Makefile and Dockerfile and Manifests for deploying nfd-topology-updater. - This patch includes unit tests - As part of the Topology Aware Scheduling work, this patch captures the configured Topology manager scope in addition to the Topology manager policy. Based on the value of both attribues a single string will be populated to the CRD. The string value will be on of the following {SingleNUMANodeContainerLevel, SingleNUMANodePodLevel, BestEffort, Restricted, None} Co-Authored-by: Artyom Lukianov <alukiano@redhat.com> Co-Authored-by: Francesco Romani <fromani@redhat.com> Co-Authored-by: Talor Itzhak <titzhak@redhat.com> Signed-off-by: Swati Sehgal <swsehgal@redhat.com>
2021-05-13 10:55:33 +00:00
if err = instance.Run(); err != nil {
klog.Exit(err)
}
}
func parseArgs(flags *flag.FlagSet, osArgs ...string) (*topology.Args, *resourcemonitor.Args) {
args, resourcemonitorArgs := initFlags(flags)
_ = flags.Parse(osArgs)
if len(flags.Args()) > 0 {
fmt.Fprintf(flags.Output(), "unknown command line argument: %s\n", flags.Args()[0])
flags.Usage()
os.Exit(2)
}
if len(resourcemonitorArgs.KubeletConfigURI) == 0 {
if len(utils.NodeName()) == 0 {
fmt.Fprintf(flags.Output(), "unable to determine the default kubelet config endpoint 'https://${NODE_NAME}:%d/configz' due to empty NODE_NAME environment, "+
"please either define the NODE_NAME environment variable or specify endpoint with the -kubelet-config-uri flag\n", kubeletSecurePort)
os.Exit(1)
}
resourcemonitorArgs.KubeletConfigURI = fmt.Sprintf("https://%s:%d/configz", utils.NodeName(), kubeletSecurePort)
}
topologyupdater: Bootstrap nfd-topology-updater in NFD - This patch allows to expose Resource Hardware Topology information through CRDs in Node Feature Discovery. - In order to do this we introduce another software component called nfd-topology-updater in addition to the already existing software components nfd-master and nfd-worker. - nfd-master was enhanced to communicate with nfd-topology-updater over gRPC followed by creation of CRs corresponding to the nodes in the cluster exposing resource hardware topology information of that node. - Pin kubernetes dependency to one that include pod resource implementation - This code is responsible for obtaining hardware information from the system as well as pod resource information from the Pod Resource API in order to determine the allocatable resource information for each NUMA zone. This information along with Costs for NUMA zones (obtained by reading NUMA distances) is gathered by nfd-topology-updater running on all the nodes of the cluster and propagate NUMA zone costs to master in order to populate that information in the CRs corresponding to the nodes. - We use GHW facilities for obtaining system information like CPUs, topology, NUMA distances etc. - This also includes updates made to Makefile and Dockerfile and Manifests for deploying nfd-topology-updater. - This patch includes unit tests - As part of the Topology Aware Scheduling work, this patch captures the configured Topology manager scope in addition to the Topology manager policy. Based on the value of both attribues a single string will be populated to the CRD. The string value will be on of the following {SingleNUMANodeContainerLevel, SingleNUMANodePodLevel, BestEffort, Restricted, None} Co-Authored-by: Artyom Lukianov <alukiano@redhat.com> Co-Authored-by: Francesco Romani <fromani@redhat.com> Co-Authored-by: Talor Itzhak <titzhak@redhat.com> Signed-off-by: Swati Sehgal <swsehgal@redhat.com>
2021-05-13 10:55:33 +00:00
return args, resourcemonitorArgs
}
func initFlags(flagset *flag.FlagSet) (*topology.Args, *resourcemonitor.Args) {
args := &topology.Args{}
resourcemonitorArgs := &resourcemonitor.Args{}
flagset.BoolVar(&args.Oneshot, "oneshot", false,
"Update once and exit")
flagset.BoolVar(&args.NoPublish, "no-publish", false,
"Do not publish discovered features to the cluster-local Kubernetes API server.")
flagset.StringVar(&args.KubeConfigFile, "kubeconfig", "",
"Kube config file.")
topologyupdater: Bootstrap nfd-topology-updater in NFD - This patch allows to expose Resource Hardware Topology information through CRDs in Node Feature Discovery. - In order to do this we introduce another software component called nfd-topology-updater in addition to the already existing software components nfd-master and nfd-worker. - nfd-master was enhanced to communicate with nfd-topology-updater over gRPC followed by creation of CRs corresponding to the nodes in the cluster exposing resource hardware topology information of that node. - Pin kubernetes dependency to one that include pod resource implementation - This code is responsible for obtaining hardware information from the system as well as pod resource information from the Pod Resource API in order to determine the allocatable resource information for each NUMA zone. This information along with Costs for NUMA zones (obtained by reading NUMA distances) is gathered by nfd-topology-updater running on all the nodes of the cluster and propagate NUMA zone costs to master in order to populate that information in the CRs corresponding to the nodes. - We use GHW facilities for obtaining system information like CPUs, topology, NUMA distances etc. - This also includes updates made to Makefile and Dockerfile and Manifests for deploying nfd-topology-updater. - This patch includes unit tests - As part of the Topology Aware Scheduling work, this patch captures the configured Topology manager scope in addition to the Topology manager policy. Based on the value of both attribues a single string will be populated to the CRD. The string value will be on of the following {SingleNUMANodeContainerLevel, SingleNUMANodePodLevel, BestEffort, Restricted, None} Co-Authored-by: Artyom Lukianov <alukiano@redhat.com> Co-Authored-by: Francesco Romani <fromani@redhat.com> Co-Authored-by: Talor Itzhak <titzhak@redhat.com> Signed-off-by: Swati Sehgal <swsehgal@redhat.com>
2021-05-13 10:55:33 +00:00
flagset.DurationVar(&resourcemonitorArgs.SleepInterval, "sleep-interval", time.Duration(60)*time.Second,
"Time to sleep between CR updates. Non-positive value implies no CR updatation (i.e. infinite sleep). [Default: 60s]")
flagset.StringVar(&resourcemonitorArgs.Namespace, "watch-namespace", "*",
"Namespace to watch pods (for testing/debugging purpose). Use * for all namespaces.")
flagset.StringVar(&resourcemonitorArgs.KubeletConfigURI, "kubelet-config-uri", "",
"Kubelet config URI path. Default to kubelet configz endpoint.")
flagset.StringVar(&resourcemonitorArgs.APIAuthTokenFile, "api-auth-token-file", "/var/run/secrets/kubernetes.io/serviceaccount/token",
"API auth token file path. It is used to request kubelet configz endpoint, only takes effect when kubelet-config-uri is https. Default to /var/run/secrets/kubernetes.io/serviceaccount/token.")
flagset.StringVar(&resourcemonitorArgs.PodResourceSocketPath, "podresources-socket", hostpath.VarDir.Path("lib/kubelet/pod-resources/kubelet.sock"),
topologyupdater: Bootstrap nfd-topology-updater in NFD - This patch allows to expose Resource Hardware Topology information through CRDs in Node Feature Discovery. - In order to do this we introduce another software component called nfd-topology-updater in addition to the already existing software components nfd-master and nfd-worker. - nfd-master was enhanced to communicate with nfd-topology-updater over gRPC followed by creation of CRs corresponding to the nodes in the cluster exposing resource hardware topology information of that node. - Pin kubernetes dependency to one that include pod resource implementation - This code is responsible for obtaining hardware information from the system as well as pod resource information from the Pod Resource API in order to determine the allocatable resource information for each NUMA zone. This information along with Costs for NUMA zones (obtained by reading NUMA distances) is gathered by nfd-topology-updater running on all the nodes of the cluster and propagate NUMA zone costs to master in order to populate that information in the CRs corresponding to the nodes. - We use GHW facilities for obtaining system information like CPUs, topology, NUMA distances etc. - This also includes updates made to Makefile and Dockerfile and Manifests for deploying nfd-topology-updater. - This patch includes unit tests - As part of the Topology Aware Scheduling work, this patch captures the configured Topology manager scope in addition to the Topology manager policy. Based on the value of both attribues a single string will be populated to the CRD. The string value will be on of the following {SingleNUMANodeContainerLevel, SingleNUMANodePodLevel, BestEffort, Restricted, None} Co-Authored-by: Artyom Lukianov <alukiano@redhat.com> Co-Authored-by: Francesco Romani <fromani@redhat.com> Co-Authored-by: Talor Itzhak <titzhak@redhat.com> Signed-off-by: Swati Sehgal <swsehgal@redhat.com>
2021-05-13 10:55:33 +00:00
"Pod Resource Socket path to use.")
flagset.StringVar(&args.ConfigFile, "config", "/etc/kubernetes/node-feature-discovery/nfd-topology-updater.conf",
"Config file to use.")
topologyupdater: Bootstrap nfd-topology-updater in NFD - This patch allows to expose Resource Hardware Topology information through CRDs in Node Feature Discovery. - In order to do this we introduce another software component called nfd-topology-updater in addition to the already existing software components nfd-master and nfd-worker. - nfd-master was enhanced to communicate with nfd-topology-updater over gRPC followed by creation of CRs corresponding to the nodes in the cluster exposing resource hardware topology information of that node. - Pin kubernetes dependency to one that include pod resource implementation - This code is responsible for obtaining hardware information from the system as well as pod resource information from the Pod Resource API in order to determine the allocatable resource information for each NUMA zone. This information along with Costs for NUMA zones (obtained by reading NUMA distances) is gathered by nfd-topology-updater running on all the nodes of the cluster and propagate NUMA zone costs to master in order to populate that information in the CRs corresponding to the nodes. - We use GHW facilities for obtaining system information like CPUs, topology, NUMA distances etc. - This also includes updates made to Makefile and Dockerfile and Manifests for deploying nfd-topology-updater. - This patch includes unit tests - As part of the Topology Aware Scheduling work, this patch captures the configured Topology manager scope in addition to the Topology manager policy. Based on the value of both attribues a single string will be populated to the CRD. The string value will be on of the following {SingleNUMANodeContainerLevel, SingleNUMANodePodLevel, BestEffort, Restricted, None} Co-Authored-by: Artyom Lukianov <alukiano@redhat.com> Co-Authored-by: Francesco Romani <fromani@redhat.com> Co-Authored-by: Talor Itzhak <titzhak@redhat.com> Signed-off-by: Swati Sehgal <swsehgal@redhat.com>
2021-05-13 10:55:33 +00:00
klog.InitFlags(flagset)
topologyupdater: Bootstrap nfd-topology-updater in NFD - This patch allows to expose Resource Hardware Topology information through CRDs in Node Feature Discovery. - In order to do this we introduce another software component called nfd-topology-updater in addition to the already existing software components nfd-master and nfd-worker. - nfd-master was enhanced to communicate with nfd-topology-updater over gRPC followed by creation of CRs corresponding to the nodes in the cluster exposing resource hardware topology information of that node. - Pin kubernetes dependency to one that include pod resource implementation - This code is responsible for obtaining hardware information from the system as well as pod resource information from the Pod Resource API in order to determine the allocatable resource information for each NUMA zone. This information along with Costs for NUMA zones (obtained by reading NUMA distances) is gathered by nfd-topology-updater running on all the nodes of the cluster and propagate NUMA zone costs to master in order to populate that information in the CRs corresponding to the nodes. - We use GHW facilities for obtaining system information like CPUs, topology, NUMA distances etc. - This also includes updates made to Makefile and Dockerfile and Manifests for deploying nfd-topology-updater. - This patch includes unit tests - As part of the Topology Aware Scheduling work, this patch captures the configured Topology manager scope in addition to the Topology manager policy. Based on the value of both attribues a single string will be populated to the CRD. The string value will be on of the following {SingleNUMANodeContainerLevel, SingleNUMANodePodLevel, BestEffort, Restricted, None} Co-Authored-by: Artyom Lukianov <alukiano@redhat.com> Co-Authored-by: Francesco Romani <fromani@redhat.com> Co-Authored-by: Talor Itzhak <titzhak@redhat.com> Signed-off-by: Swati Sehgal <swsehgal@redhat.com>
2021-05-13 10:55:33 +00:00
return args, resourcemonitorArgs
}