mirrors/node-feature-discovery
1
0
Fork 0
mirror of https://github.com/kubernetes-sigs/node-feature-discovery.git synced 2024-12-14 11:57:51 +00:00
Code Activity
Node feature discovery for Kubernetes
2 commits 13 branches 62 tags 147 MiB
Go 95.7%
Shell 2.4%
Makefile 1.1%
Smarty 0.4%
Dockerfile 0.2%
Other 0.1%
Find a file
Balaji Subramaniam d96f50e7c6 1. Enabled RDT discovery. 
2. Updated the ds to a job for node labelling.
3. Added script to run a job in each node.
4. Updated Readme.
2016-07-12 02:30:49 +02:00
rdt-discovery 1. Enabled RDT discovery. 2016-07-12 02:30:49 +02:00
Dockerfile 1. Enabled RDT discovery. 2016-07-12 02:30:49 +02:00
featurelabeling-job.json.template 1. Enabled RDT discovery. 2016-07-12 02:30:49 +02:00
label-nodes.sh 1. Enabled RDT discovery. 2016-07-12 02:30:49 +02:00
LICENSE 1. Enabled RDT discovery. 2016-07-12 02:30:49 +02:00
main.go 1. Enabled RDT discovery. 2016-07-12 02:30:49 +02:00
README.md 1. Enabled RDT discovery. 2016-07-12 02:30:49 +02:00

README.md

#CPU Feature Discovery for Kubernetes This software enables Intel Architecture (IA) feature discovery for Kubernetes. It detects the CPU features, such as Intel Resource Director Technology (RDT), available in each node of the Kubernetes cluster and labels them appropriately.

##Getting Started

System Requirements

At a minimum, you will need:

  1. [Docker] docker-down (only required to build and push docker images)
  2. [GCC] gcc-down (only required to build software to detect Intel RDT feature set)
  3. Linux (x86_64)
  4. [kubectl] kubectl-setup (properly setup and configured to work with a Kubernetes cluster)

###Installation ####Downloading the Source Code

git clone https://github.com/intelsdi-x/dbi-iafeature-discovery <repo-name>

####Building from Source The build steps described below are optional. The default docker image in Docker hub at intelsdi/nodelabels can be used to decorate the Kubernetes node with IA features. Skip to usage instructions if you do not need to build your own docker image.

Build the Intel RDT Detection Software Using make (Optional):

cd <rep-name>/rdt-discovery
make

Build the Docker Image (Optional):

cd <repo-name>
docker build -t <user-name>/<image-name> .

Push the Docker Image (Optional)

docker push

Change the Job Spec to Use the Appropriate Image (Optional): Change line #40 in featurelabeling-job.json.template to the appropriate image name in the <user-name>/<image-name> format used for building the image above or use the default intelsdi/nodelabels image.

###Usage Deploying Kubernetes Job for Node Feature Labeling:

./label-nodes.sh

The above command will label each node in the cluster with the features from cpuid. The labeling format used is

<key, val> = <"node.alpha.intel.com/<feature-name>", "true">

Note that only features that are available in that node are labelled.

##Documentation This software determines the Intel Architecture (IA) features associated with a processor using cpuid. The following set of features are discovered and labelled by this software.

###Intel Resource Director Technology (RDT) Features

Label Feature(s) Represented
RDTMON Intel Cache Monitoring Technology (CMT) and Intel Memory Bandwidth Monitoring (MBM)
RDTL3CA Intel L3 Cache Allocation Technology
RDTL2CA Intel L2 Cache Allocation Technology

###Other Features (Partial List)

Label Feature(s) Represented
ADX Multi-Precision Add-Carry Instruction Extensions (ADX)
AESNI Advanced Encryption Standard (AES) New Instructions (AES-NI)
AVX Advanced Vector Extensions (AVX)
AVX2 Advanced Vector Extensions 2 (AVX2)
BMI1 Bit Manipulation Instruction Set 1 (BMI)
BMI2 Bit Manipulation Instruction Set 2 (BMI2)
SSE4.1 Streaming SIMD Extensions 4.1 (SSE4.1)
SSE4.2 Streaming SIMD Extensions 4.2 (SSE4.2)
SGX Software Guard Extensions (SGX)

##License This is an Open Source software released under the Apache 2.0 License.