Node Feature Discovery up version. (r8, dsr8).
Added the following pages for the NFD up version installation. 1. Install Node Feature Discovery (NFD) StarlingX Application 2. Uninstall Node Feature Discovery (NFD) StarlingX Application 3. Removed the page "Run the node feature discovery helm chart" from the repo and index as well. Change-Id: I9beda664b01a0d8644cd7683f5045d6f275cf5cb Signed-off-by: Dinesh Neelapu <dinesh.neelapu@windriver.com>
This commit is contained in:
Dinesh Neelapu
parent
5a35c349cf
commit
46dace05d5
@ -364,13 +364,14 @@ Host hardware management
|
||||
|
||||
|
||||
-----------------------------------------
|
||||
Run the node feature discovery helm chart
|
||||
Node Feature Discovery |prod| Application
|
||||
-----------------------------------------
|
||||
|
||||
.. toctree::
|
||||
:maxdepth: 1
|
||||
|
||||
running-the-node-feature-discovery-helm-chart
|
||||
install-node-feature-discovery-nfd-starlingx-application-70f6f940bb4a
|
||||
uninstall-node-feature-discovery-nfd-starlingx-application-a582bfdadfb8
|
||||
|
||||
-------------
|
||||
Provision BMC
|
||||
|
||||
@ -0,0 +1,101 @@
|
||||
.. _install-node-feature-discovery-nfd-starlingx-application-70f6f940bb4a:
|
||||
|
||||
======================================================================
|
||||
Install Node Feature Discovery (NFD) |prod| Application
|
||||
======================================================================
|
||||
|
||||
Node Feature Discovery (NFD) detects hardware features available on each node
|
||||
in a kubernetes cluster and advertises those features using Kubernetes node
|
||||
labels. This procedure walks you through the process of installing the |NFD|
|
||||
|prod| Application.
|
||||
|
||||
.. rubric:: |prereq|
|
||||
|
||||
Before attempting to install the |NFD| |prod| application, ensure that your
|
||||
system meets the following requirements:
|
||||
|
||||
#. Ensure that you have Kubernetes version v1.24.4 installed. If not, follow
|
||||
the steps to upgrade the Kubernetes components. For detailed instructions on
|
||||
how to upgrade, see :ref:`manual-kubernetes-components-upgrade`.
|
||||
|
||||
|
||||
#. If you have previously installed the |NFD| Helm Chart, you need to uninstall
|
||||
it before proceeding with the new installation. To uninstall, use the
|
||||
following command:
|
||||
|
||||
.. code-block:: none
|
||||
|
||||
helm uninstall node-feature-discovery -n <namespace>
|
||||
|
||||
Replace <namespace> with the namespace where you have the |NFD| Helm Chart
|
||||
installed.
|
||||
|
||||
|
||||
.. rubric:: |proc|
|
||||
|
||||
Follow these steps to install the |NFD| |prod| Application:
|
||||
|
||||
#. Locate the |NFD| |prod| App tarball in the
|
||||
``/usr/local/share/applications/helm`` directory.
|
||||
|
||||
For example:
|
||||
|
||||
.. code-block:: none
|
||||
|
||||
/usr/local/share/applications/helm/node-feature-discovery-<version>.tgz
|
||||
|
||||
Replace <version> with the latest version number.
|
||||
|
||||
#. Upload the application using the following command.
|
||||
|
||||
.. code-block:: none
|
||||
|
||||
~(keystone_admin)]$ system application-upload /usr/local/share/applications/helm/node-feature-discovery-<version>.tgz
|
||||
|
||||
Replace <version> with the latest version number.
|
||||
|
||||
#. Verify that the |NFD| |prod| application has been uploaded successfully.
|
||||
|
||||
.. code-block:: none
|
||||
|
||||
~(keystone_admin)]$ system application-list
|
||||
|
||||
#. Apply the application using the following command.
|
||||
|
||||
.. code-block:: none
|
||||
|
||||
~(keystone_admin)]$ system application-apply node-feature-discovery
|
||||
|
||||
#. Monitor the status of the installation using one of the following commands.
|
||||
|
||||
.. code-block:: none
|
||||
|
||||
~(keystone_admin)]$ watch system application-list
|
||||
|
||||
OR
|
||||
|
||||
~(keystone_admin)]$ watch kubectl get pods -n node-feature-discovery
|
||||
|
||||
|
||||
.. rubric:: |result|
|
||||
|
||||
After a successful installation, the |NFD| |prod| deploys as the follows:
|
||||
|
||||
- A worker pod is created for each node in the Kubernetes cluster. These worker
|
||||
pods gather information about node features periodically and update the
|
||||
labels associated with the nodes based on the detected features.
|
||||
|
||||
- Additionally, there is one master pod that gathers all the updated data from
|
||||
worker pods and reporting it to Kubernetes for efficient feature management.
|
||||
|
||||
The |NFD| |prod| application is now installed on your system, and the hardware
|
||||
features available on each node in the Kubernetes cluster will be detected and
|
||||
advertised using Kubernetes node labels. You can check the labels using the
|
||||
following command for any configuration:
|
||||
|
||||
.. code-block:: none
|
||||
|
||||
kubectl label node/<node's name> --list | grep "feature.node.kubernetes.io"
|
||||
|
||||
Replace <node's name> with the specific node name such as controller-0,
|
||||
controller-1, and so on.
|
||||
@ -1,86 +0,0 @@
|
||||
|
||||
.. dea1561393939185
|
||||
.. _running-the-node-feature-discovery-helm-chart:
|
||||
|
||||
=========================================
|
||||
Run the Node Feature Discovery Helm Chart
|
||||
=========================================
|
||||
|
||||
Node feature discovery detects hardware features available on each node in a
|
||||
Kubernetes cluster, and advertises those features using Kubernetes node
|
||||
labels.
|
||||
|
||||
An **stx-platform** helm repository containing platform related charts,
|
||||
including the **node-feature-discovery** helm chart, is created
|
||||
automatically once the system has been installed.
|
||||
|
||||
.. rubric:: |proc|
|
||||
|
||||
#. Confirm that the **stx-platform** is available.
|
||||
|
||||
.. code-block:: none
|
||||
|
||||
~(keystone_admin)$ helm repo list
|
||||
NAME URL
|
||||
stable https://kubernetes-charts.storage.googleapis.com
|
||||
local http://127.0.0.1:8879/charts
|
||||
starlingx http://127.0.0.1:8080/helm_charts/starlingx
|
||||
stx-platform http://127.0.0.1:8080/helm_charts/stx-platform
|
||||
|
||||
The repository is created automatically after system installation has
|
||||
completed. There may be a delay of a few minutes after a new install
|
||||
before it is available.
|
||||
|
||||
#. Create a helm override file to specify custom values for any of the
|
||||
following:
|
||||
|
||||
.. table::
|
||||
:widths: auto
|
||||
|
||||
+--------------------+--------------------------------------------------------------------------------------------------------------------+----------------------------------------------------------------------------------------------------------------------+
|
||||
| Override Name | Description | Default |
|
||||
+====================+====================================================================================================================+======================================================================================================================+
|
||||
| image: | The docker image to use for the pods. | k8s.gcr.io/nfd/node-feature-discovery |
|
||||
| repository: | | |
|
||||
+--------------------+--------------------------------------------------------------------------------------------------------------------+----------------------------------------------------------------------------------------------------------------------+
|
||||
| worker: | The Kubernetes nodeSelector to use for the worker pods. | All nodes. |
|
||||
| nodeSelector: | This is a key/value pair to match against node labels and select which nodes should run the node feature discovery.| Changing the {} to {foo: bar} will result in the worker pods only running on Kubernetes nodes labelled with foo:bar. |
|
||||
| {} | | |
|
||||
+--------------------+--------------------------------------------------------------------------------------------------------------------+----------------------------------------------------------------------------------------------------------------------+
|
||||
| fullnameOverride: | The base name to use for the Kubernetes resources (pods, daemonsets, etc.). | A blank entry here will result in the base name being "node-feature-discovery". |
|
||||
+--------------------+--------------------------------------------------------------------------------------------------------------------+----------------------------------------------------------------------------------------------------------------------+
|
||||
|
||||
For example, to set the base name to "nfd" and to run the worker pod only
|
||||
on nodes with the "foo: bar" label:
|
||||
|
||||
.. code-block:: none
|
||||
|
||||
~(keystone_admin)$ cat <<EOF > /home/sysadmin/my-discovery-overrides.yaml
|
||||
fullnameOverride: nfd
|
||||
worker:
|
||||
nodeSelector:
|
||||
foo: bar
|
||||
EOF
|
||||
|
||||
#. Run the node feature discover helm chart. It is recommended to run it in
|
||||
the "kube-system" namespace so the pods run on the platform CPUs.
|
||||
|
||||
- To accept the default values, run :command:`helm` without specifying an
|
||||
overrides file:
|
||||
|
||||
.. code-block:: none
|
||||
|
||||
~(keystone_admin)$ helm -n kube-system upgrade -i node-feature-discovery stx-platform/node-feature-discovery
|
||||
|
||||
- If you have defined default overrides in a yaml file, specify the
|
||||
file when running :command:`helm`.
|
||||
|
||||
For example:
|
||||
|
||||
.. code-block:: none
|
||||
|
||||
~(keystone_admin)$ helm -n kube-system upgrade -i node-feature-discovery stx-platform/node-feature-discovery --values=/home/sysadmin/my-discovery-overrides.yaml
|
||||
|
||||
One worker pod is created per node, which runs once per minute to update
|
||||
the node features. One master pod gathers up all the data and reports it
|
||||
to Kubernetes.
|
||||
@ -0,0 +1,30 @@
|
||||
.. _uninstall-node-feature-discovery-nfd-starlingx-application-a582bfdadfb8:
|
||||
|
||||
======================================================================
|
||||
Uninstall Node Feature Discovery (NFD) |prod| Application
|
||||
======================================================================
|
||||
|
||||
Follow these steps to remove the Node Feature Discovery (NFD) |prod|
|
||||
Application.
|
||||
|
||||
.. rubric:: |proc|
|
||||
|
||||
#. Remove |NFD| pods and resources using the following command:
|
||||
|
||||
.. code-block:: none
|
||||
|
||||
~(keystone_admin)]$ system application-remove node-feature-discovery
|
||||
|
||||
|
||||
#. Remove |NFD| helm chart and application using the following command:
|
||||
|
||||
.. code-block:: none
|
||||
|
||||
~(keystone_admin)]$ system application-delete node-feature-discovery
|
||||
|
||||
|
||||
After completion of the uninstallation process, the |NFD| application will be
|
||||
completely removed from your system. All related pods, resources, and
|
||||
configurations associated with |NFD| will no longer be present, ensuring a
|
||||
clean removal of the application.
|
||||
|
||||
@ -89,6 +89,7 @@
|
||||
.. |NA| replace:: :abbr:`NA (Not Applicable)`
|
||||
.. |NAT| replace:: :abbr:`NAT (Network Address Translation)`
|
||||
.. |NF| replace:: :abbr:`NF (Network Function)`
|
||||
.. |NFD| replace:: :abbr:`NFD (Node Feature Discovery)`
|
||||
.. |NFs| replace:: :abbr:`NFs (Network Functions)`
|
||||
.. |NFS| replace:: :abbr:`NFS (Network File System)`
|
||||
.. |NIC| replace:: :abbr:`NIC (Network Interface Card)`
|
||||
|
||||
Loading…
Reference in New Issue
Block a user