Unable to stop the Unified Agent that is deployed on a managed cluster

Use the instructions provided in this guide to help diagnose the issue with stopping the Unified Agent that is deployed on a managed cluster.

Problem

You deployed the Unified Agent in a managed cluster on OpenShift Container Platform. Afterwards, you chose to disable the Unified Agent plug-ins. However, data is still being sent to the target endpoint.

Cause

The problem is caused by a configuration issue with an OpenShift Container Platform's NodeSelector. A NodeSelector that is configured incorrectly can block the rolling-update mechanism. A default NodeSelector for the managed cluster is defined in /etc/origin/master/master-config.yaml. For example:

projectConfig:
  defaultNodeSelector: node-role.kubernetes.io/compute=true

Resolving the problem

To solve the issue, you must overwrite the default NodeSelector：

1. Edit the Unified Agent project namespace.

   # oc edit namespace <project_name>
   

   Where <project_name> is the project where you installed the Unified Agent. For example: cp4mcm-cloud-native-monitoring.
2. Find the annotations section and add the following annotation:

   openshift.io/node-selector: ""
   

   For more information about setting NodeSelectors, see Deploying Applications to specific Nodes.