Description of problem:

After updating the NumaResourcesOperator CR to expand the set of included MCPs no additional daemonsets were created and the nodes in the additional MCPs were not represented with noderesourcetopologies CRs.

The state of the cluster when the operator CR was updated:

• 4 custom mcps for worker nodes.
• All worker mcps were paused
• All had pending updates, ie the MCP was Updated=False

When the MCPs were unpaused and updated the daemonsets were then created and noderesourcetopology crs created.

Same seems to hold true when removing an MCP from the NumaResourcesOperator config. If they are paused and not updated the daemonsets do not get removed until all of the MCPs are unpaused and ready (updated=true). They do get scaled down (but not deleted) when the MCP is unpaused.

Version-Release number of selected component (if applicable):

nrop 4.18

How reproducible:

one out of one attempts

Steps to Reproduce:

spec:
  logLevel: Normal
  nodeGroups:
  - config:
      infoRefreshMode: Periodic
    machineConfigPoolSelector:
      matchLabels:
        machineconfiguration.openshift.io/role: worker-1
  - config:
      infoRefreshMode: Periodic
    machineConfigPoolSelector:
      matchLabels:
        machineconfiguration.openshift.io/role: worker-2
  - config:
      infoRefreshMode: Periodic
    machineConfigPoolSelector:
      matchLabels:
        machineconfiguration.openshift.io/role: worker-3
  - config:
      infoRefreshMode: Periodic
    machineConfigPoolSelector:
      matchLabels:
        machineconfiguration.openshift.io/role: worker-4

Actual results:

Expected results:

Additional info: