Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-30245

[4.13] Replace nodelister with master nodelister everywhere

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Done-Errata
    • Icon: Undefined Undefined
    • 4.14.z
    • 4.13, 4.12, 4.14, 4.15, 4.16
    • Etcd
    • None
    • No
    • 1
    • False
    • Hide

      None

      Show
      None

      This is a clone of issue OCPBUGS-29717. The following is the description of the original issue:

      Description of problem:

      From profiling on cert rotation we know that the node informer is
      called every couple of seconds on node heartbeats. This PR will ensure
      that all our node listers only ever listen/inform on the master node
      updates to reduce the frequency of unnecessary sync calls.
      
      Also related to the issue, increasing the amount of node status updates:
      OCPBUGS-29713
      OCPBUGS-29424    

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

      4.16 down to 4.12, we need to check all versions    

      How reproducible:

      always

      Steps to Reproduce:

          1. create a cluster
          2. look at some metric (eg sum(rate(apiserver_request_total{resource="nodes"}[5m]))))
          3. observe some improvement over previous state
          

      Actual results:

      increased amount of CPU usage for CEO / QPS to apiserver

      Expected results:

      less amount of CPU consumed for CEO / QPS to apiserver   

      Additional info:

      already fixed in 4.16 with
      https://github.com/openshift/cluster-etcd-operator/pull/1205
      
      creating this ticket for backporting     

              dwest@redhat.com Dean West
              openshift-crt-jira-prow OpenShift Prow Bot
              Ge Liu Ge Liu
              Votes:
              0 Vote for this issue
              Watchers:
              5 Start watching this issue

                Created:
                Updated:
                Resolved: