Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-41555

SingleReplica HCPs can not upgrade on cluster with nodes in a single zone

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Done-Errata
    • Icon: Normal Normal
    • 4.16.z
    • 4.15, 4.16, 4.17.0, 4.18.0
    • HyperShift
    • Moderate
    • No
    • False
    • Hide

      None

      Show
      None
    • Hide
      * Previously, if the Hosted Cluster (HC) `controllerAvailabilityPolicy` value was `SingleReplica`, networking components with `podAntiAffinity` would block the rollout. With this release, the issue is resolved. (link:https://issues.redhat.com/browse/OCPBUGS-41555[*OCPBUGS-41555*])
      -----------
      Fixes an issue where podAntiAffinity on networking components would block rollout if the Hosted Cluster controllerAvailabilityPolicy is SingleReplica
      Show
      * Previously, if the Hosted Cluster (HC) `controllerAvailabilityPolicy` value was `SingleReplica`, networking components with `podAntiAffinity` would block the rollout. With this release, the issue is resolved. (link: https://issues.redhat.com/browse/OCPBUGS-41555 [* OCPBUGS-41555 *]) ----------- Fixes an issue where podAntiAffinity on networking components would block rollout if the Hosted Cluster controllerAvailabilityPolicy is SingleReplica
    • Bug Fix
    • In Progress

      This is a clone of issue OCPBUGS-39313. The following is the description of the original issue:

      This is a clone of issue OCPBUGS-29497. The following is the description of the original issue:

      While updating an HC with controllerAvailabilityPolicy of SingleReplica, the HCP doesn't fully rollout with 3 pod stuck in Pending

      multus-admission-controller-5b5c95684b-v5qgd          0/2     Pending   0               4m36s
      network-node-identity-7b54d84df4-dxx27                0/3     Pending   0               4m12s
      ovnkube-control-plane-647ffb5f4d-hk6fg                0/3     Pending   0               4m21s
      

      This is because these deployment all have requiredDuringSchedulingIgnoredDuringExecution zone anti-affinity and maxUnavailable: 25% (i.e. 1)

      Thus the old pod blocks scheduling of the new pod.

              cewong@redhat.com Cesar Wong
              openshift-crt-jira-prow OpenShift Prow Bot
              Feilian Xie Feilian Xie
              Votes:
              0 Vote for this issue
              Watchers:
              4 Start watching this issue

                Created:
                Updated:
                Resolved: