Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-22221

DNS pods stuck in CrashLoopBackOff on upgrade of cluster using Hybrid Networking

XMLWordPrintable

    • Moderate
    • No
    • Sprint 244, Sprint 245, Sprint 246
    • 3
    • Rejected
    • False
    • Hide

      None

      Show
      None

      Description of problem:

      When upgrading clusters to any 4.13 version (from either 4.12 or 4.13), clusters with Hybrid Networking enabled appear to have a few DNS pods (not all) falling into CrashLoopBackOff status. Notably, pods are failing Readiness probes, and when deleted, work without incident. Nearly all other aspects of upgrade continue as expected

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

      4.13.x

      How reproducible:

      Always for systems in use

      Steps to Reproduce:

      1. Configure initial cluster installation with OVNKubernetes and enable Hybrid Networking on 4.12 or 4.13, e.g. 4.13.13
      2. Upgrade cluster to 4.13.z, e.g. 4.13.14
      

      Actual results:

      dns-default pods in CrashLoopBackOff status, failing Readiness probes

      Expected results:

      dns-default pods are rolled out without incident

      Additional info:

      Appears strongly related to OCPBUGS-13172. CU has kept an affected cluster with the DNS pod issue ongoing for additional investigating, if needed.

            jtanenba@redhat.com Jacob Tanenbaum
            rhn-support-jorbell Jordan Bell
            Jean Chen Jean Chen
            Votes:
            0 Vote for this issue
            Watchers:
            11 Start watching this issue

              Created:
              Updated:
              Resolved: