Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-38891

Add release note to 4.14 stating that each node during upgrade can have 300 ms of network downtime during network upgrade stage

XMLWordPrintable

    • Moderate
    • None
    • 2
    • T&PS 2024 #9, T&PS 2024 #10
    • 2
    • False
    • Hide

      None

      Show
      None
    • Hide
      There is a known issue when upgrading the Cluster Network Operator from version 4.13 to 4.14. The conversion to the new OVN Kubernetes interconnect multi-zone architecture can lead to packet drops that cause brief network outages. The impact is on East/West traffic in local gateway mode with `routingViaHost` set to `true`.
      Show
      There is a known issue when upgrading the Cluster Network Operator from version 4.13 to 4.14. The conversion to the new OVN Kubernetes interconnect multi-zone architecture can lead to packet drops that cause brief network outages. The impact is on East/West traffic in local gateway mode with `routingViaHost` set to `true`.
    • Bug Fix
    • In Progress
    • Hide
      Docs Status: 2024/09/30: peer review complete; needs change management
      Docs Status: 2024/09/09: Text proposal on QE review
      Show
      Docs Status: 2024/09/30: peer review complete; needs change management Docs Status: 2024/09/09: Text proposal on QE review

      Hi,

      We should add a release note to 4.14 stating that the pods on each node during the `network` upgrade stage will encounter up to 300ms of network outages. This happens in a rolling fashion node by node when the ovn-kube pods are restarted and migrate to interconnect multizone.
      The reason is the transition from non-interconnect to interconnect.

      Thanks,

      Andreas

            slovern@redhat.com Shane Lovern
            akaris@redhat.com Andreas Karis
            Zhanqi Zhao Zhanqi Zhao
            Votes:
            0 Vote for this issue
            Watchers:
            9 Start watching this issue

              Created:
              Updated:
              Resolved: