Uploaded image for project: 'OpenShift Virtualization'
  1. OpenShift Virtualization
  2. CNV-79408

HCO not handling controlPlaneTopology: HighlyAvailableArbiter

XMLWordPrintable

    • CNV I/U Operators Sprint 285
    • None

      Description of problem:

      When CNV is installed on a Two Node +Arbiter cluster configuration, it is not deploying the proper number of virt-api and virt-controller pods. It would appear that the HCO is not properly handling the "controlPlaneTopology: HighlyAvailableArbiter" from "Infrastructure cluster" resource.

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

      4.20.3

      How reproducible:

      Every time

      Steps to Reproduce:

      1. Build a Two Node plus Arbiter cluster
2. Install OpenShift Virt using all defaults

      Actual results:

      Take note of the number of "virt-api" and "virt-controller" pods that are deployed by the HCO. You will see that only 1 virt-api pod and 1 virt-controller pod are created. 
You will also receive two alerts as a part of monitoring that will say that the number of virt-api and virt-controllers is wrong

      Expected results:

      I would expect that because there are two nodes in the cluster that can run workloads, two virt-api and virt-controller pods would be deployed since this is a "HighlyAvailable" cluster.

      Additional info:

      See https://redhat-internal.slack.com/archives/C068X44C8VB/p1770733245892159 for a larger discussion of this bug.

              ocohen@redhat.com Oren Cohen
              mdeneve@redhat.com Mark DeNeve
              Ramón Lobillo Ramón Lobillo
              Votes:
              0 Vote for this issue
              Watchers:
              5 Start watching this issue

                Created:
                Updated: