Uploaded image for project: 'AMQ Broker'
  1. AMQ Broker
  2. ENTMQBR-6187

[Operator] Operator fails to automatically upgrade the broker pod

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Done
    • Icon: Minor Minor
    • AMQ 7.10.0.OPR.3.GA
    • AMQ 7.9.4.OPR.1.GA, AMQ 7.9.4.OPR.2.GA, AMQ 7.9.4.OPR.3.GA, AMQ 7.9.4.OPR.4.GA
    • operator
    • None
    • False
    • False
    • User Experience
    • Hide

      The custom resource reconciliation must be triggered and the simplest way to do so is to scale up the deployment - change the deployment size from 1 to 2 and oc apply the custom resource or do the same via editing the custorm resource instance's yaml via OpenShift's web console.

      Show
      The custom resource reconciliation must be triggered and the simplest way to do so is to scale up the deployment - change the deployment size from 1 to 2 and oc apply the custom resource or do the same via editing the custorm resource instance's yaml via OpenShift's web console.
    • Hide
      1. Install operator v7.9.2-opr-1 using operator hub (OLM) into single namespace with support of automatic upgrades
      2. Deploy broker_activemqartemis_cr.yaml custom resource with single pod and upgrades enabled
      3. Wait for the broker pod to start and become ready
      4. Check the logs to verify 7.9.2 version is deployed
      5. Trigger the operator upgrade to v7.9.3-opr-1
      6. Wait few minutes if operator upgrades the broker pod.
      Show
      Install operator v7.9.2-opr-1 using operator hub (OLM) into single namespace with support of automatic upgrades Deploy broker_activemqartemis_cr.yaml custom resource with single pod and upgrades enabled Wait for the broker pod to start and become ready Check the logs to verify 7.9.2 version is deployed Trigger the operator upgrade to v7.9.3-opr-1 Wait few minutes if operator upgrades the broker pod.

      there is a bug in reconciliation process. While testing bundles for 7.9 I followed the usual steps: Deployed older (released) index image, installed operator for single namespace with automatic upgrades, thrown in artemis CR with single pod to make sure everything gets upgraded when index image is updated. Operator upgraded from 7.9.2 to 7.9.3 but somehow didn't pick up on the updates for the broker image it self. Pod got stuck on 7.2.9 version. I tried to scale to two pods and operator picked up on the changes ... Deployed second pod as 7.9.3 and then upgraded the first one ... So bundles are fine (info is correct) but there is a miner bug in operator itself ... not sure if it's known issue .
      In 7.8.4 -> to 7.8.5 upgrades this wasn't an issue ...

        1. broker_activemqartemis_cr.yaml
          1 kB
        2. operator-Initial-inciden.log
          58 kB
        3. operator-reproduced-state.log
          5 kB

            gtully@redhat.com Gary Tully
            rvais Roman Vais
            Roman Vais Roman Vais
            Votes:
            0 Vote for this issue
            Watchers:
            5 Start watching this issue

              Created:
              Updated:
              Resolved: