logical volume manager storage operator version 4.18 is incorrectly installed on OpenShift V4.17 or older clusters

Solution Verified - Updated

Environment

  • Red Hat OpenShift Container Platform (RHOCP) <= 4.18
  • lvm-storage operator (LVMS) 4.18

Issue

Resolution

Resolution depends on the RHOCP Version:

  • RHOCP 4.16 and 4.17:

    • lvms V4.18 on RHCOP 4.16 and 4.17 is now fully supported
    • switch the subscription to the stable-4.18 channel to ensure to get future Z stream updates for CVE and bug fixes of lvms 4.18 on OCP clusters version 4.16 and 4.17.

  • RHOCP < 4.16: Version is not affected

    • The upgrade failed with error

      InstallComponentFailed:error validating existing CRs against new CRD's schema for "lvmclusters.lvm.topolvm.io": error validating lvm.topolvm.io/v1alpha1, Kind=LVMCluster "openshift-storage/lvmcluster": updated validation is too restrictive: [].status.deviceClassStatuses[0].nodeStatus[0].deviceDiscoveryPolicy: Required value

    • Procedure to remove the failed installation attempt of the V4.18 lvms operator from a OCP <=4.14 cluster. Note: lvm-storage will continue to run through the procedure, remaining fully functional.
      IMPORTANT: Do NOT remove LVMCluster CR

      1. Identify failed 4.18 InstallPlan and delete it. It's in the namespace the operator is installed in. Example:

        oc delete installplan/install-zdlsp -n openshift-lvm-storage

      2. Delete ClusterServiceVersion objects for the failed V4.18 version and the correct 4.14 version. The latter is in state “Replacing” and thus needs to be removed. Example:

        oc delete csv/lvms-operator.v4.18.4 -n openshift-lvm-storage
oc delete csv/lvms-operator.v4.14.4 -n openshift-lvm-storage

      3. Delete the operator Subscription. IMPORTANT: Do NOT delete dependent objects when offered using the Console! Example:

        oc delete subscription/lvms-operator -n openshift-lvm-storage

      4. Re-Install the operator as per the documentation using the 4.14 channel. Be sure to install into the same (existing) namespace.

Root Cause

  • A repository management error led to the accidental publication of v4.18 redhat-operators catalog content into legacy tags (v4.12 - v4.16). This caused a catalog-to-platform version mismatch.

  • For clusters with automatic OLM updates enabled, the OLM interpreted the v4.18 content as a valid "recommended update" path. Consequently, operators were upgraded to versions incompatible with the underlying OpenShift version, leading to a failure state that persisted even after the catalog was corrected, Reference OCP Bug: This content is not included.OCPBUGS-76865.

SBR
Product(s)
Components
Category

This solution is part of Red Hat’s fast-track publication program, providing a huge library of solutions that Red Hat engineers have created while supporting our customers. To give you the knowledge you need the instant it becomes available, these articles may be presented in a raw and unedited form.