I am trying to deploy Portworx Enterprise on and OpenShift 4.10 cluster on IBM Cloud, deployed through the IBM Cloud catalog tile. It is a brand new cluster, and are no issues on with the cluster, but Portworx fails to deploy successfully. As a result, Portworx is unusable on the cluster.
Portworx pods on the cluster never enter a healthy state, and Portworx never becomes available on the cluster, no matter how long you wait. There are errors in the logs and events on the pod, including:
- “insmod: WARNING: could not insert module /tmp/btrfs.all.el8.kos/x86_64/4.18.0-372.9.1.el8.x86_64/fs/btrfs/btrfs.ko: Unknown symbol in module. Check dmesg for more information.”
- “Failed to execute patch fs for kernel 4.18.0-372.26.1.el8_6.x86_64, exiting…”
- “Readiness probe failed: HTTP probe failed with statuscode: 503”
Steps to recreate:
- Create an OpenShift 4.10 VPC cluster. (Have tried both RHEL7 and RHEL8 workers, and the result is the same in both cases).
- Create a Portworx Enterprise service instance from the IBM Cloud Catalog, pointed at this cluster. Select the internal KVDM database option. I have tried with all 3 different versions of Portworx available in the IBM Cloud offering, and all end up with the same result (Portworx 2.9, 2.10, and 2.11).
- The Portworx service instance shown in the IBM Cloud resource list shows that it deployed successfully. However, the Portworx deployment & pods inside of the cluster never become healthy. All pods in the portworx and portworx-api DaemonSets never enter a healthy/ready state, and both events and logs show failures.
Is there a workaround that can be used to get Portworx healthy on the cluster?