Multiple pods not starting

Products

• VMware Integrated OpenStack

Issue/Introduction

• Nova services are down.
• nova-compute logs are similar to:
  Jan 1 15:28:37 controller-######## nova-compute[885]: 2025-11-11 15:28:37.644 1 CRITICAL nova [req-########-####-####-####-############  - - - - -] Unhandled error: nova.exception.NovaException: Unable to find a valid datastore
• maraidb pods are in CrashLoopBackOff status.
• Describing the mariadb pods shows similar to:
  Warning FaiedAttachVolume 12s (x25 over 34m) attachdetach-controller AttachVolume Attach failed for volume *pvc-########-####-####-####-############* datastore <datastore name>' not found

 

Environment

7.3

Cause

One or more configured datastores have been renamed.  Openstack/vCenter can no longer find the configured datastore names.

Resolution

1. Restore the naming of these datastores back to what was configured.
2. Stop the VIO Deployment
   #viocli stop services
3. Check the controllers in vCenter and make sure only the base disk is attached.
4. If there is more than one disk attached, detach the other disks    Note: Do not delete the data, we just want to detach the disk(s)
5. Once all extra disks are detached from all controllers, start the deployment again.
   #viocli start services

Kubernetes will now start and attach the database, rabbitmq and vspc disks as it sees fit.  The deployment should now return to a running status.