Created attachment 1829193 [details]
must-gather from a cluster that failed to resume

Description of problem:

We utilize Hive clusterpools to drive our CI/CD process - similarly to their use in OSCI.  While clusters are in a pool - their nodes are shut down and the cluster is "hibernated" to save cost.  When the cluster is needed - it is brought back online by hive following the documented process - https://docs.openshift.com/container-platform/4.8/backup_and_restore/graceful-cluster-restart.html.  We've observed greatly increased resume failures on OCP 4.8 when the cluster has been offline for more than 24 hours.  The Console and Auth fail to come back online nearly 100% of the time.  


Version-Release number of selected component (if applicable): 4.8.Z


How reproducible: 80-100% occurance


Steps to Reproduce:
1.Hibernate/Power off a clusters nodes
2.Leave for 24-48 hours
3.Follow the procedure to gracefully restart a cluster

Actual results:
Cluster Console and Auth are not reachable.  Cluster is reachable via CLI with Kubeconfig only.

Expected results:
Cluster resumes successfully.

Additional info:
Must-gather attached.