BMYPC5004

A pod PodDisruptionBudget (PDB) is set to zero.

Severity

Warning

Problem description

It indicates that the PodDisruptionBudget (PDB) of a pod is zero. During the upgrade, each node restarts in a rolling fashion. Any pod with a PDB set to zero blocks the node restarts because of the incomplete node drain operation. It might cause the upgrade to get stuck until you delete the pod manually. The upgrade operation restarts after the pod gets deleted and the node drains completely.

Recommended actions

Do the following steps to resolve the issue:
  1. To avoid this issue, delete the pod and set the PDB to zero.
  2. If you proceed with the upgrade without deleting the pod, the upgrade gets stuck during the node restart where the pod is running. Delete the pod whenever the IBM Storage Fusion user interface shows the upgrade as blocked.
  3. Select a method based on how disruptive the pod deletion is for its associated workloads. If you can tolerate a pod in unscheduled state during the upgrade, then pod deletion before the upgrade causes it to run unblocked.
  4. If you require active replicas of the pod throughout the upgrade, then you must manually manage the upgrade.
  5. Every time the upgrade gets blocked because a nod drain is in progress, delete the pod that causes the issue and allow it to reschedule on another node.

The upgrade restarts after the pod deletion and node drain operations are complete.