Skip to main content

PVC Mount Error

Introduction​

In some cases, we encounter an error where a pod cannot mount its PVC on a node. This document describes the runbook for troubleshooting and resolving this issue.

Step 1: Scale Down the Workload​

Scale the replicas of the Deployment or StatefulSet that references the affected PVC to zero.
In other words, make sure all related pods are stopped.

Tip

Before scaling down, note the node name where the affected pod was scheduled.

Delete the following pods that are running on the same node as the affected pod:

unklog -> logging-system
node-metrics-exporter -> monitoring-system
multus -> kube-system
  • Delete the unklog pod related to that node (namespace: logging-system)
  • Delete the node-metrics-exporter pod related to that node (namespace: monitoring-system)
  • Delete the multus pod related to that node (namespace: kube-system)

Step 3: Scale Up and Verify​

Scale up the affected Deployment/StatefulSet again.

Then:

  1. Describe the new pod.
  2. Confirm the PVC mount error is resolved.