CloudCasa pods get evicted during restore

Summary

A PVC restore may fail with a message similar to the following, in job activity:

“Failed to copy PVC test-data in namespace test-ns. Error: failed to wait for mover configmap update. Error is wait for ConfigMap update was interrupted. Error is Mover pod status is not RUNNING and TimeoutExpired. Status is Pending”

Around the same time, you may see this message (or similar) in Kubernetes events for cloudcasa-io namespace:

“The node was low on resource: ephemeral-storage. Container cloudcasa-io-kubemover was using 1252Ki, which exceeds its request of 0”

Solution

This problem occurs due to insufficient ephemeral-storage space that is used for PVC data caching during restore.

To resolve the issue, ensure that the root filesystem of each Kubernetes node has at least 20 GB of free space.