Applicable versions

• All Longhorn v1.7.x versions
• Longhorn v1.8.x versions prior to v1.8.3
• Longhorn v1.9.0

Symptoms

When both the storage network and the storage-network-for-rwx-volume-enabled setting are enabled, restarting the CSI plugin pod causes workloads using migratable RWX volumes to restart unexpectedly.

Root Cause

This issue is due to a bug in the storage network support for RWX volumes. Since Longhorn RWX volumes are not true NFS-based volumes, restarting the CSI plugin should not trigger a workload restart.

Workaround

This bug is resolved in:

• Longhorn v1.8.3 and later
• Longhorn v1.9.2 and later

If you are using an affected version, consider the following workarounds:

1. Disable the storage-network-for-rwx-volume-enabled setting to revert RWX volumes to use the cluster network.
2. Disable the auto-delete-pod-when-volume-detached-unexpectedly setting.

   Warning Be sure to review the limitation of this approach before applying it.

Related information

• GitHub Issue: #1158