Offline Replica Rebuilding

Starting with v1.9.0, SUSE Storage supports offline replica rebuilding. This feature allows degraded volumes to automatically rebuild replicas while the volumes are detached.

Global Setting offline-replica-rebuilding

  • When enabled, SUSE Storage automatically initiates offline rebuilding for eligible volumes.

  • For more information about this global setting, see settings.

Per-Volume Override

  • You can override the global offline-replica-rebuilding setting for each volume individually. This can be done through the SUSE Storage UI or by editing the Volume Custom Resource. To use kubectl, run the following command and then modify the spec.offlineRebuilding field:

    kubectl -n longhorn-system edit volume <volume-name>

    Replace <volume-name> with the name of your specific volume.

  • When the per-volume spec.offlineRebuilding field is set to enabled or disabled, this setting takes precedence over the global configuration. The default value for spec.offlineRebuilding is ignored.

The following table shows how the global and per-volume settings interact:

Global Setting (offline-replica-rebuilding)

Per-Volume Setting (spec.offlineRebuilding)

Offline Rebuilding Enabled

true

ignored

Yes

false

ignored

No

true

enabled

Yes

false

enabled

Yes

true

disabled

No

false

disabled

No

Rebuilding Process

  • When offline replica rebuilding is triggered, SUSE Storage attaches the volume without activating the frontend, rebuilds any missing replicas, and then detaches the volume after the rebuilding process is complete.

  • This process can be interrupted if the associated workload scales up and requires the volume.

Limitations

This feature does not apply to faulted volumes.