This message was deleted.
# longhorn-storage
a
This message was deleted.
q
restarting the pod that uses the PVC took care of triggering the rebuild of the missing replica. I also don't seem to see anytmore the warnign for an unknown replica.
What I see on the host is that the actual device supporting the longhorn volume reports I/O superblock errors.
error is only for that specific longhorn volume (called sdh on the host).
So all is back to normal now and I need to investigate the root cause. It happened with another user/PV too.
f
This log typically means that there is something wrong with the replica or users manually delete the replica hence Longhorn will clean it up. This is a symptom rather than the cause of the replica failure. After replica failure or when the volume becomes degraded, Longhorn will automatically try to reuse the old replica. If it does not work, Longhorn will launch and rebuild new replica instead. Ideally you don’t need to handled failed replica cleanup and the following rebuilding at all.
103 Views