Closed chiraganand-e2e closed 2 years ago
Those are probably both problems in DRBD rather than LINSTOR itself. One appears to be DRBD becoming unresponsive during the resync, the other one is probably a consequence of that problem, causing DRBD not to send out events for changes of the resources' status.
Setup:
There was a failure on one of the linstor satellite nodes in our cluster so it had to be rebuilt from scratch (PV -> VG -> thinpool). When we brought the node back online Linstor realised that it has to create all resources on this node because there were none and the resource group had place-count as 3. All the resources were created (40 of them) but 5 got stuck at SyncTarget 97% or 98%. This information was given to us by
linstor resource list
but we looked atdrbdadm status
on the node itself these resources just showed as Inconsistent without any SyncTarget information.After restart, the satellite Linstor controller showed these resources as Inconsistent.
Eventually, we had to remove these 5 resources from this node and create them again.
I feel there are two bugs here: