I had a RAID battery die in out of our nodes, and replaced it but the cache got corrupt. I put the node into storage repair and a ghost node was created for the cluster. I fixed the RAID on the node with the corrupt cache, added it back into the mgmt group and the cluster. All the volumes re-striped but the are still in a unprotected state and the ghost node is still there.
I saw in another form that I should of exchanged the node instead of readding the node to the cluster.
So what would be the best approach for this, would it be remove the node I fixed from the cluster and try to do the exchange, or can I just delete the ghost node with the RIP MAC. My other fear of removing the ghost node with the RIP MAC would that cause any issues with the repaired node with the same MAC?
All the nodes are running 9.5 and there are multiple clusters under the mgmt group.
Just in case anyone every comes across this, I removed the ghost node out of the cluster, and the volumes re-striped. I left the ghost node in the mgmt group and when the re-striping was done the ghost node went away.
Also I was told I could of removed the repaired node which would cause the volumes to re-strip, then once the re-striping was finised I could of exchaned the repaired node with the ghost node then let the volumes re-stripe again.
I went for the first option since our nodes are old I did not want to put them through two re-stripes.
