After a botched ZFS rpool mirror-1 replacement it somehow converted itself to a full ZFS raid-0 on that same pool. I've tried different ways in fixing it but at the end it now makes sense just move the VMs and replication off of that node and rebuild it using new node name and IPs.
I've read the instructions on how to properly remove the node from the cluster but the part that bothers me is that I should power off the node before I use the remove node command on the other node. Since the effected node it still can boot up properly and may cause problems in the cluster after the node is removed.
So I'm wondering would it make more sense to go ahead and blow away that node and reinstall using new node name and different IPs so it'll produce a node down in the cluster and I can then safely use the remove node command?
I'm doing this entirely via iDrac9 and things can happen during reboot when it fail to boot into my virtual boot media. Trying to make sure it doesn't happen.
I've read the instructions on how to properly remove the node from the cluster but the part that bothers me is that I should power off the node before I use the remove node command on the other node. Since the effected node it still can boot up properly and may cause problems in the cluster after the node is removed.
So I'm wondering would it make more sense to go ahead and blow away that node and reinstall using new node name and different IPs so it'll produce a node down in the cluster and I can then safely use the remove node command?
I'm doing this entirely via iDrac9 and things can happen during reboot when it fail to boot into my virtual boot media. Trying to make sure it doesn't happen.
Last edited:
