I have a 3-node cluster with HA set up. On one node, I have a Raid-Z with a NAS container serving my media. That ZFS subvolume is a mountpoint; the root volume is in Ceph.
I shut down the hardware for maintenance, and it migrated the NAS container to one of the other nodes, where it was immediately broken because its storage isn't available. (Which is why I set migration to 0 in the HA settings.)
Now I can't start the container, and I can't move it back, because migration fails saying that the storage isn't available on the source node. (The one that never should have been migrated to in the first place, because guess what, the storage isn't available there.)
If I try to detach the volume to move the VM so I can then reattach it, it threatens to delete everything in it.
How can I force this container to migrate back to where it's supposed to be?
And what is the right way for me to have set it up so that it can never move away from its storage? (One would expect any such migration to fail, so this should be a no-brainer, no?)
I shut down the hardware for maintenance, and it migrated the NAS container to one of the other nodes, where it was immediately broken because its storage isn't available. (Which is why I set migration to 0 in the HA settings.)
Now I can't start the container, and I can't move it back, because migration fails saying that the storage isn't available on the source node. (The one that never should have been migrated to in the first place, because guess what, the storage isn't available there.)
If I try to detach the volume to move the VM so I can then reattach it, it threatens to delete everything in it.
How can I force this container to migrate back to where it's supposed to be?
And what is the right way for me to have set it up so that it can never move away from its storage? (One would expect any such migration to fail, so this should be a no-brainer, no?)