Just a short follow-up.
I've been way too busy to do any more testing. Servers had to go to the client and time was tight, so we skipped ZFS and replication. VM-s are running on HW RAID + LVM. All is good, although we don't have replication. We're doing backups every 2 hours from one node to the other. I haven't tested if that causes slowdowns after restore, to be honest, but that gave us more confidence and that0's what we went for.
In the end this remained unresolved.
First chance I get to test this issue on another pair of servers I will check it again and post back here... But it could be months before that happens.
Thanks @LnxBil and @spirit for joining with insights !
I've been way too busy to do any more testing. Servers had to go to the client and time was tight, so we skipped ZFS and replication. VM-s are running on HW RAID + LVM. All is good, although we don't have replication. We're doing backups every 2 hours from one node to the other. I haven't tested if that causes slowdowns after restore, to be honest, but that gave us more confidence and that0's what we went for.
In the end this remained unresolved.
First chance I get to test this issue on another pair of servers I will check it again and post back here... But it could be months before that happens.
Thanks @LnxBil and @spirit for joining with insights !