With more specialised hardware, you can make use of xfs file system and build a raidz (hdd partial replication). So if a hdd fail, you can replace it with a new one and it will restore data. All this, without service interruption. Also with xfs you unleach the power of snapshot. So when you have a big change to make in a VM/CT, you can make a snapshot before, so if something go wrong, you can rollback to this snapshot. Be aware that it can be dangerous to use snapshot with lightning nodes, because you could publish old channel state and undergo a punishment transaction. A cluster of server can be use not only to backup but also for high availability and hardware failover. You choose where your VM\CT run and where it failover.