I use a combination of LVM thin volume snapshots and btrfs snapshots. While i could do fine without btrfs snapshots i like to have the ability to browse old ver

I use a combination of LVM thin volume snapshots and btrfs snapshots. While i could do fine without btrfs snapshots i like to have the ability to browse old versions online in order to diff or restore them. The thin volume pool is mirrored using lvm's rather new raid1 volume types (not the mirror type). If the filesystem fails (which happened twice already) i can merge the working LVM snapshot back and loose at max a day worth of work. However if the LVM fails i have an offline copy with the same setup (LVM and btrfs) where i use btrfs send/receive functionality. But if this happens i will loose some days of work. If both LVM VGs fail im lost. The offline copy also holds more btrfs snapshots than the online one in order to reduce disk usage on the online storage. I do it all manually but i am planning to do the btrfs snapshots automatically. However, the LVM snapshots wont ever happen automatically as i don't want to snapshot a bad filesystem and i am unsure how to reliably detect filesystem errors without using the system.