On 28/12/16 16:32, sdm wrote:
"Yes, unfortunately btrfs RAID5/6 still suffers from the write hole (10/2015). The one missing piece, from a reliability point of view, is that it is still vulnerable to the parity RAID "write hole", where a partial write as a result of a power failure may result in inconsistent parity data." - http://superuser.com/questions/701111/is-btrfs-vulnerable-to-raid-write-hole...
Write hole is behaviour mdadm can't get around and it does the best it can considering the implementation. btrfs and write atomicity still appears to be a work in progress. OP's problem most likely isn't that btrfs is the culprit. Sounds more like mdadm and an unclean shutdown (power loss). If he can't afford a hardware RAID controller (A real one, not a $30 one) and ECC memory, then if it were me, I would opt for btrfs built-in software RAID 1 over mdadm RAID 1+btrfs.
This will - hopefully - soon be a fixed problem as far as mdadm/raid is concerned. Dunno which kernel it will end up in but there's a new feature being on its way - "PPL" or "Partial Parity Log". If switched on, it writes a log for a stripe being updated so it knows the stripe didn't write successfully, and it'll recover it. I don't understand the details, you'll have to read the linux-raid list for that. Cheers, Wol -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org