Right, then it's apparent why it doesn't want to mount - the check which fails the mount was not present until rpm-4.4.63-1. It's only a consistency check. However, the more pressing question the volume got in a situation where device size doesn't equal to what is recorded in the superblock. I guess there are 2 courses of action: 1) Revert the offending patch, but it's supposed to harden the the kernel against such "corruption" ( the quotes are due to my not being sure that it's actually a corruption) 2) If you have enough freespace on a different drive, try to recreate the filesystem via btrfs send/receive. Obviously this might work in your case but we don't know how widespread this bug could be, it seems so far no one has complained. David, Jeff what do you make out of this ?