On Thu, Apr 4, 2019 at 12:18 PM Zirconium Hacker <jared.e...@gmail.com> wrote: > > Yes, I have qgroups enabled on that filesystem. I'll disable qgroups > the next chance I get, they aren't worth the trouble. > > After rebooting, BTRFS finished the balance. I then foolishly started > a selective metadata balance, which never got past the first block > group; it just thrashed the disk for several hours. > It could not be canceled, so I rebooted. > > Then I mounted the filesystem with skip_balance and clear_cache. > The mount failed. The relevant part of dmesg is attached.
I can't parse that, maybe Qu has an idea. In the meantime I suggest btrfs-progs 4.20.2 and running: # btrfs check --readonly /dev/ Attach that output. Balance is supposed to be COW. And skip_balance and clear_cache are supposed to be safe. So I'd say you've found a bug but I don't know what it is or how to work around it. -- Chris Murphy
