2017-07-16 18:06 GMT+02:00 Marc MERLIN <m...@merlins.org>: > On Sun, Jul 16, 2017 at 04:01:53PM +0200, Giuseppe Della Bianca wrote: >> > On Fri, Jul 14, 2017 at 06:22:16PM -0700, Marc MERLIN wrote: >> > > Dear Chris and other developers, >> ]zac[ >> > Others on this thread with the same error: did anyone recover from this >> > without wiping the filesystem? >> > >> > Is there a chance a balance might work around the bug so that whatever >> > layout I have, changes, and stops the bug from occuring? >> ]zac[ >> >> Any attempt, even just delete files, has worsened the situation. >> I advise not to waste time in repairs, and directly recreate the filesystem. > > I see. So, this is a condition where the filesystem is clear as far as: > - check > - check lowmem > - scrub > are all concerned (at least in my case), but it's in a state where > touching something around a sensitive area causes the bug. > If so, this blows, and I'm not really wanting to recreate a clean 12TB > filesystem "just because", especially since this could just happen > again after I've rebuilt it. >
IMHO, rebuild from scratch, 1-2 times a year, the snapshot receive filesystem is inevitable. For this reason, my snapshot receive filesystems have only this purpose and are not bigger than 1-2 TB. >> while read proc; do >> if [ $progResult == 0 ]; then >> echo -e \nbtrfs tools already running >> >> progResult=222 >> fi >> >> echo $proc" >> done < <(ps -ef | grep -e "btrfs >> \{1,\}\(subvolume\|send\|receive\|delete\)") > > Yeah, I probably hit that. I think you can also add scrub to that list. > Yes. I did not add scrubs because my scrub are always read-only. And I think that race condition is between snapshot receive and subvolume delete. I also suggest: - Use btrfs subvolume delete with -c - Try to add a sleep after subvolume delete and receive. > Marc Gdb -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html