Hi again Chris, Le 11/03/2020 à 00:12, Chris Murphy a écrit : > > Maybe some extent tree corruption. I'm thinking that the older kernel > without an extensive tree checker won't care, and eventually things > may get worse and unworkable again. But it might work for a while? > What kernel version was it at the time of the problem? I see btrfs-progs 4.15.1 Actually I checked and the kernel itself was 5.3, only the tools are 4.15 - Yes that's Mint, you can install a newer kernel but the toolchain is the distro's... The same goes for Ubuntu or Debian... But a 5.3 kernel is actually pretty recent. >> You could change fstab to include these two mount options: > > flushoncommit,notreelog > > The first probably won't slow things down, might help improve > consistency if there's a crash; where the second will likely make > certain things slower because it turns off fsync optimization and > requires full sync each time. Uh... The kid's got an old computer and I cannot consider anything that would slow things down more than they already are... As I could, thanks to you, mount the FS, I rsync'd it completely to another one and saw that only 2 files from the python3 package and a couple dozens from a flatpak cache were corrupt or missing. No big deal. So I reformatted the original disk with its original uuid, recreated the subvols, restored everything back into place, put the disk in another laptop and it booted like a charm. I then just reinstalled python3, ran system updates and voilà. Now I only need to drop the SSD back into the little girl's ole PC. Thanks again, without your kind assistance I would have been doomed to a full reinstall and the kid would had lost her files - being a kid's machine, no regular backups... However a FS that badly dies is an issue anyway. Not for a rant, but I've considered BRFS to be « slow but extremely reliable » for at least 6 years. And I certainly wouldn't have expected it to die because of a system crash or power fail. With time it became faster BUT in the past year I've already completely lost 4 BTRFS filesystems on different machines, most from the 5.2 kernel bugs and I don't consider this acceptable from an FS anybody would deem “reliable”. This plus the ugly “zero free space” bug in 5.3... I actually wonder what's going on with BTRFS development... Kind regards and thanks again. ॐ -- Swâmi Petaramesh <swami@xxxxxxxxxxxxxx> PGP 9076E32E
