On Fri, Feb 15, 2019 at 2:31 PM Martin Pöhlmann <mpdeimos@xxxxxxxxx> wrote: OK I just looked at the dmesg you attached, and you might have found a bug. 1. Try -o ro,nologreplay because it's safe, and you can update your backups if needed. 2. If ro,nologreplay doesn't work, use 'btrfs restore' to extract important files. The point here is to be prepared to lose the file system before it's actually lost. 3. Try to mount normally (no mount options) with a different kernel. If it were me I would try in order: 5.0.0-rc6 4.19.23 4.14.101 4. If none of those kernels will mount it normally, wait if you can. But once you need to move on you can try to zero the log as described and then try a normal mount. If mount still fails but with different messages, report back. It might be possible once zero log is run, to use a backup root, or do a check using a specific root tree. But if you're at this point, it's close to check --repair time and that can sometimes make things worse. -- Chris Murphy
