Ask Your Question

Btrfs errors on boot, can't boot anymore

asked 2018-02-16 21:52:31 -0600

horizonbrave gravatar image

updated 2018-02-16 22:13:11 -0600

Hi Fedora 27 user here with btrfs partitioning on a ssd (default scheme choose at installation). All of a sudden my laptop can't boot, tried all kernel entries in GRUB. If I true the rescue we entry it says: "Cannot open access to console, the root account is locked."

The errors on the screen are: bad tree block start fail to open runtime journal (systemd) error removing orphan entry, stopping orphan cleanup pending csums is 16994304 remounting read-write after error is not allowed

Not sure what could it have trigger it, did a couple of cold reboots though recently (machine freezed). Many thanks for suggestions

PS please see attached screenshot here

edit retag flag offensive close merge delete

1 Answer

Sort by ยป oldest newest most voted

answered 2018-02-17 00:38:19 -0600

cmurf gravatar image

updated 2018-02-17 00:43:09 -0600

These look like hardware problems. /etc/crypttab has IO error, journal has IO error, the orphan cleanup error also is IO failure. Orphans are normal on forced power offs during writes but normally get cleaned up at mount time, but that can't happen here because of IO errors. Depending on the hardware a bunch of writes might have happened in the wrong order, and all of this might actually not be hardware failure but rather a (drive) firmware bug. Pending csums is really high, that's data extent checksums which is how btrfs knows if data being read in is correct.

Absolutely for sure do not use 'btrfs check --repair' a.k.a. btrfsck. Doing it without --repair is fine, which will just spit out a bunch of errors but not try to fix anything. If you have data you want to recover from the drive, if backups aren't current, you need to avoid changing the file system. I recommend posting on the upstream btrfs list for help with data recovery, it's a bit involved and the tools are not exactly intuitive.

Order of things to do:

  1. Try to get it to mount, even with -o ro, to try and get data you want that isn't backed up off the drive.
  2. Get some information about the setup, drive make/model, firmware version, kernel and btrfs-progs versions, and btrfs check output (without --repair) and some history including mount options used and if the fstrim time is enabled, and kernel messages to see what might have gone wrong. This is best done on the btrfs list rather than back and forth on ask.fedoraproject.
  3. Scrape user data off if normal, or ro mount attempts fail. List can help
  4. Update drive firmware if applicable (don't do it before getting data off).
  5. Wipe the btrfs partition and reinstall Fedora. (the errors you're reporting are almost certainly catastrophic and not worth fixing - just get the data off and then start over).

Btrfs list is at find linux-btrfs in the list, click it, then click on subscribe link and follow the instructions. It's possible to post without subscribing but not everyone will do reply to all so...I suggest subscribing then unsubscribe later.

edit flag offensive delete link more

Question Tools

1 follower


Asked: 2018-02-16 21:52:31 -0600

Seen: 363 times

Last updated: Feb 17 '18