6

I have a btrfs file system that cannot be mounted anymore. I resetted the computer and after boot a Java process crashed, showing a core dump with much btrfs stuff on the stack. After that happened, the computer was basically unusable, as starting processes didn't work anymore.

So since the computer itself was unusable, I decided to "rescue" the file system with a Ubuntu 14.04 Desktop disc.

I tried btrfs rescure chunk-recover, and it asked me if the chunk tree should be rebuild. I answered yes. Now, it seems like it is completely broken. The FS cannot be mounted anymore, and now the btrfs check output looks like this:

root@ubuntu:/dev# btrfs check /dev/mapper/encrypted 
checksum verify failed on 20971520 found B89CA074 wanted EDB30E17
checksum verify failed on 20971520 found B89CA074 wanted EDB30E17
checksum verify failed on 20971520 found B89CA074 wanted EDB30E17
checksum verify failed on 20971520 found B89CA074 wanted EDB30E17
Csum didn't match
Couldn't read chunk root
Couldn't open file system

Trying to mount gives this error:

[90746.734393] btrfs: device fsid ee55dbb6-e359-42cc-bf76-9973ed5cf4ff devid 1 transid 6036 /dev/dm-0
[90746.844751] btrfs: device fsid ee55dbb6-e359-42cc-bf76-9973ed5cf4ff devid 1 transid 6036 /dev/dm-0
[91441.004757] btrfs: device fsid ee55dbb6-e359-42cc-bf76-9973ed5cf4ff devid 1 transid 6036 /dev/mapper/encrypted
[91441.006465] btrfs: disk space caching is enabled
[91441.008592] btrfs: dm-0 checksum verify failed on 20971520 wanted EDB30E17 found B89CA074 level 0
[91441.009034] btrfs: dm-0 checksum verify failed on 20971520 wanted EDB30E17 found B89CA074 level 0
[91441.009065] btrfs: failed to read chunk root on dm-0
[91441.025106] btrfs: open_ctree failed

I tried btrfs check with --repair, --init-csum-tree and --init-extent-tree, but it changed nothing.

4
  • Asking the obvious: Hard drive defective? If you have a bad drive, no wonder there are checksum errors. If so, replacing that drive should probably be the first thing to do.
    – basic6
    Jul 8, 2015 at 19:04
  • As of now the question is one year old, and the HDD is still working fine.
    – Sibbo
    Jul 9, 2015 at 7:39
  • So no smart errors, relevant smart values look ok and badblocks read-only test found no error? Or has nothing been checked and you're assuming it still works fine because it still runs? Bad drives usually start corrupting data long before they die. On the other hand, maybe your drive is healthy, but does btrfs dev stats / list any errors and what does btrfs fi df / say?
    – basic6
    Jul 9, 2015 at 10:54
  • As a last resort, btrfs rescue chunk-recover may do something if it's not just the log that's corrupted. Jul 29, 2022 at 21:29

1 Answer 1

4

Your checksums are not matching. That's not good.

If you haven't fixed this yet, you can try upgrading your btrfs-tools and using btrfs check --fix-crc and it should resolve your checksum issues. You might want to take a btrfs-image of your metadata first.

2

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .