3

I'm attempting to perform a backup to an external hard drive. Both the native drive and external drive use a BTRFS filesystem. I do this by first creating a read-only snapshot of my home directory:

# btrfs subvolume snapshot -r /home /.snapshots/home/BACKUP

Followed by a send/receive to the backup drive:

# btrfs send /.snapshots/home/BACKUP | btrfs receive /.backup/home

I get this output:

At subvol /.snapshots/home/BACKUP
At subvol BACKUP
ERROR: failed to clone extents to matt/.local/share/Anki2/Matt/collection.anki2: Invalid argument

where the error occurs after a minute or so. I tried with several different snapshots and got the same error on each. Moreover, I couldn't find that file in any of the snapshots. I thought this might be some sort of disk corruption error, so I booted into a USB OS image and ran

# btrfs check --repair /dev/mapper/...

on the unmounted filesystem. This reported no errors. After booting back into the filesystem, I tried the backup again and got the same error. Any thoughts on what could be causing this?

EDIT

As per Emmanuel Rosa's comment, I tried

# btrfs scrub start -B /

which exited without errors. The issue is still present after the scrub.

EDIT 2

I wasn't able to figure out what was wrong with this. I ended up just deleting the local file and then it worked again.

EDIT 3

While deleting files temporarily suspends the issue, new files are often created that create the same error, so this really isn't resolved. sqlite files seem to be particularly problematic.

Improve this question
4
  • Oh, boy. btrfs check --repair is the LAST thing one should run. See man btrfs-check. btrfs scrub may be useful. Check man btrfs-scrub to see if it fits your use case. Commented Apr 12, 2020 at 12:51
  • Thanks for the heads up about btrfs-check repair and for pointint out scrub. Unfortunately, scrub didn't find any errors and I got the same error. Commented Apr 12, 2020 at 17:21
  • Do any of the btrfs subvolume {find-new,list,show} commands show any surprises for your subvolumes? Commented Apr 14, 2020 at 13:41
  • As far as I can tell those all look normal. Commented Apr 14, 2020 at 15:48

1 Answer 1

Reset to default
1

This was apparently a bug in the Linux kernel, which was fixed by this commit. I discovered this by posting a question on the btrfs mailing list. Upgrading the kernel resolved the issue.

Improve this answer
1
  • The relevant kernel version is 5.5.3, but from the mailing list thread it sounds like there are some other circumstances that can also lead to this error (e.g. setting snapshots read-write after they are received, and then using them as parents for other receives) Commented Sep 26, 2024 at 12:05

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.