koverstreet
Unable to correct errors with fsck due to 'interrupted system call' + superblock not marked as containing replicas journal
Hello there! I've recently encountered an annoying bug, so hopefully this is the right place for it. I'm using bcachefs tool version v0.1-357-g6c42566 on Arch Linux, with a 5.13 kernel with bcachefs patches from https://github.com/Frogging-Family/linux-tkg/.
So, onto the bug: My system will lock up fairly frequently, and I've been able to reproduce it reliably by attempting to delete or move a certain folder on my hard drive. After restarting, remounting read-only, and running fsck, I was only able to fix one of the errors; the other two consistently gave me "Interrupted system call" errors and returned a -4. ~Unfortunately, I don't know yet how to get output from a read-only system here, but I'll try and see if I can get it on a USB or something, if you think that'll be helpful.~ Got it!
Output available here: fsck_output.txt
On a possibly related issue, I got this error from trying to have bcachefs list my drives: EDIT: maybe related to #110?
$ sudo bcachefs list /dev/nvme0n1p1 /dev/sda3
superblock not marked as containing replicas journal: 1/1 [1], exiting
Unable to continue, halting
Error in recovery: cannot allocate memory (1)
filesystem contains errors: please report this to the developers
mount with -o fix_errors to repair
error opening /dev/nvme0n1p1: Input/output error
I did try mounting with -o fix_errors, but it didn't seem to do anything.
Thanks for your hard work and I apologize for the info gaps. This is my first issue here, so, if you could, please let me know what information to provide and how to get it. Thanks again!
Also, is it a known thing that systemd-fsck treats bcachefs partitions as ext2? I have some output here:
$ /usr/lib/systemd/systemd-fsck /dev/nvme0n1p1
/dev/nvme0n1p1 is mounted.
WARNING!!! The filesystem is mounted. If you continue you ***WILL***
cause ***SEVERE*** filesystem damage.
Do you really want to continue<n>? yes
fsck.ext2: Permission denied while trying to open /dev/nvme0n1p1
You must have r/w access to the filesystem or be root
fsck failed with exit status 8.
Ignoring error.
$ /usr/lib/systemd/systemd-fsck /dev/sda3
fsck.ext2: Permission denied while trying to open /dev/sda3
You must have r/w access to the filesystem or be root
fsck failed with exit status 8.
Ignoring error.
Also, putting them both together doesn't seem to work, but I think this one's a known issue(https://github.com/systemd/systemd/issues/8234)
$ /usr/lib/systemd/systemd-fsck /dev/nvme0n1p1:/dev/sda3
Failed to stat /dev/nvme0n1p1:/dev/sda3: No such file or directory
Thanks as always!
So, I have an update. After trying some stuff (that I can't really remember), fsck no longer shows the individual errors, but unfortunately states that the disk still has errors (that I guess can't be corrected). Right now, the filesystem has significant delays, as well as hard/soft lockups, so I might have to reformat the disk soon. Any ideas, @koverstreet?
Here's the latest fsck output for reference: fsck_output_2.txt