I’ve been running bazzite for a while now but after upgrading to the 202041104 build my file system completely locks up after I start running an ostree upgrade. No matter if I use ujust update or go directly into rpm-ostree upgrade, I always get this message
── 22:14:10 - System update ────────────────────────────────────────────────────
Pulling manifest: ostree-image-signed:docker://ghcr.io/ublue-os/bazzite:stable
Checking out tree 4b95952... done
error: Removing extensions/rpmostree/private/commit: unlinkat(LoopPassManager.h): Das Dateisystem ist nur lesbar
System update failed:
0: Command failed: `/usr/bin/rpm-ostree upgrade`
1: `/usr/bin/rpm-ostree` failed: exit status: 1
Location:
src/steps/os/linux.rs:228
Retry? (y)es/(N)o/(s)hell/(q)uit
And then I can’t do much of anything anymore, even my home directory is apparently not writable anymore. I basically get error messages from whatever program I want to run that it can’t write to its files.
I’m right now not on the 20241104 build anymore but the one before, thinking that would solve it, but it doesn’t. It has the same problem. I also went back to a build with F40 that I had pinned from a while back, same issue.
I don’t really know what might be wrong, can anyone help?
Thanks a lot!
This is most likely silent file system corruption, (TLDR:) you probably either need to reinstall or get a new SSD/HardDrive. Try booting into a Live CD, you can get one at the Fedora Workstation download page. Download Fedora Media Writer, Select Workstation and write it to a flash drive. Boot to it and check your filesystem with sudo btrfs check -p /dev/blah1 where blah1 is your disk (you can find it with the lsblk command.), depending on the result, try reinstalling your system, if that happens again, make sure to try installing your system in a different storage medium (another SSD, HDD, whatever else you are using.)
I tried different things and identified the corrupted file (it’s part of an older deployment) but even messing with it and googling around there doesn’t seem to be a way to fix it or get rid of it( rm can’t delete it and it can’t be overwritten or moved). At least it doesn’t appear to be a hardware issue.
What I tried was
Boot into a live Fedora WS environment as suggested
use sudo btrfs check -p <drive>. The log told me there are broken inode references for a file
Tried all kinds of things to delete that file, nothing worked
Tried sudo btrfs check --repair <drive partition> since I read that is a last resort. The command gets stuck in a loop though, apparently it can’t repair it either.
Seems all I have left is to install fresh, it’s what other people had to do that I read about with similar issues. I’ll do that when I have time, for now I still have a running system since the corruption is not in the current deployment.
You probably shouldnt try to fix it through any corruption fixing tool that BTRFS has, as that will probably just worsen the problem Yeah, reinstalling is really the best thing you can do right now. Good luck, hoping it works better from now on!