Alexey Korepanov

fsck showed some unreachable inodes, I don't know if this is related. Both hard drives report healthy with 0 errors in SMART. ``` bcachefs (60a8973f-529a-439b-88e8-1e41a166893e): check_unreachable_inodes...unreachable inode: inum: 628425:4294967295 mode=100644...

I've been on bcachefs since before it was merged in the kernel, but I stopped on 6.12 because this is the current long term support version, supposedly more stable than...

If data corruption is not a super critical thing, then I wonder what is? (no disrespect, I absolutely admire your work on bcachefs)

> I'm seeing filesystem errors, but where's the data corruption? I'm not sure if/how the data corruption shows in system logs, but I observe this: > After the copy processes...

6.12 has still quite some time before EOL, if we mean the same L 

After I upgraded from 6.12 to 6.14 and mounted the filesystem with fsck option, I saw a lot of `missing backpointer`, `dirent points to inode that does not point back:`,...

I still see the errors on 6.14.9, although this time without messages in system log. I use zfs now because it has `zpool scrub` which shows any disk errors better...

I've been trying to check whether my hardware is faulty... I ran the same test without bcachefs, temporarily using my 128GB swap partition instead of an image file (but on...

I have been trying to create a minimal reproducer for this corruption in a virtual machine, but so far without luck. Yet according to the video and https://lkml.org/lkml/2025/6/4/1186 Kent may...

Uh, I named the issue "Data corruption...", I thought that was clear. I can't easily try your master branch yet: I'm limited to kernels