I’m currently in the process of migrating an older bcachefs pool (version 16). It fscks without issues on 3c39b422acd3346321185be0ce263809e2a9a23f, corresponding to the kernel it was last mounted on. Now I am running another fsck on 5ef62f56ab50c5799f713e3a42f5c7ad7e8283d3 and it reports a lot of fixable errors (such as bad val size (48 < 56)), which are automatically applied without -p flag, which I assume are due to the FS upgrade.
However at some point it stops execution entirely, with the most recent message being:
btree_node_read_work: rewriting btree node at btree=alloc level=0 3:2107798:0 due to error
I’m currently in the process of migrating an older bcachefs pool (version 16). It fscks without issues on 3c39b422acd3346321185be0ce263809e2a9a23f, corresponding to the kernel it was last mounted on. Now I am running another fsck on 5ef62f56ab50c5799f713e3a42f5c7ad7e8283d3 and it reports a lot of fixable errors (such as bad val size (48 < 56)), which are automatically applied without -p flag, which I assume are due to the FS upgrade.
However at some point it stops execution entirely, with the most recent message being:
More of the log can be seen here: https://thermalpaste.technogothic.net/p/Mpcy02.txt
checking in GDB, it appears that https://github.com/koverstreet/bcachefs-tools/blob/5ef62f56ab50c5799f713e3a42f5c7ad7e8283d3/linux/sched.c#L41-L42 is trying to wait for the current task to no longer be running, which never occurs (likely because the task is waiting for itself to stop running?
output of
bt full
: