Closed kdave closed 10 months ago
Orphan cleanup failures can be skipped at mount/remount, with a warning. The process is started again when a root path is open.
The device stats bdev /dev/sdd3 errs: wr 1299065141, rd 10234262, flush 6762148, corrupt 109308019, gen 0
show that something really bad happened so it's most likely not due to normal operations.
Before RW mount there are several other things done that must be done (relocation recovery, orphan root lookup, free space tree, uuid tree, ...), orphan cleanup is just one of them. If the real reason is ENOSPC this would fail one of the others anyway. Given the device stats, it's not clear if the ENOSPC situation is due to space management bug or because the filesystem git hit by a hw error, or something else.
Report from IRC, kernel 6.5.9, mount fails with ENOSPC and transaction abort