On 2022-08-23 03:37, Song Liu wrote:
Thomas, have you tried to bisect with the fio repro?
Yes, just finished:
d32d3d0b47f7e34560ae3c55ddfcf68694813501 is the first bad commit commit d32d3d0b47f7e34560ae3c55ddfcf68694813501 Author: Christoph Hellwig Date: Mon Jun 14 13:17:34 2021 +0200
nvme-multipath: set QUEUE_FLAG_NOWAIT The nvme multipathing code just dispatches bios to one of the blk-mq based paths and never blocks on its own, so set QUEUE_FLAG_NOWAIT to support REQ_NOWAIT bios.
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?i...
So another NOWAIT issue -- similar to the bad commit which is causing the mdraid issue I already found (https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?i...).
Reverting the commit, i.e. deleting
blk_queue_flag_set(QUEUE_FLAG_NOWAIT, head->disk->queue);
fixes the problem for me. Well, sort of. Looks like this will disable io_uring. fio reproducer fails with
$ fio reproducer.fio filename0: (g=0): rw=randwrite, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T) 4096B-4096B, ioengine=io_uring, iodepth=1 fio-3.30 Starting 1 thread fio: io_u error on file /srv/machines/fio/filename0.0.0: Operation not supported: write offset=12648448, buflen=4096 fio: pid=1585, err=95/file:io_u.c:1846, func=io_u error, error=Operation not supported
My MariaDB reproducer also doesn't trigger the problem anymore, but probably for the same reason -- it cannot use io_uring anymore.