Hello Song,
OK, I will add a cover letter with more descriptions & resend these patches.
Though the test scripts almost same, there are two different bugs.
patch 1/2 fixes --grow wrong behaviour. (the bug happened after second --grow cmd executing)
patch 2/2 fixes md-cluster deadlock. (the deadlock happened before second --grow cmd)
The patch 1/2 bug was came from one of SUSE customers. When I finished bugfix and ran test script to verify,
I triggered patch 2/2 bug.
test script of patch 2/2 adds "--bitmap-chunk=1M" in creating mdadm & mkfs.xfs after setup array.
These two steps make array to do more resync work. More resync time give lager time window (more opportunities)
to trigger deadlock.
Thanks.
On 11/7/20 8:17 AM, Song Liu wrote:
> On Thu, Nov 5, 2020 at 5:11 AM Zhao Heming <heming.zhao@xxxxxxxx> wrote:
>>
>> Test script (reproducible steps):
>> ```
>> ssh root@node2 "mdadm -S --scan"
>> mdadm -S --scan
>> mdadm --zero-superblock /dev/sd{g,h,i}
>> for i in {g,h,i};do dd if=/dev/zero of=/dev/sd$i oflag=direct bs=1M \
>> count=20; done
>>
>> echo "mdadm create array"
>> mdadm -C /dev/md0 -b clustered -e 1.2 -n 2 -l mirror /dev/sdg /dev/sdh
>> echo "set up array on node2"
>> ssh root@node2 "mdadm -A /dev/md0 /dev/sdg /dev/sdh"
>>
>> sleep 5
>>
>> mdadm --manage --add /dev/md0 /dev/sdi
>> mdadm --wait /dev/md0
>> mdadm --grow --raid-devices=3 /dev/md0
>>
>> mdadm /dev/md0 --fail /dev/sdg
>> mdadm /dev/md0 --remove /dev/sdg
>> #mdadm --wait /dev/md0
>> mdadm --grow --raid-devices=2 /dev/md0
>> ```
>
> I found it was hard for me to follow this set. IIUC, the two patches try to
> address one issue. Please add a cover letter and reorganize the descriptions
> like:
>
> cover-letter: error behavior, repro steps, analysis, and maybe describe the
> relationship of the two patches.
> 1/2 and 2/2: what is being fixed.
>
> Thanks,
> Song
>
> [...]
>
