dmesg output for my test scenario with you patch applied:
[534716.791252] md126: detected capacity change from 0 to 41943040
[534716.797684] md: md126 still in use.
[534716.803334] md: md126 stopped.
[534716.829098] md: md127 stopped.
[534718.036483] md126: detected capacity change from 20971520 to 0
[534741.743609] md/raid:md126: device nvme0n1 operational as raid disk 0
[534741.762739] md/raid:md126: raid level 4 active with 1 out of 2 devices,
algorithm 5
[534741.822765] md: reshape of RAID array md126
[534747.144197] md: md126: reshape interrupted.
[534747.149098] md: reshape of RAID array md126
[534747.566093] md: md126: reshape interrupted.
[534747.570979] md: reshape of RAID array md126
[534793.916521] md: md126: reshape done.
and w/o:
[534907.642262] md126: detected capacity change from 0 to 20971520
[534907.648697] md: md126 still in use.
[534907.654340] md: md126 stopped.
[534907.679414] md: md127 stopped.
[534911.985080] md126: detected capacity change from 20971520 to 0
[534922.920777] md/raid:md126: device nvme0n1 operational as raid disk 0
[534922.940442] md/raid:md126: raid level 4 active with 1 out of 2 devices,
algorithm 5
[534922.995454] md: reshape of RAID array md126
[534975.643237] md: md126: reshape done.
[534975.669424] md126: detected capacity change from 41943040 to 20971520
Not sure what is causing errors you see.
btw, I'm working on md-next 5.11.0 kernel from 02/24
On 3/22/2021 6:16 PM, Nigel Croxon wrote:
----- Original Message -----
From: "Oleksandr Shchirskyi" <oleksandr.shchirskyi@xxxxxxxxxxxxxxx>
To: "Nigel Croxon" <ncroxon@xxxxxxxxxx>, linux-raid@xxxxxxxxxxxxxxx
Cc: "Mariusz Tkaczyk" <mariusz.tkaczyk@xxxxxxxxxxxxxxx>, "Jes Sorensen" <jes@xxxxxxxxxxxxxxxxxx>
Sent: Monday, March 22, 2021 12:21:11 PM
Subject: Re: [PATCH] mdadm: fix reshape from RAID5 to RAID6 with backup file
Hello Nigel,
I have collected more info regarding this issue.
I can confirm what Mariusz said, it's a regression caused by patch 4ae96c802203
The reason for failure during the reshape, is that in this patch sync_max
value is set to max, but the function wait_for_reshape_imsm, used in some
reshape scenarios, relies on this parameter, and doesn't expect, that value
can be max. This leads to reshaping fail.
Here's an example of a debug log from this method, when the issue is hit:
mdadm: wait_for_reshape_imsm: wrong next position to set 4096 (2048)
mdadm: imsm_manage_reshape: wait_for_reshape_imsm returned error!
With this patch reverted, the issue is not observed. See my logs below:
# mdadm -CR imsm0 -e imsm -n4 /dev/nvme[0-3]n1 && mdadm -CR volume -l0
--chunk 64 --size=10G --raid-devices=1 /dev/nvme0n1 --force
# mdadm -D /dev/md/volume
/dev/md/volume:
Container : /dev/md/imsm0, member 0
Raid Level : raid0
Array Size : 10485760 (10.00 GiB 10.74 GB)
Raid Devices : 1
Total Devices : 1
State : clean
...
# mdadm -G /dev/md/imsm0 -n2
# mdadm -D /dev/md/volume
/dev/md/volume:
Container : /dev/md/imsm0, member 0
Raid Level : raid4
Array Size : 10485760 (10.00 GiB 10.74 GB)
Used Dev Size : 10485760 (10.00 GiB 10.74 GB)
Raid Devices : 3
Total Devices : 2
State : clean, degraded
...
# git revert 4ae96c802203ec3cfbb089240c56d61f7f4661b3
Auto-merging Grow.c
[master 1166854] Revert "mdadm: fix reshape from RAID5 to RAID6 with backup
file"
1 file changed, 2 insertions(+), 5 deletions(-)
# mdadm -Ss; wipefs -a /dev/nvme[0-3]n1
# make clean; make; make install-systemd; make install
# mdadm -CR imsm0 -e imsm -n4 /dev/nvme[0-3]n1 && mdadm -CR volume -l0
--chunk 64 --size=10G --raid-devices=1 /dev/nvme0n1 --force
# mdadm -G /dev/md/imsm0 -n2
# mdadm -D /dev/md/volume
/dev/md/volume:
Container : /dev/md/imsm0, member 0
Raid Level : raid0
Array Size : 20971520 (20.00 GiB 21.47 GB)
Raid Devices : 2
Total Devices : 2
State : clean
...
#
On 3/16/2021 4:59 PM, Nigel Croxon wrote:
----- Original Message -----
From: "Mariusz Tkaczyk" <mariusz.tkaczyk@xxxxxxxxxxxxxxx>
To: "Jes Sorensen" <jes@xxxxxxxxxxxxxxxxxx>, "Nigel Croxon" <ncroxon@redhat=
.com>, linux-raid@xxxxxxxxxxxxxxx, xni@xxxxxxxxxx
Sent: Tuesday, March 16, 2021 10:54:22 AM
Subject: Re: [PATCH] mdadm: fix reshape from RAID5 to RAID6 with backup fil=
e
Hello Nigel,
Blame told us, that yours patch introduce regression in following
scenario:
#mdadm -CR imsm0 -e imsm -n4 /dev/nvme[0125]n1
#mdadm -CR volume -l0 --chunk 64 --raid-devices=3D1 /dev/nvme0n1 --force
#mdadm -G /dev/md/imsm0 -n2
At the end of reshape, level doesn't back to RAID0.
Could you look into it?
Let me know, if you need support.
Thanks,
Mariusz
I=E2=80=99m trying your situation without my patch (its reverted) and I=E2=
=80=99m not seeing success.
See the dmesg log.
[root@fedora33 mdadmupstream]# mdadm -CR volume -l0 --chunk 64 --raid-devic=
es=3D1 /dev/nvme0n1 --force
mdadm: /dev/nvme0n1 appears to be part of a raid array:
level=3Dcontainer devices=3D0 ctime=3DWed Dec 31 19:00:00 1969
mdadm: Creating array inside imsm container md127
mdadm: array /dev/md/volume started.
[root@fedora33 mdadmupstream]# cat /proc/mdstat=20
Personalities : [raid6] [raid5] [raid4] [raid0]=20
md126 : active raid0 nvme0n1[0]
500102144 blocks super external:/md127/0 64k chunks
md127 : inactive nvme3n1[3](S) nvme2n1[2](S) nvme1n1[1](S) nvme0n1[0](S)
4420 blocks super external:imsm
unused devices: <none>
[root@fedora33 mdadmupstream]# mdadm -G /dev/md/imsm0 -n2
[root@fedora33 mdadmupstream]# cat /proc/mdstat=20
Personalities : [raid6] [raid5] [raid4] [raid0]=20
md126 : active raid4 nvme3n1[2] nvme0n1[0]
500102144 blocks super external:-md127/0 level 4, 64k chunk, algorithm=
5 [2/1] [U_]
md127 : inactive nvme3n1[3](S) nvme2n1[2](S) nvme1n1[1](S) nvme0n1[0](S)
4420 blocks super external:imsm
unused devices: <none>
dmesg says:
[Mar16 11:46] md/raid:md126: device nvme0n1 operational as raid disk 0
[ +0.011147] md/raid:md126: raid level 4 active with 1 out of 2 devices, a=
lgorithm 5
[ +0.044605] md/raid0:md126: raid5 must have missing parity disk!
[ +0.000002] md: md126: raid0 would not accept array
-Nigel
--
Regards,
Oleksandr Shchirskyi
