> On 21 Nov 2018, at 23:47, Artur Paszkiewicz <artur.paszkiewicz@xxxxxxxxx> wrote:
>
> Successfully adding a journal device will change the consistency policy
> to journal, there is no need to use --grow --consistency-policy= after
> that.
Ahh, thank you, some things I saw hinted at this being an automatic change.
> So the problem is that your array had a write-intent bitmap before you
> added the journal device. Using bitmap and journal was blocked some time
> ago in the kernel: 230b55fa8d64 ("md: forbid a RAID5 from having both a
> bitmap and a journal."), but apparently not in mdadm. Assemble the array
> with --update=no-bitmap, that should fix it.
Fantastic, this has in fact fixed it.
/dev/md1:
Version : 1.2
Creation Time : Fri Jul 10 16:02:39 2015
Raid Level : raid6
Array Size : 8788666368 (8381.53 GiB 8999.59 GB)
Used Dev Size : 2929555456 (2793.84 GiB 2999.86 GB)
Raid Devices : 5
Total Devices : 6
Persistence : Superblock is persistent
Update Time : Wed Nov 21 23:50:34 2018
State : clean
Active Devices : 5
Working Devices : 6
Failed Devices : 0
Spare Devices : 0
Layout : left-symmetric
Chunk Size : 512K
Consistency Policy : journal
Name : nas:1 (local to host nas)
UUID : 4b9d2327:07ac21d0:b134c2fb:0270e37a
Events : 196101
Number Major Minor RaidDevice State
0 8 35 0 active sync /dev/sdc3
1 8 49 1 active sync /dev/sdd1
6 8 67 2 active sync /dev/sde3
5 8 17 3 active sync /dev/sdb1
4 8 1 4 active sync /dev/sda1
7 259 5 - journal /dev/nvme0n1p1
So is this a case where the procedure should have been to remove the bitmap
first, or should mdadm never have allowed me to attempt what I did?
Thanks again.
Matt.
