> Read the man page (spare-group parameter), mdadm allows for one spare
> drive to be used
> for multiple arrays.
I have read the manuals, but it doesn't work as described, or likely I
do something wrong. Are there any working examples ?
My config is simple -
[root@memverge2 anton]# cat /proc/mdstat
Personalities : [raid6] [raid5] [raid4]
md118 : active raid5 nvme0n1[4] nvme3n1[0](S) nvme7n1[3] nvme5n1[1]
3125362688 blocks super 1.2 level 5, 512k chunk, algorithm 2 [3/3] [UUU]
bitmap: 0/12 pages [0KB], 65536KB chunk
md117 : active raid5 nvme6n1[4] nvme4n1[1] nvme2n1[0]
3125362688 blocks super 1.2 level 5, 512k chunk, algorithm 2 [3/3] [UUU]
bitmap: 0/12 pages [0KB], 65536KB chunk
unused devices: <none>
[root@memverge2 anton]#
[root@memverge2 anton]# cat /etc/mdadm.conf
ARRAY /dev/md/117 level=raid5 num-devices=3 metadata=1.2
UUID=0fab82cd:36301f6c:6ec78c95:7f092d4c spare-group=group1
devices=/dev/nvme2n1,/dev/nvme4n1,/dev/nvme6n1
ARRAY /dev/md/118 level=raid5 num-devices=3 metadata=1.2 spares=1
UUID=2d7290e5:c3ccbfe9:004cb182:3e325714 spare-group=group1
devices=/dev/nvme0n1,/dev/nvme3n1,/dev/nvme5n1,/dev/nvme7n1
[root@memverge2 anton]#
[root@memverge2 anton]# systemctl restart mdmonitor
[root@memverge2 anton]#
[root@memverge2 anton]# mdadm /dev/md117 --fail /dev/nvme6n1
[root@memverge2 anton]#
[root@memverge2 anton]# mdadm -D /dev/md117
/dev/md117:
Version : 1.2
Creation Time : Tue Jan 28 11:08:21 2025
Raid Level : raid5
Array Size : 3125362688 (2.91 TiB 3.20 TB)
Used Dev Size : 1562681344 (1490.29 GiB 1600.19 GB)
Raid Devices : 3
Total Devices : 3
Persistence : Superblock is persistent
Intent Bitmap : Internal
Update Time : Tue Jan 28 15:29:47 2025
State : clean, degraded
Active Devices : 2
Working Devices : 2
Failed Devices : 1
Spare Devices : 0
Layout : left-symmetric
Chunk Size : 512K
Consistency Policy : bitmap
Name : memverge2:117 (local to host memverge2)
UUID : 0fab82cd:36301f6c:6ec78c95:7f092d4c
Events : 849
Number Major Minor RaidDevice State
0 259 5 0 active sync /dev/nvme2n1
1 259 0 1 active sync /dev/nvme4n1
- 0 0 2 removed
4 259 1 - faulty /dev/nvme6n1
[root@memverge2 anton]#
And there is no raid117 recovering using spare from raid118.
Anton
пн, 27 янв. 2025 г. в 23:18, Dragan Milivojević <galileo@xxxxxxxxxxx>:
>
> Read the man page (spare-group parameter), mdadm allows for one spare
> drive to be used
> for multiple arrays.
>
> On Mon, 27 Jan 2025 at 15:15, Anton Gavriliuk <antosha20xx@xxxxxxxxx> wrote:
> >
> > How to add a spare disk to raid50 which consists of several raid5
> > (7+1) ? It would be more economical and flexible than adding spare
> > disks to each raid5 in raid50.
> >
> > Anton
> >
