Dilemma.
One of the mirrors on your Supermicro has gone. No SNMP management to pick this up? See another article on this website.
It’s evident from the output of mdstat that something is wrong if you look below at [U_]:
cat /proc/mdstat Personalities : [raid1] md0 : active raid1 sdb1[0] 9766303680 blocks super 1.2 [2/1] [U_]
On a working server this should be [UU]:
cat /proc/mdstat Personalities : [raid1] md0 : active raid1 sdb1[0] sdc1[1] 3906885440 blocks super 1.2 [2/2] [UU]
So you’ve zeroed the drive, which took around 24 hours, and now you have replaced it. But what next?
“Partition the Disk: If your RAID array requires partitions rather than using entire disks, you will need to partition the new disk similarly to the other disks in the array. You can use fdisk
or parted
to create the partitions. For a simple RAID setup, you might mirror the partition layout of another disk in the array.”
My network engineer suggested gdisk instead, so these instructions will use that:
gdisk /dev/sdx
- Type
n
to create a new partition. - Enter the partition number (or press Enter to use the default).
- Specify the first and last sectors (or press Enter to use the default full disk).
- For the partition type, type
fd00
which stands for Linux RAID. - Finally, type
w
to write the new partition table to the disk and confirm the changes.
Add partition to mirror
sudo mdadm –manage /dev/md0 –add /dev/sdx1
Monitor RAID
cat /proc/mdstat
Monitor with more Detail
mdadm -D /dev/m mdadm --detail /dev/md0
Check Sync Action
cat /sys/block/md0/md/sync_action check
More information
Lots of goodies here: https://unix.stackexchange.com/questions/28636/how-to-check-mdadm-raids-while-running