Topic: Raid & mdadm (again...) SOLVED!

[jumba]

Sorry folks, but I feel somewhat confused with determining the correct command to resynchronising a failed RAID.

I've read http://wiki.contribs.org/Raid over and over again, but I simply fail to see the clear logic in how to determining which partition to rejoin

The output from cat /proc/mdstat on this particular server reads:

Code: [Select]

# cat /proc/mdstat
Personalities : [raid1]
md2 : active raid1 sda2[0]
     156143680 blocks [2/1] [U_]

md1 : active raid1 sda1[0] sdb1[1]
     104320 blocks [2/2] [UU]

unused devices: <none>

...so please could someone confirm if my assumption here is correct?

If I understand correctly, the command to execute should be:

Code: [Select]

mdadm --add /dev/md2 /dev/sdb2

(It's not my server, so I really want to be 100% sure...)

[christian]

That looks correct to me and it is the command I would issue.

Also watch the rebuild via /proc/mdstat to ensure it is rebuilding correctly. What is not clear is why the disk got kicked out of the array so if there is a real issue, the disk may get kicked out again.

per the wiki entry you referenced, you can monitor mdstat via

Code: [Select]

watch -n .1 cat /proc/mdstat

[jumba]

Thank you Christian for your reply!

It feels much better just to have someone else confirming ones ideas...

The "reason" for the RAID error in this particular casa is most likely several power failures. (The particular server doesn't have UPS)

But of course we'll monitor the rebuild as well as keeping a close watch on that disk for a while.

[Stefano]

The "reason" for the RAID error in this particular casa is most likely several power failures. (The particular server doesn't have UPS)

well.. time to spend money then..

Ciao
Stefano and.. ups will save your day/work/data

[jumba]

Sure, it's not my server, I just suggested that purchase...