CentOS RAID failed

[ocean-addict]

Hi

If anyone can help with this It would be much appreciated.

I have a CentOS machine with 3 disks in a RAID 5 array, when I boot it says there are not enough disks (2/3 failed), unable to mount md0, unable to read superblock. This is more or less what it says obviously..

The problem is that they used software RAID, and one of the disks that failed might be the one holding the software.

I tried running the rescue from the cd, but it says that no linux partitions are found.. Do I need to mount the parrtitions first or something?

From reading this you may see that I am not very clued up with software RAID, or RAID in general. Please help!

Thanx

 

[stickytoffee]

Before booting linux does it pick up all the drives or not in the BIOS?

 

[ocean-addict]

I can see all 3 drives in the BIOS. From the rescue mode on the CentOS cd, i run fdisk -l and all 3 drives are listed. I run mdadm --examine on all 3 drives and they all show that they are part of the active raid array, but query /dev/md0 tells me that md0 is a md device which is not active and is too small to be an md component.

I have new hard drives if that would help to rebuild the array.. but i cant figure out which drive is actually faulty, if any..

thanx again

 

[stickytoffee]

Are the partition types on the individual partitions set to “fd” (linux raid autodetect)?

What is the output of:

1. mdadm -D /dev/md0
2. cat /proc/mdstat

?

 

[ocean-addict]

they are all set to fd..

mdadm -D /dev/md0:

mdadm: mdd device /dev/md0 does not appear to be active

cat /proc/mdstat:

personalities : raid0 raid1 raid5 raid6
unused devices : none

 

[stickytoffee]

You should try activate the raid by running:

# mdadm -A /dev/md0 /dev/hd[abc]1
# mdadm -D /dev/md0

The above assumes that you are using hda1/hdb1/hdc1 as partitions....

See if that gives any errors?

 

[ocean-addict]

mdadm -A...

mdadm: /dev/md0 assembled from 1 drive - not enough to start the array.

Im giving up, were busy digging through backups...

Thanx for all your help tho

 

[stickytoffee]

No probs

If you still want to try ... try adding the "faulty" drives back to the array manually:

mdadm /dev/md0 -a /dev/hd[ab]1

Substitute [ab] for your drive letters.... this might help...

 

[ocean-addict]

mdadm: cannot get array info for /dev/md0

 

[koffiejunkie]

You need to specify all the drives when assembling the raid. What does the following give you:

fdisk -l


(that's a lower case L)

 

[ubercal]

Hi

If anyone can help with this It would be much appreciated.

I have a CentOS machine with 3 disks in a RAID 5 array, when I boot it says there are not enough disks (2/3 failed), unable to mount md0, unable to read superblock. This is more or less what it says obviously..

The problem is that they used software RAID, and one of the disks that failed might be the one holding the software.

I tried running the rescue from the cd, but it says that no linux partitions are found.. Do I need to mount the parrtitions first or something?

From reading this you may see that I am not very clued up with software RAID, or RAID in general. Please help!

Thanx

RAID 5 needs at least 2 hdds to function.You only have 1 functional HDD which suggests that you have lost all your data

 

[koffiejunkie]

Doesn't seem all that urgent to me...

 

[ocean-addict]

Doesn't seem all that urgent to me...

Sorry, went on leave for a couple days. A new raid has already been built from new drives and the old data copied across, so thanx for all the help, but like ubercal said, i reckon with 2 drives down it was a bit of a difficult one..

 

[koffiejunkie]

Consider RAID6 - it's like RAID5 but with two parity drives. Or in that case, if you're only dealing with 4 discs (in which case RAID6 would give you 2 discs worth of space), consider RAID10, which would give you the same amount of space, but much better performance).