software RAID data recovery on Rocks Cluster

[markmh]

Hello,

We have a software RAID 1 array, and one of the disks (sda) has failed, and a few hundred gigs of data have disappeared. The type is a software not a hardware RAID. Unfortunately, we may not have the resources to get a replacement disk.

-Priority1: Our priority is to recover the data. It contains the research of a few people. What is a good way to access the data safely?
Is it possible to use the following strategy to gain access to the data, then back it up onto an external disk?:

umount /compute
mkdir /compute-other
mount /dev/sdb /compute
mount /dev/sda /compute-other

The drive that we have been working on (composed of the RAID1 system) is called /compute. Would mounting the good drive on /compute allow us to gain access to the data and back it up?


-Priority2: Also, to get our system back up and running, without destroying our data, our hope is that we can restore the RAID using the current (failed or kicked out, I am not sure which) sda disk. It has been suggested to me to do this in our system:

mount /dev/md0 /compute
mdadm /dev/md0 --add /dev/sda

but online, I have also seen:
raidhotadd /dev/mdX /dev/sdX

Unfortunately, my knowledge is quite rudimentary and I am not sure which of these options to use, or what is necessary to be able to use them correctly. For instance, do I need to put the system into recovery mode or something? Or run
/sbin/mdadm --assemble --scan
as I have also seen suggested?
A preliminary test suggests to me that we do not have raidhotadd, but we do have the mdadm program.

Any help would be highly appreciated. Thank you.

(more info, FYI, if useful:
system info: CentOS release 4.4 (final) operating system
distribution is Rocks release 4.2.1 (Cydonia)

see attached text file for the contents of /proc/mdstat /etc/fstab and /var/log/messages )

[fuxin]

We have a software RAID 1 array, and one of the disks (sda) has failed

How do you know that one of the disks is corrupt?

What is a good way to access the data safely?

Is the data there or what do you mean by "safely access"? Can you access the /compute directory at all?

[markmh]

How do you know that one of the disks is corrupt?

Is the data there or what do you mean by "safely access"? Can you access the /compute directory at all?

The log message from /var/log/messages indicates this is so (this message is in the text file attached to my post). And there has been an issue with this same sda disk in the past.

The compute directory show up, in fact, all of the user's directories inside /compute show up, the problem is that it thinks they are all empty, after the disk problem starting on the 22nd.

Thanks!