Controller settings... RAID5 recovery???

[Calab]

A while ago we had a RAID5 array fail (thanks Seagate!!!) and now we're trying to recover our data. I'm an old hand with the computer, but never had to deal with this type of situation before.

Setup:
 P4 machine runnning Windows 2003 Server
 Adaptec 21610SA PCI-x 16 port SATA RAID controller in a 32bit PCI slot
 Four 500g Seagate hard drives connected to ports 0-3 of the controller.
 1.3TB RAID 5 array created with the defaults of the Adaptec controller.

What happened:
 The machine had been working fine for months and we had a drive failure. No problem... I disconnected the failed drive and checked the array. It was running in a degraded state - so I know I pulled the correct drive. We installed a new drive into the PC and started the rebuild of the array.

During the rebuild, the new drive failed. From that point forward the array was no longer viable. The Adaptec controller will not do anything with this array.

At this point I have the orignal four array drives, and the new drive that failed partway through the rebuild. The drives have not been written to since the failure.

My problem:
 I don't know the particulars used by this RAID controller. I'm guessing that it created a Left Asynchronous array, using a 64KB stripe and an offset of 64KB. I have no way to know if this is actually correct and attempts to find out from Adaptec have failed.

What do I do now?

[Joep]

Hello,

Well, if your question is regarding our iRecover: iRecover will try to work out array parameters on its own. So, the idea is, iRecover needs access to the individual disks (non-RAIDed) and I suggest you leave the only partially rebuilt one out. iRecover does not repair the array but will allow you to copy data from the array to a safe location.

[Calab]

Thank you!  That was what I wanted to know.

I do have a couple suggestions for this software...
- Allow the user to see the serial number of the drives when they are choosing drives
- Put a "Analysis started at 00:00 on 09-01-01" message up telling the user when the analysis actually started.
- Number each step of the analysis uniquely. This will allow a user to say, "I got to step 4" instead of "I got to the [2] Full - identifying data" process, for example.

I've already started an analysis with the partial drive included. It's at 4% of [2] Full - identifying data.

[Joep]

- Allow the user to see the serial number of the drives when they are choosing drives

It's shown on the right