topic Re: Saving RAID1 data from failed server in ProLiant Servers (ML,DL,SL)

Saving RAID1 data from failed server

Alinn — Tue, 30 Jun 2009 15:27:28 GMT

Hi,

I have a DL360 G3 server with a failed motherboard with a RAID 1 configuration.

I wish to save date from the disks, using a ML370 server.

I already have 4 disks on ML370 G4 and 2 free slots.

If I put the disks from the DL360 on the free slots, the controller doesn't see any logical drive and ACU say that the new drives are "unassigned".

What should I do to can save data from the old array?

Assigning the new disk with ACU or if I define a new logical drive from the controller's firmware will erase the actual data from them?

If I put the disks on the first two slots of ML370 replacing the actual ones, will make the old array to be available without aditional (destructive) configuration operations?

The order in with the disk was in the old server is important?

What is the possible solution to recover data from drives?

Re: Saving RAID1 data from failed server

Gary Benavides Meza — Tue, 30 Jun 2009 16:03:30 GMT

Each of the hard drives on the server gets an SCSI ID, so it would be recommended to replace the hard drives on the same bay where they were populated previously and on a server with the same storage controller.

The Proliant Servers with smart array controllers have a feature called Drive Roaming, which allows administrators to move one or more hard drives that are members of a configured logical drive to a different bay position as long as the new bay position is accessible by the same controller (not an array with OS). The RAID information is stored in the drives and not in the controller, so moving the disks to another server will not result in data lost, but keep this in mind:

- Make sure that the server where you are moving the hard drives has the same storage controller
- If the RAID array to be moved contains the OS there should not be any other RAID array in the destination array controller.
(This is because the bootable array will always be the array "A" and if you have an array already in the controller the new array will be tagged as array "B" and if you need to boot from it the array "B" it will not be possible until it's marked as array "A", even though that does not mean that the data will be lost)
- For proper OS functionality the destination server should be another just like the source, this to avoid BSOD and other app. issues, however if it's only data there would be no problem at all

Re: Saving RAID1 data from failed server

Alinn — Tue, 30 Jun 2009 20:35:12 GMT

Indeed, the drivers contain OS and data but I don't indend to boot from it.

I only need to see the old drives as a logical disk on array and the image/rescue software do the rest.

Seem that the "same controller type" is mandatory.

I put the disks on the first 2 bays of the ML370 without any other harddisks and the bios of the controller doesn't recognise them as logical array.

I don't know what SCSI controller is using DL360 vs. ML370 but it just doesn't work..

Any other sugestion to read the data from them?

Re: Saving RAID1 data from failed server

Alinn — Tue, 30 Jun 2009 20:37:19 GMT

PS:
1. "same controller type" = same controller model. The type is the same (Ultra320)

2. To recover data I boot from rescue/linux CD, only if I could see the logical drive..