RAID5 Array 'failed'

[sidewinder]

Got a Dell Poweredge SC 1420, about 4 years old.

3x 1Tb SATA discs in RAID5 with CERC SATA 1.5/6ch card

 

Started a high pitched beep yesterday, I powered the server down, rebooted and all was OK (at that point I should have checked things, didnt even click that it could be the RAID card making the noise) then was using today when server froze and the beeping started again.

 

Now every time it boots up it sticks on the RAID initialization screen, saying array failed. But it detects all 3 drives. If I go into RAID management, 2 of the 3 drives appear slightly greyed out in the array. So they appear to not have failed, just they wont come online.

 

Ive tried the obvious things like checking connections, which hasnt made any difference.

 

There is a restore option but from reading around there is no guarantee the data will be intact after that so Id rather leave as a last resort. Just wondered if there is a way to force the disks online, even if its just long enough for me to copy data off? Or any other ideas?

 

Unfortunately it isnt backed up (due to lack of space) though neither is it vital (ghost images, recreating will be a pain but not impossible)

[Michael]

If 2 disks fail at once in RAID5 then the data is lost I'm afraid.

[sidewinder]

Yeah I know that, but is that likely to be what has happened? Since they are all detected, and just due to the extremely low chance of 2 failing together

[3s-gtech]

Detection may well be due to the controller on the disk, they could be detected even if the disk itself has failed. Could you run a boot disc with some HD test utilities on it?

[sidewinder]

Tried running some utils from UBCD but none seemed to recognise any of the drives, I assume because of the RAID card (and they're probably desktop level tools)

[Michael]

I suspect a compatible RAID driver is still required. It is possible for two drives to fail at the same time. Rare admittedly, but it does happen.

[SC-UK]

It is possible for two drives to fail at the same time. Rare admittedly, but it does happen.

 

I've seen before, especially with the system being 4 years old, that a drive failure causes a power spike to one of the other drives, causing that to fail as well. Seems to be a problem with "older" PSU's or on board RAID power supplies.

[richardp]

something like this happened to me last year where all the disks looked fine but just would not go beyond initialisation. it turned out to be the raid card itself that was faulty. This was on a Dell Poweredge 2850 which when I contacted Dell about getting a replacement for the built in card was told it was a very old model. They then sent me a pci-e addin card after I had ordered a pci-x card which I got the details for from their website. They were adamant it was the card I had ordered from the code number, finally they admitted that the dell website was wrong and took the card back and sent out the right one. Did not leave me feeling all that impressed especially when the card cost £390 !!! as the server was out of warranty.

 

Richard

[sidewinder]

lol...finally plucked up the courage to use the 'restore' option and instead of taking hours like I thought, in 3 seconds it had brought the drives online and now the server boots fine!

 

Currently backing up the ghost images since there is the chance that something is still faulty

[timzim]

I had similar problems with a CERC SATA 1.5/6ch on a Dell PowerVault 745. It happened a few times and (if i remember right) in the end a firmware update on the CERC card stopped it happening again.

[sidewinder]

Cool, well as soon as the backup has finished I'll do a firmware update