Urgent : Array initialized - what to do ?

[pkreli]

Server is ML350 G4.

Win2K3 crashed with a blue screen. Restarted. Message came Array controller not configured or something like that then started initializing.....

Now cannot boot.


What can I do to prevent data loss ? Or is it already lost ? Urgent advice needed please.

BTW, it is 4 x 72GB disks in RAID 1+0

[TTr]

> then started initializing.....
How did it start, all by itself or you initialized the array? What exactly did it do in the initialization?
> Now cannot boot
If you look in the array now, what is its status? What kind of configuration does it have? Logical devices, raid type etc.
Be prepared for the worst.

[pkreli]

Ok here's what had happened. Windows had crashed with blue screen after I hot-plugged a spare drive into an empty bay (Slot 4) planning to use as a spare drive. After windows crashed, I powered down, removed the drive I had added and restarted. (The other four disks are in their original positions, I hadn't moved them).

Then came the message that the array was not configured and it started initializing automatically. I waited and of course it wouldn't load the OS.

Now I rebooted to see what message would come up. After reaching array controller section, it says :

"Smart Array 642 Controller Initialising
0 Logical Drives

1785-Slot 3 Drive Array Not configured
No drives detected"



Apparently, the logical disk is lost !
Is there any way to recover the data ?
Please advise.

[TTr]

Was the disk that you plugged in the same as the other four? I don't understand why inserting a hot pluggable disk crashed the server unless there was something seriously wrong with the disk. It more than crashed the server, it looks like it wiped out the array information which was is on the disks. What kind of raid controller is it? Look in it and see if there is anything that explains this. I am not sure that the array information on the disks is totally wiped out or the controller just went bad. If you have another same controller, try the disks on that one.

[pkreli]

Well the card is Smart Array 642 (as shown in the message I posted above).

No I hadn't touched the original existing drives (4 drives in bay 0 - 3). Bay 4 and 5 are empty so I'd plugged the extra drive at bay 4.

Actually the spare drive was in fact one of the array drive itself before it was replaced two days before because its SMART message said it may fail soon. I'd plugged it in to erase it before disposing or using it as spare until it really failed. But Windows crashed soon after I plugged it into the empty bay.

[TTr · Accepted solution]

As you know, all array configuration information for your raid1+0 is kept on the disks and this includes the number of physical disks and sizes, number of logical drives and sizes, stripe size etc. So you removed a disk, let's say the disk in slot2 and replaced it with a new disk and allowed the array to rebuild. So now you have two disks that are ID-ed as the same disk from slot2, one in slot2 and one spare on the shelf. Two days later you plugged in the spare disk in slot4 and the raid controller read its contents and saw the disk in slot2 twice. I don't know if that crashed your controller and caused it to wipe out all raid configuration from the disks but I would suspect that it could happen.

One other question. When you removed the disk from 2 days back and installed a replacement one, did the controller start rebuilding the raid1+0 array to include the new disk and do you know if the rebuild process completed 2 days later when you added the 5th disk?

I would try the following (I am not suggesting you do this but you decide before you try it). As you know raid1+0 is 2 mirrored pairs. I would plug in drive0 and drive1 or drive0 and drive2 only and start the server hoping to end up with one member from each mirrored pair and see if the raid controller is able to read any disk information.

I don't know and I don't think anybody can tell what state the disks are in. May be a company that specializes in data recovery can help. It depends how valuable your data was. If you have HP support you should call them first.

[pkreli]

Thanks a lot TTr for your support so far. I am really grateful.

Yes I'am sure the rebuilding was finished the same day within a few hours (that was on Tuesday).

Thanks for your idea about the recovery. I was thinking along the same lines but not exactly sure.

Another idea I am thinking, in case the array information has been wiped from all four disks, could it be possible for me to get the information from the spare disk (that I had put in the 5th bay)?

I am hoping the logical drive info is still on that drive as I pulled it out right after Windows crashed and I shut down the power. Then maybe I could use this info to rebuild the logical drive exact as before ?

But if I go this way, how can I connect it to the array ? Can remove the other four (to prevent further damage) and just put the 'spare' one in the array to have a peek ? I'm fairly sure that one still contains the array info.

[TTr]

At this point anything that you try is uncertain. You can try the 5th disk too but I have not seen of a way to propagate the array configuration from one disk to the others. The data may still be on the disks but the array configuration may not. This is on a special area on the header of the disk and not much if any is public knowledge about it. As I mentioned earlier the data recovery specialist may be the only ones that can help.

[Lesley Persyn]

To exclude array/backplane/cable issue's, you should test with another server. Just plug the disks in, and hope you get everything back.

[pkreli]

Yes it turned out to be backplane. It got fried it seems. All four disks were ok. Everything was back to normal after replacing the backplane.

Now I am wondering whether I should ever try inserting any drive into the two empty bays.