RAID Array is not rebuilding

[JSchow]

I have a HP/Compaq ML530 G2 with a Smart Array 532 w/ 32MB of RAM. The server has a RAID 5 Array /w 4x 36.4 GB (10k, Ultra3) Drives that are Hot Swappable.
One of the drives in the RAID 5 Array has failed and I have purchase a new drive to replace the failed drive. When I put the new drive in the server, I expected the array to start rebuilding and fix the RAID 5 array. However, the array did not and has not started a rebuild. I have verified that the 532 controller has the latest firmware, the OS has the most current driver, and I am using v8.35 ACL. I have restarted the server several times and the array still has not started to repair itself.
Can anyone tell me why and how I can fix this? A screen shot is attached to this post.
The only thing that I can tell is that the SA 532 is looking for a drive in slot 6, but the drive that failed was in slot 2 in the 1st bay. There is no physical slot 6 in bay one. And since the drives start at SCSI ID 0... it can only go to ID 5 if there are 6 physical slots.
Any help would be appreciated.

[wobbe]

This is very strange indeed. SCSI id should be in the second optional drive cage that you don't have. Faulty cabling or perhaps a defective backpane?

[Anonymous]

I see two unassigned drives in the system.
It should be possible to add a hotspare to the RAID and after rebuild add it as a member.

[JSchow]

I thought of trying to add one of the two drives as a online spare and then having that drive rebuild the array... but no such luck. I don't have that option anywhere in the Array Configuration Manager.
Is there anyway for me to "assign" a different SCSI ID to drive? My understanding is that the drive gets its SCSI ID from the slot in the cage.

And I also though of the backplane going bad, but if that were true, I would think that I wouldn't be able to see the new drive in the "bad" slot.

[gregersenj]

I have seen this problem in a few post here, and once at a customer.
It seem to happen on old fw revisions.

There's no easy fix, to my best knowledge.

It is possible to fix it though.
You will have to put, all the drives in a compatible drive cage, wich kan hold more that 6 drives. And then put a drive in the bay, that has ID6.
When it has finished rebuilding, you can put the drives back in the server, and the Id will be changed.

Another option has been suggestet - Add a hotspare, and live with that.

If you can't / won't do one of the above options.
then the only option it to delete the array. Then recreate, re-install and restore

br
/jag

[JSchow]

It's not that I haven't tried adding the new drives as spares to the array... trust me I have. It just won't let me add them until the array has been repaired.
Any idea if I would have more options available to me if I booted to the a SMART Start CD? I have tried booting to the OCRA menu of the RAID card, but all that I see is view/create/delete an array.

[Ken Krubsack]

Just out of curiosity, you DID buy an HP/Compaq drive, right? Not an IBM or Dell or some other 3rd-party vendor?

I could see this happening with a drive that the array controller doesn't recognize or that might be bad itself. I s'pose another option is that the backplane is flaking out and therefore the controller is never seeing the drive in the first place. It could also just flat be a bad replacement drive.

You do see lights come on the drive when you plug it in, I presume?

[JSchow]

I am sure that I purchased a HP/Compaq drive.
I took the replacement part number right off of the dead drive. Looks and works the same as the dead drive.

I forgot to mention above that I purchased two new drives. (the two shown in the attached pic as unassigned drives) One as a replacement for the dead drive and another to be used as a spare in case another one were to die. I tried both new HD's in the same slot. Same thing happens to both replacement drives. They show up as unassigned drives with SCSI ID 2.

So, I'm guessing that need to backup the server, delete the array, and restore?

[cnb]

Hi,

>So, I'm guessing that need to backup the server, delete the array, and restore?

Yep. A known issue. I've seen this before and there's no easy work-around other than what has already been stated. Happened to us with a non-existent SCSI ID 5 (in a 4-slot backplane).


An advisory on an MSA array has similar issues:

http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?lang=en&cc=us&taskId=110&prodSeriesId=440223&prodTypeId=12169&objectID=c01136618


A real P.I.T.A!


Rgds,

[JSchow]

Thanks everyone for you help and info.