ProLiant Servers (ML,DL,SL)
1753767 Members
5656 Online
108799 Solutions
New Discussion

Remove the P400 SmartArray to read/recover individual SAS disks of a RAID 5 array?

 
Scott-
Visitor

Remove the P400 SmartArray to read/recover individual SAS disks of a RAID 5 array?

I happen to personally own an ML370 G5, and recently moved internationally. Upon unpacking, the machine reports that 2 hard drives of the RAID 5 array it had (of 8 SAS 72GB disks) have failed -- and so it shuts down both the LSi UltraSCSI 320   and the P400 SmartArray during boot. And as luck would have it, the box which contains the backup is one of two boxes (out of several hundred) which (so far) ;have been lost in the move!!

In this scenario, I think the first thing to do is to check if the "failed" drives really have failed; and even if they have, to see if ordinary recovery tools can restore them at least temporarily. I also think I should verify that the problem isn't with the slots that the drives occupied. The problem is that the drives are 72GB SAS drives (attached to an Lsi UltraSCSI 320 Series 2000 card),.  I can't just pop the drives into some free slot on another machine at home and run testdisk or spinwrite or ddresescue -- I have no other SAS controller at home. 

My questions is this: Is there some way that I could backup the P400/LSi configuration, then eliminate the RAID/array, leaving me with 8 separate SAS disks:? The first step in any procedure I have seen online is to get another machineto which I can attach the disks to examine/image them. My question is is there a relatively risk-free way to use the machine itself as the "other machine".  It seems especially attractive to me to do this because at POST I'm told that the drives in bay 6 and 8 have failed, and ORCA tells that the (failed) RAiD 5 is made up of disks 1, 2, 3, 4,5, 6, and 8. But all 8 bays are populated. So it's even possible that I already have a spare drive lying around! Could I use ORCA to define a logical vlume that is just this disk in bay 7 (without RAID)? And if so, could I hot-swap other physical disks into that bay 7 and have the controller just present me with the disk as a single un-RAIDed disk? 

Would it work to use the 2nd redundant ROM as a configuration back-up. The procedure would be: remove all 8 SAS disks, then boot up and clear the RAID at boot via ORCA.  Then hot-plug individual drives and examine/ddrescue/etc. as I see fit; and if I manage to ddrescue up a clone of one of the "failed" drives onto my 'unused disk 7, I would then put the 7 disks back in place, and now boot using the 2nd backup (redundant) ROM as the machine ROM?

At first, I hoped that I could just remove the P400 and see the 8 separate SAS disks without the 'benefit" of the P400,  but examinig the server showed me that in fact that 8 SAS disks are directly attached to the p400, and NOT to the LSi Ultra320 SCIS Series 2000 card. So I'm led to ask if I can 'turn off' the RAID from the combination of P400 and LSi Ultra320, and then 'turn it back on' once I have a working 7th disk.

Both  ORCA and iLO report a failed Logical Drive, but as far as I can tell, one result of this is that the P400 is disabled by ORCA when it detects the failed RAID. As a result, ACU doesn't even see the LSi controller, let alone the RAID.  Perhaps someone can tell me what I can do to re-enable the P400/SCSI without RAID once the OS is up and running (the thing boots both Linux and Windows FWIW). 

I realize that I could spend some money and get an SAS controller card to hook the disks up, but 1. It seems very wasteful to me to spend such money, if I could just use the SCSI card I have. I just don't want to mess things up further. And I do want to exercise the slots that the failed drives came from, so I'd like in any case to be able to use the SCSI card and cage without the RAID. 

If I do need to buy a standalone SAS controller to examine the disks, if anyone could recommend a suitable (cheap) one, I'd be very grateful. I doubt I'll need the card for any other purpose than this. 

Perhaps I'm too naive, but i have a hard time believing that TWO disks out of 8 completely failed as a result of sitting idle for 6 months. And I also would hope that it is possible to do something as simple as back up the configuration of some RAID and restore it to the same hardware after I use the hardware with RAID not configured. ANd I'll be prettyannoyed if I shell out for an SAS controller, discover that in fact that disks are fine, and then go on to learn that it's some hardware in the ML370 that is the problem.

Thank you very much for any advice/suggestions.

Thanks,

scott