HP MSA 1000 controller failure - beyond useful life?
The end question is, specifically what would be the expected useful life of an HP MSA 1000 controller board? I had one "go nuts" recently, and it was suggested that it was past its expected life. I think the MSA 1000 is about six years in service. Yes, hard drives fail, but a controller board that is only electronics, in a fairly dust free, static and temperature controlled server room, on a dedicated power circuit, connected via a UPS (standby, with over and under voltage protection) shouldn't "wear out" in about six years should it?
A single drive in the array went bad, not an unusual event. A hot spare was present and configured. The controller did not pick up the failed drive status, and did not pull the hot spare online and rebuild the array automatically as it should have.
Instead, the O/S (Windows Server 2003) ended up seeing the volume as heavily corrupted, and resulted in many errors as end users tried to access shares and the shared files on the volume. The volume could be seen, but the longer things went the more corruption seemed to occur. A rebuild of the volume was a final result, along with replacing the controller board in the MSA 1000.
For more info, the array was one LUN, with RAID6 (ADG). Ten physical disks. Rebuild priority was not set to zero. Everything I know of should indicate the array should have easily survived a single drive failure.
A single drive in the array went bad, not an unusual event. A hot spare was present and configured. The controller did not pick up the failed drive status, and did not pull the hot spare online and rebuild the array automatically as it should have.
Instead, the O/S (Windows Server 2003) ended up seeing the volume as heavily corrupted, and resulted in many errors as end users tried to access shares and the shared files on the volume. The volume could be seen, but the longer things went the more corruption seemed to occur. A rebuild of the volume was a final result, along with replacing the controller board in the MSA 1000.
For more info, the array was one LUN, with RAID6 (ADG). Ten physical disks. Rebuild priority was not set to zero. Everything I know of should indicate the array should have easily survived a single drive failure.
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
Thanks once again for the comments. I don't recall any capacitors or other components that might wear on the MSA1000 controller board. Makes sense about the battery, as it is a rechargeable. I had no indication that the battery failed in this instance.
Getting ready to shutdown the topic. Any further comments anyone, particularly for a high performance, server grade HP storage array?
Getting ready to shutdown the topic. Any further comments anyone, particularly for a high performance, server grade HP storage array?
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Replacement for MSA1000 is P2000 G3 range, or for higher performance with more disks or SSDs the P6000 range.
We had more failures with the P2000 range than we ever had with the MSA1000 series.
6-8 complete failures last year, which HP claimed were overheating backplane faults or psu faults something weird.
all purchased at different times, only problem visiting Engineers do not seem know these baby units.
6-8 complete failures last year, which HP claimed were overheating backplane faults or psu faults something weird.
all purchased at different times, only problem visiting Engineers do not seem know these baby units.
Yeah, I don' treally like them either, couldn't understand why HP didn't use a pair of Smart Array controllers plus a management chip like they did with the MSA1000 rather than OEM'ing something from Dot Hill.
ASKER
Thanks everyone. I split the points widely, as this is a rather general discussion. It does provide valuable feedback to support my case that this MSA1000 unit should not be determined "beyond useful life" at its current time-in-service.
ASKER
Any other thoughts anyone? Although it pertains specifically to a HP MSA 1000, it's a bit general also, whether electronics should "wear out" in a arguably short timeframe.