Performance
Showing results for 
Search instead for 
Do you mean 

autoraid performance and configuration

SOLVED
Go to Solution
Frequent Advisor

autoraid performance and configuration

I have an HP AutoRaid 12H. Here is a listing of my configuration after losing a drive.
----------------------------------------------------
Array State = WARNING
Server name = able1-e00
Array type = 3
Mfg. Product Code = IJMTU00004

--- Disk space usage --------------------
Total physical = 34,733 MB *
Allocated to LUNs = 19,456 MB *
Used as Active Hot spare = 0 MB *
Used by non-included disks = 0 MB *
Used for Redundancy = 15,277 MB *
Unallocated (avail for LUNs) = 0 MB *
-----------------------------------------

THE FOLLOWING WARNING(S) EXIST:
- Active Hot Spare Unavailable


Here is a listing after the drive is replaced. Array State = READY
Server name = able1-e00
Array type = 3
Mfg. Product Code = IJMTU00004

--- Disk space usage --------------------
Total physical = 43,416 MB *
Allocated to LUNs = 19,456 MB *
Used as Active Hot spare = 8,683 MB *
Used by non-included disks = 0 MB *
Used for Redundancy = 10,503 MB *
Unallocated (avail for LUNs) = 4,774 MB *
-----------------------------------------
Is there really an active hot spare configured? When the disk died, the performance of the autoraid was terrible. Shouldn't the Hot Spare go into action, and the autoraid keep going like normal? Does this have anything to do with Raid 0/1 vs. Raid 5 configuration? Thanks for your help.
1 REPLY
Honored Contributor Honored Contributor

Re: autoraid performance and configuration

Hello Raymond,

I know MPE performance rather well, but I don't know much about the 12H, so I asked the disk team to comment. Their response:

Most likely the slowdown was because of two reasons: The rebuild priority is probably set to high, which means it rebuilds and handles I/Os as it has "spare" time during the rebuild. Also, the "Unallocated" space is used by the autoraid to make itself faster and more efficient. With the disk bad and the hotspare space being used to rebuild, there wasn't any "unallocated" space available. The rebuild priority can be changed. The customer needs to decide what is more important, accessing autoraid or getting rebuild finished and having redundancy in case of another disk failure.
A problem well defined is half solved.