On 5/9/2010 10:50 PM, Nick Holland wrote:
Look at the blinky lights on the hard disks? I know, macho admins love to look at magical system parameters, but I usually solve such problems by looking at the disk activity lights (and why I dislike Sun and Macintosh systems). I suspect you are i/o bound. (ok, that's not my most clever diagnosis of the day...)
Thanks for the reply. The lights on HP disks aren't that great; I suspect it'll look like it's on most of the time even with light I/O.
* softdeps? (I know, a few people hyperventilate over softdeps on mail servers, but really...if your mail server crashes enough to worry about uncommitted-to-disk messages, you probably have issues much bigger than softdeps. This is mostly an academic issue; in real life, you will lose more mail through your mail filters than you will through crashes and softdeps if your servers are at all reliable).
I have softdeps enabled on all but the root partition.
* cache active on the RAID card? no cache=sucky performance!
I believe it's 64mb read cache, I'll have to see if write cache is enabled (i.e. battery installed). You're right, write cache should help.
* Good RAID config? (raid1+0 rocks. raid5 just pretends to, until you lose a disk, raid1 trades some write performance for redundancy).
RAID 1+0.
* could you have a degraded RAID set? (i.e., if you have RAID5, but blew out a disk that you didn't notice)
Possible but unlikely; hw.sensors doesn't show a failure, and the disk was good when I built the machine.
* Some Compaq/HP RAID systems don't perform overly well on OpenBSD (not sure about other OSs).
Perhaps, I'll see if I can find a similar machine and run some benchmarks. Since most of what I have are ciss and cac, does anyone have approximate disk benchmarks for U320 10k disks in RAID 1+0? I've had good results with OpenBSD and ciss SAS controllers, but I think they're different chips than the ciss SCSI controllers.
