Folks,
I've got a machine I'm working on, and I'm seeing some unusual
problems, so I wanted to get a sanity check.
The machine in question is a Dell 1300 with two 450Mhz processors
(512KB L2 cache each) and 1GB ECC RAM. In addition to the on-board
Adaptec AIC-7890 controller (to which the system disk is attached, a
Quantum Atlas IV 9GB), I have two Adaptec 2940U2W controllers.
Each of these two controllers is attached to a different
interface on a Hitachi/Comparex D1400 mainframe-style
refrigerator-size drive array, with twenty 18GB Seagate 10kRPM
"Cheetah" disks (plus one hot spare not currently being used) spread
across five internal SCSI channels, and with 1GB ECC battery
backed-up write-back cache (mirrored between the two controllers, so
effective size is halved).
On this machine, I have installed FreeBSD 3.4-STABLE (I cvsup'ed
late last week, so it should be fairly up-to-date).
Needless to say, we would expect this system to pretty much
scream at disk I/O. Of course, this is precisely what it is not
doing.
Fortunately, all the work I have been doing over the last few
days is in preparation for a Hitachi/Comparex performance expert to
come in tomorrow and help us tune this thing for maximum performance
from their perspective, but I'm seeing something weird from the host
side and I'd like to ask if anyone else has seen anything similar.
In particular, when we run the "postmark" benchmark on a single
filesystem, we get some decent results (although not nearly as good
as we'd like). When we run the same benchmark on two different
filesystems mounted through two different controllers, well,
performance really goes into the toilet -- the result is about 1/10th
as fast as the single filesystem test.
But, when we test the same two filesystems mounted through the
same Adaptec controller, we see each test get almost exactly half the
performance of the single filesystem test.
This just doesn't make any sense to me. If anything, I would
expect the dual controller configuration to result in each postmark
benchmark getting nearly as much throughput as the single filesystem
test, so that the aggregate is almost twice as much. But instead,
it's ten times as slow?!?
This just boggles my mind, and I don't have the first clue where
to go looking to see what the problem might be.
Is there any additional information I could provide that might
prove helpful? Could there be conflicts of some sort between the two
Adaptec 2940U2W controllers?
Thanks for any and all advice, observations, and assistance you
can provide!
--
These are my opinions and should not be taken as official Skynet policy
_________________________________________________________________________
|o| Brad Knowles, <[EMAIL PROTECTED]> Belgacom Skynet NV/SA |o|
|o| Systems Architect, Mail/News/FTP/Proxy Admin Rue Col. Bourg, 124 |o|
|o| Phone/Fax: +32-2-706.13.11/726.93.11 B-1140 Brussels |o|
|o| http://www.skynet.be Belgium |o|
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/
Unix is like a wigwam -- no Gates, no Windows, and an Apache inside.
Unix is very user-friendly. It's just picky who its friends are.
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-stable" in the body of the message
