| Author |
Message |
Tom
Guest
|
 Posted:
Sun Oct 10, 2004 2:10 am Post subject:
Strange? RAID Benchmark Results |
|
|
Can anyone clear up why I may be getting the following results?
Powermac G4 933MHz
OS 10.3.5
Adaptec 39160 SCSI controller
4 meter Dell (Amphenol connectors) VHDCI to 68-pin SCSI Cable
Gateway (Advanced Logic Research) 6-slot LVD backplane
200W AT Power supply
Seagate Barracuda 50.1GB ST150176LC Drives
Using the OS X Utility Xbench to test and built in OS X Raid software
to stripe multiple drives:
TEST Drive 1 Drive 2 Drive 3 RAID 1+2 RAID 2+3
---- ------ ------ ------ -------- --------
1 9.28 22.58 8.77 11.61 25.76
2 8.95 21.74 8.86 11.20 24.76
3 18.53 16.80 18.72 18.92 22.69
4 22.85 22.85 22.88 45.69 40.76
TEST RAID 1+3 RAID 1+2+3
---- --------- ----------
1 11.55 12.57
2 11.06 12.07
3 16.20 16.02
4 45.71 60.05
All values are in MB/sec
Test 1 = Sequential Uncached Write [4K Blocks]
Test 2 = Sequential Uncached Write [256K Blocks]
Test 3 = Sequential Uncached Read [4K Blocks]
Test 4 = Sequential Uncached Read [256K Blocks]
Right now the drives are sitting out in the open in slots 1-3 of the
backplane with the external 68-pin cable directly attached to the
backplane (I'm cheating since I don't have an internal-to-external
cable yet).
Do I have problems with Drives 1 and 3?
I'm not letting them sit for very long before I test, and Drive 2 is
between the other ones, perhaps it comes up to temp faster. But I
have tested Drive 1 right after boot up and then about 10 mins later
when I can feel a obvious temp change and the readings are practically
identical.
On the 256K Read test, my rates go up about 20MB/s/drive which seems
to make sense, but the write rates are terrible!
SCSI is really confusing me! ARGH! LOL Please help!
-Tom |
|
| Back to top |
|
 |
Folkert Rienstra
Guest
|
| Posted:
Sun Oct 10, 2004 2:59 am Post subject:
Re: Strange? RAID Benchmark Results |
|
|
"Tom" <tmbannon@comcast.net> wrote in message news:db5fc446.0410091410.3aefbaa3@posting.google.com
| Quote: | Can anyone clear up why I may be getting the following results?
Powermac G4 933MHz
OS 10.3.5
Adaptec 39160 SCSI controller
4 meter Dell (Amphenol connectors) VHDCI to 68-pin SCSI Cable
Gateway (Advanced Logic Research) 6-slot LVD backplane
200W AT Power supply
Seagate Barracuda 50.1GB ST150176LC Drives
Using the OS X Utility Xbench to test and built in OS X Raid software
to stripe multiple drives:
TEST Drive 1 Drive 2 Drive 3 RAID 1+2 RAID 2+3
---- ------ ------ ------ -------- --------
1 9.28 22.58 8.77 11.61 25.76
2 8.95 21.74 8.86 11.20 24.76
3 18.53 16.80 18.72 18.92 22.69
4 22.85 22.85 22.88 45.69 40.76
TEST RAID 1+3 RAID 1+2+3
---- --------- ----------
1 11.55 12.57
2 11.06 12.07
3 16.20 16.02
4 45.71 60.05
All values are in MB/sec
Test 1 = Sequential Uncached Write [4K Blocks]
Test 2 = Sequential Uncached Write [256K Blocks]
Test 3 = Sequential Uncached Read [4K Blocks]
Test 4 = Sequential Uncached Read [256K Blocks]
Right now the drives are sitting out in the open in slots 1-3 of the
backplane with the external 68-pin cable directly attached to the
backplane (I'm cheating since I don't have an internal-to-external
cable yet).
Do I have problems with Drives 1 and 3?
|
Not necessarily. Check the write cache settings on the drives.
| Quote: |
I'm not letting them sit for very long before I test, and Drive 2 is
between the other ones, perhaps it comes up to temp faster. But I
have tested Drive 1 right after boot up and then about 10 mins later
when I can feel a obvious temp change and the readings are practically
identical.
On the 256K Read test, my rates go up about 20MB/s/drive which seems
to make sense, but the write rates are terrible!
SCSI is really confusing me! ARGH! LOL Please help!
-Tom |
|
|
| Back to top |
|
|
Tom
Guest
|
| Posted:
Mon Oct 11, 2004 4:20 am Post subject:
Re: Strange? RAID Benchmark Results |
|
|
"Folkert Rienstra" <see_reply-to@myweb.nl> wrote in message news:<2sstkqF1pcs2sU1@uni-berlin.de>...
| Quote: | Do I have problems with Drives 1 and 3?
Not necessarily. Check the write cache settings on the drives.
|
That sounds reasonable, but I checked out the manuals for the drives
(Seagate ST150176LC) and I don't see any way to enable/disable the
write cache, at least not through jumpers. Seagate's website has
diagnostic software but its only Windows/Linux. I don't have access
to a PC with SCSI and I don't want to install Linux on the G4 just to
test these. I'm hunting around for some other utility I can use that
may help.
Tom |
|
| Back to top |
|
|
Ron Reaugh
Guest
|
| Posted:
Mon Oct 11, 2004 4:59 am Post subject:
Re: Strange? RAID Benchmark Results |
|
|
"Tom" <tmbannon@comcast.net> wrote in message
news:db5fc446.0410101620.49fc8771@posting.google.com...
| Quote: | "Folkert Rienstra" <see_reply-to@myweb.nl> wrote in message
news:<2sstkqF1pcs2sU1@uni-berlin.de>...
Do I have problems with Drives 1 and 3?
Not necessarily. Check the write cache settings on the drives.
That sounds reasonable, but I checked out the manuals for the drives
(Seagate ST150176LC) and I don't see any way to enable/disable the
write cache, at least not through jumpers. Seagate's website has
diagnostic software but its only Windows/Linux. I don't have access
to a PC with SCSI and I don't want to install Linux on the G4 just to
test these. I'm hunting around for some other utility I can use that
may help.
|
Right, it requires a utility to turn on write caching and it's not a jumper.
Lose the MAC. |
|
| Back to top |
|
|
Tom
Guest
|
| Posted:
Tue Oct 12, 2004 1:05 am Post subject:
Re: Strange? RAID Benchmark Results |
|
|
| Quote: | Right, it requires a utility to turn on write caching and it's not a jumper.
|
I used an old copy of FWB Hard Disk Toolkit and sure enough the write
cache was off on drives 1 and 3 but on on drive 2. I turned it on and
also had HDT "optimize" the settings for each drive (backing up the
old settings of course).
I booted back into OS 10.3 (that old software had to run in OS 9 or
earlier) so that I could use Xbench to test the drives as before.
And I'm getting basicly the same results I posted above! There must
be something else besides simply checking the "Write Cache Enabled"
box, but that was the reason I had HDT optimize the settings. I
thought it would correct the cache settings since I wasn't sure what
to set anything at. And I checked after optimization just to make
sure it didn't completely turn off the cache for some reason.
That's just crazy talk! My Mac is great. I have had no problems
occur do to the fact that its a Mac - things like my problems with
these drives have happened, but that's been due to the equipment or my
lack of understanding, not that fact that I'm using a Mac and not a
PC. As a hard drive manufacturer, Seagate's lack of multi-operating
system support is a bit ridiculous. At least they had a Linux
version, though I think it was x86 only. Anyway I was an avid PC user
growing up but started using old Macs then bought a G4 and love it. I
don't play games anymore and I can't think of anything I haven't been
able to do on the Mac that I could do on a PC.
Of course I'd certainly welcome a PC In Addition to the Mac if
somebody wants to send me a freebie! ;)
thanks for the help so far!
Tom |
|
| Back to top |
|
|
Folkert Rienstra
Guest
|
| Posted:
Tue Oct 12, 2004 9:59 am Post subject:
Re: Strange? RAID Benchmark Results |
|
|
"Tom" <tmbannon@comcast.net> wrote in message news:db5fc446.0410111305.638e16f1@posting.google.com
| Quote: | Right, it requires a utility to turn on write caching and it's not a jumper.
I used an old copy of FWB Hard Disk Toolkit and sure enough the write
cache was off on drives 1 and 3 but on on drive 2. I turned it on and
also had HDT "optimize" the settings for each drive (backing up the
old settings of course).
I booted back into OS 10.3 (that old software had to run in OS 9 or
earlier) so that I could use Xbench to test the drives as before.
And I'm getting basicly the same results I posted above! There must
be something else besides simply checking the "Write Cache Enabled"
box, but that was the reason I had HDT optimize the settings. I
thought it would correct the cache settings since I wasn't sure what
to set anything at.
And I checked after optimization just to make sure it
didn't completely turn off the cache for some reason.
|
But did you check also after reboot? SCSI has 'current' and 'saved' settings.
Current settings default to saved after a reset so if you didn't change the
saved values as well ......
| Quote: |
Lose the MAC.
That's just crazy talk! My Mac is great. I have had no problems
occur do to the fact that its a Mac - things like my problems with
these drives have happened, but that's been due to the equipment or my
lack of understanding, not that fact that I'm using a Mac and not a
PC. As a hard drive manufacturer, Seagate's lack of multi-operating
system support is a bit ridiculous. At least they had a Linux
version, though I think it was x86 only. Anyway I was an avid PC user
growing up but started using old Macs then bought a G4 and love it. I
don't play games anymore and I can't think of anything I haven't been
able to do on the Mac that I could do on a PC.
Of course I'd certainly welcome a PC In Addition to the Mac if
somebody wants to send me a freebie! ;)
thanks for the help so far!
Tom |
|
|
| Back to top |
|
|
Tim Kelley
Guest
|
| Posted:
Wed Oct 13, 2004 8:05 am Post subject:
Re: Strange? RAID Benchmark Results |
|
|
Tom wrote:
| Quote: | Can anyone clear up why I may be getting the following results?
Powermac G4 933MHz
OS 10.3.5
Adaptec 39160 SCSI controller
4 meter Dell (Amphenol connectors) VHDCI to 68-pin SCSI Cable
Gateway (Advanced Logic Research) 6-slot LVD backplane
200W AT Power supply
Seagate Barracuda 50.1GB ST150176LC Drives
|
Isn't a 200w ps pushing it for that system? |
|
| Back to top |
|
|
Tom
Guest
|
| Posted:
Wed Oct 13, 2004 10:25 pm Post subject:
Re: Strange? RAID Benchmark Results |
|
|
Tim Kelley <tim@removefornospam.it.kpt.cc> wrote in message news:<2832120.FC1TkSxNBq@el-gato.bombsquad.23rdward.org>...
| Quote: | Tom wrote:
Can anyone clear up why I may be getting the following results?
Powermac G4 933MHz
OS 10.3.5
Adaptec 39160 SCSI controller
4 meter Dell (Amphenol connectors) VHDCI to 68-pin SCSI Cable
Gateway (Advanced Logic Research) 6-slot LVD backplane
200W AT Power supply
Seagate Barracuda 50.1GB ST150176LC Drives
Isn't a 200w ps pushing it for that system?
|
If I didn't I should have specified that the 200W PS is an old AT unit
that I'm using to power the external drives. With only 3 drives at
this point I should be fine. I may be pushing it when I start using
all 6 Barracudas, but if calculations show I need more power I'll pick
up a bigger PS.
I haven't had a chance to work on these any further in the past few
days. I need to stick some fans onto the test setup to keep things
cool before I run the drives too long.
Tom |
|
| Back to top |
|
|
|
|
|
|
FAQ
Memberlist
Register
Profile
Log in to check your private messages
Log in