Hard disks, solid state drives

HDD, SSDForum → Poor performance raid5. What the. . . ?

Poor performance raid5. What the. . . ?

Pages: 1
18.04.2003 14:50:00
Hello everybody!
facts and only the facts:
Mother - Asus PR-DLS (Dual Xeon)
Perc - 2 x Xeon
Controller - Adaptec 2110S SCSI160
Screws - 3 x Seagate Chetah 10k U160 27Gb
OS - the Windows 2000 Server

Sobral RAID5 without problems. All saw everything and earned, but. . . I run the HD Tach:

In 1500 the cost of the array of green productivity is lower than that of the screw 85 IDE green! Knew would be that it will be built SLR two IDE "shkah and would live happily ever after. Where is the vaunted SCSI-performance???

way, matters on which the connectors loop hanging screws? Both the trail?

Ready to answer any questions. The array to Monday can distort as you like.

ZYZH one picture is not visible, open the URL in a separate window
18.04.2003 15:20:00
to Hdtach fucking mother! ! ! ! !
it does not work correctly with RAID-controllers (and also with screws).
18.04.2003 15:55:00
Aah. . . Well it's a little easier. . .
and then what to measure correctly?
And Adaptec Storage Manager PRO (service) gave me something about the Non-recoverabile (seems ) error.
And then about the fact that the drive has been stop and then again start. This buggy
concrete screw as I understand it?
18.04.2003 16:46:00
exactly. . . the controller is disconnected from the bus. . .
you to take the usual speed test PAR and copying it shows DSP. . . Here
and copy pair kinoshek row. . .
18.04.2003 18:26:00

Well, what is it? I plan to take it soon.

rather than if measured correctly?
Why measure at all?
18.04.2003 18:27:00
Nichchego not understand. . .
right now connected to the embedded controller screws. I tested individually. Charts - SUPER! Reading Zone 65Mb / s tightened to the floor surface, further decline. Read Burst 140! ! ! And so all the screws.
All screws individually perfectly formatted and work. A RAID dead So it turns out that the RAID controller, if buggy? ? ?

Or hand curves to create RAID "Yes, but there is possible to run three times to click on the & lt;??? OK & gt;... And an hour wait and that's all building

Nichchego not understand ......
Now try individually to test them through

controller Why

A measure then, to know that - whether you are what has made would be thrown out in 1500 tanks

18.04.2003 19:57:00

Well, if you need a simple and naglyadnenky testik, you can try ATTO Disk Benchmark. . .

array will need to mark before the test, and watch the speed, for example, for Overlapped I / O (with file size to 32MB)
19.04.2003 8:52:00
new day - the new results.

When I connect the drive via the built-in SCSI controller then we have:
Average Linear Read Speed ​​= 52Mb / s
Average Access Time = 8, 4ms
Cache Read Speed ​​= 132Mb / s

When the same disc is connected via Adaptec 2110S get :
Average Linear Read Speed ​​= 39Mb / s
Average Access Time = 8, 4ms
Cache Read Speed ​​= 48, 4Mb / s

Who knows what kind of x. . nya? ? ? Of course SCSI360 drive. The controller on mamma SCSI360. Adaptec 2110S - SCSI160, BUT Sorry, one disk performance SCSI160 no means limit can not.

Juzaju instead of HD Tach program from Drive! - It seems more honest
19.04.2003 11:46:00

RAID-controller WB mode actually decreases the speed of reading from the cache of the hard drive. But increasing the average speed of processing in the other modes.
change caching mode WT - will be faster

Drive is also not for the arrays is done. . .
19.04.2003 13:16:00

Okay. . . last question. . . uh. . . and where the regime change?
The controller BIOS something not found
And what is that such regimes - where you can read?
senks advance.
19.04.2003 14:30:00

in the BIOS of the controller!
read the documentation (search WriteBack or WriteThrough)
Pages: 1
HDD, SSDForum → Poor performance raid5. What the. . . ?

Poor performance raid5. What the. . . ?