ProLiant Servers (ML,DL,SL)
cancel
Showing results for 
Search instead for 
Did you mean: 

15k harddrives slower than 10k

 
Ayman Altounji
Valued Contributor

15k harddrives slower than 10k

I noticed the record I opened on this problem months ago has been deleted without ever receiving confirmation that Compaq is aware of the problem and intends on doing anything.

Problem is that with SCO 5.0.5 and either the Integrated RAID or 5304 Array Controller, large file copy performance with the 15k hard drives is significantly slower (40%) than exact same system on 10k drives. (3x9GB U3 drives configured RAID 0). File size is 285MB of image data. The previous argument I received was that file copy is not a valid measure of performance. It is the only valid measure if the work being done is printing large image files or processing streaming video while simultaneously receiving 200MB+ files to the same hard drive through a network connection. With color printing the file sizes will only get larger (1-2GB).
Testing with Caldera and NT Workstation shows the 15k drives perform 50% better (as expected) than the 10k drives. The same is true with SCO and the 431 Array controller. But the combination of SCO 5.0.5, Integrated RAID or 5304, and 15k drives is a big loser. Stick with 10k drives, save money, get 40% performance improvement.
2 REPLIES 2
Ayman Altounji
Valued Contributor

Re: 15k harddrives slower than 10k

What is the speed you get on a large file creation, ie

timex dd if=/dev/zero of=bigtest bs=1024 count=2000000

Can you try this on both configurations?
Highlighted
Ayman Altounji
Valued Contributor

Re: 15k harddrives slower than 10k

The answer is no. I am done testing this problem for Compaq. When offered a reasonable solution, I will try it. Until then, I wait on Compaq to acknowledge the problem and decide whether a solution is planned.