Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I have recently acquired an Intel SASMF8I raid card for my office. I have 5 x 1TB disks and have created two arrays, 3 x 1TB RAID-5 and 2 x 1TB RAID-1. The second drive is for sensitive/vital data and the RAID is for temporary and desired but not vital data. I chose this set-up because it maximises my current space on 5 disks without having to buy an extra one for RAID-10, I plan to use the other 3 ports for RAID-1 for performance.

My question is, Why am I seeing no Read Speed advantage on my RAID-5? I know RAID-5 is the most complex controller, and that speed/s are dependant on controller design, but I should see a boost in read speeds over just a single hard drive?

1TB WD10EARX, Avg. 98.0MB/s at 14.ms.
3x1TB SASMF8I RAID-5, Avg. 83.1MB/s at 15.1ms.

EDIT: The above tests were done using HD Tune, I had closed down all running programs, and of course neither of them were the system disk. I know it probably isn't the best program to use, but the same tests were run (with a 36hr initialise in between).

I have read the manual, I have configured the drive/s and checked the configuration. The card itself is plugged into a PCIex16 port (the card is PCIex8). The RAID-5 has sync'd/initialized and I have checked everything in the RAID Console application. I cannot think of anything else restricting the performance.

Any pointers as to the read speeds above? This is a hardware implementation and although the throughput can't be theoretically projected, I should see an increase in read (and maybe write/s too, though not as much) performance, shouldn't I?

share|improve this question
1  
what single drive are you measuring against? you said you created a 2x1TB mirror pair –  Rex Aug 2 '12 at 14:53
    
Are the measurements for the 1 drive and the 3 disks as a RAID 5 done on the same computer or are you testing these with 2 machines? –  Patrick Aug 2 '12 at 19:21
    
I am assuming so, but are you using the exact model of HDD for the single drive and the raid 5? I know this sounds a little redundant but I suggest running the tests again, making sure that there are no scheduled processes during either test (including updates to programs, virus scans, etc). –  Patrick Aug 2 '12 at 19:37
    
Same computer, same drive/s in both tests. I did close any significant processes, and these drives were not system disks. –  Ghostpsalm Aug 2 '12 at 23:49

1 Answer 1

We have no clue how you have measured these facts, of course.

Perhaps if you describe what you did, somebody can advise you on how to interpret the data you have, or how to run better tests to produce the data you want.

Real I/O tests can be performed with software such as IOmeter.

share|improve this answer
    
Good point! I have added the edit above. –  Ghostpsalm Aug 2 '12 at 10:06

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.