I've got a Dual Celeron 466 system (Abit BP6)
running four maxtor IDE hard drives on the ATA66 controller (using ATA33 for all
four drives). Two of the drives are 27GB 7200RPM, one is 17GB 5400RPM and
the fourth is 13GB 5400RPM. The two smaller drives are about 9.5MB/sec
with hdparm -t and the two larger drives are about 20MB/sec with
hdparm.
The array is 78GB and currently I've got a chunk
size of 32 (though I've played with other sizes from 8 to 64).
Running the array I get about 20.25MB/sec with
hdparm.
In the old days (using the raid software, circa
2.0.36) I used to get about 28-30MB/sec with the 13, 17 and a 8GB drive. I
can run hdparm -t /dev/hde & hdparm -t /dev/hdg and get a combined 37MB/sec
or so by adding the numbers together.
Does anyone have any idea why the performance is so
low? I'd expect numbers at least in the 40MB/sec range, being that the
drives are on seperate channels on the ide.
I'm running 2.2.14pre11 with the ide patch and the
latest raid patch (8/24?). ATA/66 mode is disabled for the channels (lock
up problems). I also got similar numbers on a seperate motherboard running
the stock 2.2.13 with the raid patches.
|
- Re: RAID0 performance odditity Michael Trainor
- Re: RAID0 performance odditity Erik van Zijst
- Re: RAID0 performance odditity Michael Trainor
- Re: RAID0 performance odditity Erik van Zijst
- Re: RAID0 performance odditity Robert
- Re: RAID0 performance odditity Michael Trainor
- Re: RAID0 performance odditity stephan
- Re: RAID0 performance odditity Andrea Arcangeli
- Re: RAID0 performance odditity raid
- Re: RAID0 performance oddit... Andre Hedrick