Below par performance of SATA drives on FC4-Test2

[Russell Coker]

On Monday 09 May 2005 17:42, Vamsee Krishna Gomatam <vamsee.krishna at gmail.com> 
wrote:
> /dev/sda:
>  Timing cached reads:   1676 MB in  2.00 seconds = 837.29 MB/sec
> HDIO_DRIVE_CMD(null) (wait for flush complete) failed: Inappropriate
> ioctl for device
>  Timing buffered disk reads:  120 MB in  3.00 seconds =  39.94 MB/sec
> HDIO_DRIVE_CMD(null) (wait for flush complete) failed: Inappropriate
> ioctl for device

40MB/s isn't that great by today's standards, but you omitted to mention the 
RPM speed of the disk.  For a 10,000rpm or better disk 40MB/s is below par, 
you should be seeing 50MB/s or more.

If it's a cheaper drive (designed for desktop not server) then 40MB/s might be 
performing to spec.  That doesn't mean there's a problem, disks that cost 
less and/or make less noise deliver less performance.

For a real test do something more than that.  Bonnie++ is one option for 
benchmarking that's a lot better than hdparm.

Do some benchmark tests with a serious benchmark program as well as some tests 
using dd to read large amounts of data (more than 1G).  Also run top while 
running the tests and see if excessive CPU use is occurring.

-- 
http://www.coker.com.au/selinux/   My NSA Security Enhanced Linux packages
http://www.coker.com.au/bonnie++/  Bonnie++ hard drive benchmark
http://www.coker.com.au/postal/    Postal SMTP/POP benchmark
http://www.coker.com.au/~russell/  My home page