[Novalug] benchmarking iSCSI

Miguel Gonzalez Castaños miguel_3_gonzalez at yahoo.es
Thu Sep 13 16:48:10 EDT 2007


kkauffman at headfog.com escribió:
> hdparm can still be used for timings, just not much adjustment of
> parameters.  SCSI is GENERALLY already optimized and you don't need to
> adjust it like you do the IDE interface drives, but you can do some
> adjustments.
>
> Reference stuff...
> http://people.redhat.com/alikins/system_tuning.html#scsi
> http://www.ibm.com/developerworks/linux/linux390/perf/tuning_rec_dasd_ioScheduler.html
> http://www.eslim.co.kr/pds/pds/2/5/RHEL_Tuning_Guide.pdf
>   
I don't want to do tuning, just to know how iSCSI is doing in terms of
readings/writings.

I used bonnie, and I get the following results for:

bonnie -d /mnt/virtualdata1/ -s 512 -m svn -u root

Version  1.03       ------Sequential Output------ --Sequential Input-
--Random-
                    -Per Chr- --Block-- -Rewrite- -Per Chr- --Block--
--Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP
/sec %CP
svn            512M 13594  43 15746  18  2292   3  2702   9  2905   2
482.8   4
                    ------Sequential Create------ --------Random
Create--------
                    -Create-- --Read--- -Delete-- -Create-- --Read---
-Delete--
              files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
/sec %CP
                 16  1685  91 +++++ +++ 25747  92  1576  91 +++++ +++
5134  92
svn,512M,13594,43,15746,18,2292,3,2702,9,2905,2,482.8,4,16,1685,91,+++++,+++,25747,92,1576,91,+++++,+++,5134,92

This results make sense?

Miguel




More information about the Novalug mailing list