raidtest for zfs
Pawel Jakub Dawidek
pjd at FreeBSD.org
Sat Sep 22 04:21:44 PDT 2007
On Sat, Sep 22, 2007 at 11:24:46AM +0400, Michael Monashev wrote:
> ????????????, Pawel.
>
> > What's strange about those results? Can you show us 'zpool status' and
> > explain what you expect and why?
>
> > zpool status
> pool: tank
> state: ONLINE
> scrub: none requested
> config:
>
> NAME STATE READ WRITE CKSUM
> tank ONLINE 0 0 0
> raidz1 ONLINE 0 0 0
> ad6s1 ONLINE 0 0 0
> ad8s2 ONLINE 0 0 0
> ad10s3 ONLINE 0 0 0
> ad12s1 ONLINE 0 0 0
> ad14s2 ONLINE 0 0 0
>
> errors: No known data errors
>
> I compare graid5, graid3+gcache, raidz. My target is find software
> RAID with better read performance and huge storage size. I planing
> place a lot of little (1-100kb) images on it.
>
> Here is my test results: http://michael.mindmix.ru/168-958-rezul-taty-testirovanija-graid5-graid3-gcache-i-raidz.zhtml
>
> graid3+gcache with round-robin reading is 1.25 faster than 1 disk.
> graid5 is 2.18 faster than 1 disk.
> raidz is 7.7 faster than 1 disk.
>
> 5 disks cant read faster than 5 times. raidz is 7.7 faster than 1
> disk!
>
> This is strange for me.
ZFS ZVOL doesn't present raw blocks to use, like graid3/graid5, so it's
quite possible that because of how the data was written and how ZFS
prefetches the data it's faster than one disk.
You may want to compare one raw disk to single disk pool.
--
Pawel Jakub Dawidek http://www.wheel.pl
pjd at FreeBSD.org http://www.FreeBSD.org
FreeBSD committer Am I Evil? Yes, I Am!
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-geom/attachments/20070922/1a6d22a6/attachment.pgp
More information about the freebsd-geom
mailing list