Re: NFS tuning - high performance throughput.

[Chris Penney <cpenney@gmail.com>]

On 6/15/05, Dan Stromberg <strombrg@dcs.nac.uci.edu> wrote:
> On Wed, 2005-06-15 at 14:52 -0500, Roger Heflin wrote:
> > The mount/remount (unless you time how long it takes to unmount),
> > will potentially still skew the write test, and the writes
> > will still be happening for quite a while after the dd completes.
>=20
> Or unless you sync twice?  Or write synchronously.
>=20

You can optionally use iozone (www.iozone.org) you can use '-c -e' to
include fsync() and close() times.  It also has an option to umount
the volume between iterations.

Also, I missed that he was using a 4k stripe size before.  That's got
to be his issue.  My luns are 8+1 h/w raid 5 using a 64k segment size
per disk and I use 512k (8*64) as the stripe size.  As was
intelligently noted, you really want your writes to fill a whole
stripe if you can.  If they do not my understanding is that you should
lower both your segment size and stripe width.  I think you always
want your stripe width to match the width of your raid 5 lun.

   Chris


-------------------------------------------------------
SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
from IBM. Find simple to follow Roadmaps, straightforward articles,
informative Webcasts and more! Get everything you need to get up to
speed, fast. http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs