Date: Mon, 17 Jul 1995 10:33:17 +0100 (BST) From: Doug Rabson <dfr@render.com> To: Bruce Evans <bde@zeta.org.au> Cc: current@freebsd.org Subject: Re: slow nfsv3 writes Message-ID: <Pine.BSF.3.91.950717102824.7169J-100000@minnow.render.com> In-Reply-To: <199507151413.AAA23348@godzilla.zeta.org.au>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, 16 Jul 1995, Bruce Evans wrote:
> The speed of writes seems to be slower for nfsv3 according to iozone:
>
> iozone 1 512: 9055 bytes/sec
> iozone 1 8192: 113028 bytes/sec
>
> Read speeds are good (800K and 1000K/sec) although this may be due to
> local caching (it would take too long to test a large file size at
> 9055 bytes/sec :-).
>
> I don't know exactly what the write speeds used to be. I remember
> something over 100K/sec and hoped for more. I don't remember such a
> low speed for the small block size, except for a Linux client using
> an unsuitable nfs block size.
I don't have a FreeBSD NFSv3 server to test against (I only have one
machine running current) but when I mount a v3 filesystem from an sgi, I
get:
iozone 1 512
63250 bytes/second for writing the file
725501 bytes/second for reading the file
iozone 1 8192
372827 bytes/second for writing the file
16777216 bytes/second for reading the file
It is interesting to see that the local cache appears to only work if the
read size matches the filesystem blocksize (8k in this case).
The 370k/sec is respectable but I would expect it to be much larger for
NFSv3, for a reasonably fast server.
>
> Bruce
>
Doug Rabson, Microsoft RenderMorphics Ltd. Mail: dfr@render.com
Phone: +44 171 251 4411
FAX: +44 171 251 0939
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.91.950717102824.7169J-100000>