Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 21 Mar 2001 14:43:15 +0100 (CET)
From:      Jan Conrad <conrad@th.physik.uni-bonn.de>
To:        Matt Dillon <dillon@earth.backplane.com>
Cc:        Alfred Perlstein <bright@wintelcom.net>, Gordon Tetlow <gordont@bluemtn.net>, Rich Morin <rdm@cfcl.com>, <freebsd-stable@FreeBSD.ORG>
Subject:   Re: NFS performance
Message-ID:  <Pine.BSF.4.33.0103211332070.1867-100000@merlin.th.physik.uni-bonn.de>
In-Reply-To: <200103201911.f2KJBAR96066@earth.backplane.com>

next in thread | previous in thread | raw e-mail | index | archive | help
ok - to sum up a bit..

- for a good LAN, use UDP
- use v3 (this is what I thought)
- use standard data sizes

but still...

Our network connection between client and server is going directly over a
BaySwitch 450 24T, all interfaces set to 100baseTX, half-duplex.
So the LAN is good.
We run 4 nfsd's and 4 nfsiods on each machine.

I did the mount as 'mount_nfs ....'
Even so 'mount' does'nt show, I suppose I made a v3 mount (from the
source code of mount_nfs).
BTW, is there any way to figure that out?

If I do a 'dd if=/dev/zero of=zero2 bs=16k count=64x400' I get
419430400 bytes transferred in 147.402451 secs (2845478 bytes/sec)
(varying from 2.5 to 2.8 Mb/s)
maybe the box was loaded yesterday at my first try...

Locally the server does 10Mb/s write and 33Mb/s read.
So isn't 2.8 Mb/s a bit slow, still?

Do you know what actually determines the writing speed in a case like
our's? Network or disk?

BTW, what is that '-q' switch to mount_nfs actually doing?

-Jan

-- 
Physikalisches Institut der Universitaet Bonn
Nussallee 12
D-53115 Bonn
GERMANY








To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.33.0103211332070.1867-100000>