Date: Thu, 28 Sep 2000 07:55:24 -0700 From: Cy Schubert - ITSD Open Systems Group <Cy.Schubert@uumail.gov.bc.ca> To: Matthew Seaman <m.seaman@inpharmatica.co.uk> Cc: Andrew BOGECHO <andrewb@cs.mcgill.ca>, freebsd-stable@FreeBSD.ORG Subject: Re: nfs errors Message-ID: <200009281455.e8SEtsI40382@cwsys.cwsent.com> In-Reply-To: Your message of "Thu, 28 Sep 2000 09:39:33 BST." <39D303C5.1E440669@inpharmatica.co.uk>
next in thread | previous in thread | raw e-mail | index | archive | help
Sorry to all for a null reply on this subject. Was going to reply to
Andrew BOGECHO changed my mind and hit send instead of abort. Hope the
rest of my day is better than this.
I've been using a FreeBSD client (my desktop system at work) with
Solaris servers for close to three years (or when ever 2.2.6 came out)
when I replaced the Alpha that was on my desk with a FreeBSD system.
My amd.conf looks like:
/defaults type:=host;opts:=rw,intr,proto=udp,vers=2,timo=10,retran
s=5;\
fs:=${autodir}/${rhost};rhost:=${key}
some_alpha_systems opts:=rw,intr,nosuid,nodevs,grpid,soft,timo=
10,\
retrans=5,rsize=1436,wsize=1436
* opts:=rw,intr,nosuid,nodevs,grpid,soft,timo=10,retrans=5
I've had no problems using Solaris servers and FreeBSD clients. Most
of the Solaris systems are on our raised floor which is two hops away
and three are in another city, 5 hops away. Of course NFS requests to
the other city are slower but are just as reliable.
The Alpha systems that I've had problems with are 4 hops away and a 5
minute cab ride to that datacentre. Reducing the read and write buffer
size to below the Ethernet frame size fixed many timeouts. Tru64-UNIX
has problems with PATH MTU discovery. Between this and the fact that I
run a packet filter on my desktop system, to protect it from the rest
of my employer's network, making sure NFS packets aren't fragmented
helps a lot.
You might have some network issues, e.g. routers fragmenting packets in
some weird way or your hosts not reassembling the packets properly.
Try reducing your read/write buffer sizes.
Regards, Phone: (250)387-8437
Cy Schubert Fax: (250)387-5766
Team Leader, Sun/DEC Team Internet: Cy.Schubert@osg.gov.bc.ca
Open Systems Group, ITSD, ISTA
Province of BC
In message <39D303C5.1E440669@inpharmatica.co.uk>, Matthew Seaman
writes:
>
> Hi,
>
> We also have a mixture of FreeBSD and Solaris servers, and I had a very
> bad
> patch of this error occuring between our mailserver (FBSD) and our NFS machin
> e
> (Solaris). It turns out that we had a marginal ethernet cable, and replacing
> it has reduced the occurrence of such errors to once every few days (which is
> managable). The annoying thing was that the FreeBSD machine seemed to suffer
> a lot more from the fault than other machines around our network.
>
> I also found that NFSv2 rather than v3 seemed to work a bit better give
> n the
> dodgy connection.
>
> Andrew BOGECHO wrote:
> > We have recently been moving more and more to FreeBSD. The enjoyment
> > of such a stable platform allows us to see that we are making the
> > right decision.
> >
> > We do have one error message that shows up all the time. It does not
> > affect us or our users, but it does make me curious. I have searched
> > everywhere, and have only found one reference to this error, but no
> > repies.
> >
> > The error message is as follows:
> >
> > /kernel: nfs send error 32 for server blah:/partitiona
> >
> > The above error message occurs sometimes several times a minute, a
> > few times an hour, there is no preset frequency, but it is often.
> > It also affects all nfs mounted paritions.
>
> Matthew
>
> --
> Certe, Toto, sentio nos in Kansate non iam adesse.
>
> Dr. Matthew Seaman, Inpharmatica Ltd, 60 Charlotte St, London, W1T 2NU
> Tel: +44 20 7631 4644 x229 Fax: +44 20 7631 4844
>
>
> To Unsubscribe: send mail to majordomo@FreeBSD.org
> with "unsubscribe freebsd-stable" in the body of the message
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?200009281455.e8SEtsI40382>