[ILUG] Too many NFS ops between RH 5.2 machine and F740 filer.

[Dave Airlie]

the only thing I can see of note is 2.0.38 may not be the best NFS client
in the world ever....

I've no proof of this it is merely provided as conjecture.

Dave.

On Sat, 5 Jan 2002, Richard Archbold wrote:

> Hi Guys,
>
> I have a mail server which NFS mounts its spool and conf files from a NetApp
> 740 filer. The filer has about 3 times as many NFS ops as of the other
> filers that I look after and when I do a nfsstat -l on the filer I get the
> following line:
>
> 10.0.0.214      skinner-e0      NFSOPS =      20970 (54%)
>
> I have other mail servers of similar spec doing similar things that don't
> create half as many NFS ops. I have check /var/log/messages and any other
> log files I can think of but can't see any errors.
>
> The machine itself is newish, it is an Encom machine, PIII 500 MHz, with
> half a gig of RAM and an 8 Gig scci hard disk. The machine is running Red
> Hat 5.2 with kernel 2.0.38. The version of ONTAP on the filer is 5.3.1. The
> mount options are as follows:
>
> rw,nolock,hard,intr,rsize=4096,wsize=4096
>
> I guess what I am looking for is someway to debug the NFS packets, are some
> of them being refused/corrupted/re-sent ? Also, can someone tell me what the
> NFS client software is on the RH machine ? Would upgrading the version of
> ONTAP help alleviate the problems or provide me with any better tools for
> debuging the problem ? I am a bit new to troubleshooting NFS problems so if
> some of the questions seem a bit silly I apologize in advance.
>
> Anyway, all help is appreciated.
>
> Regards,
>
> Richard.
>
>
>

-- 
David Airlie, Software Engineer
http://www.skynet.ie/~airlied / airlied at skynet.ie
pam_smb / Linux DecStation / Linux VAX / ILUG person