NFS problems with RH 9

Hoffmann, Tony Tony.Hoffmann at
Wed Aug 20 03:12:00 CDT 2003

Are the patches on your Solaris machine up to date.   There are issues
with solaris client->linux server which require solaris nfs patches.

-----Original Message-----
From: Seth Mos [mailto:knuffie at]
Sent: Tuesday, August 19, 2003 12:09 AM
To: Venkatesh Nagarajan; linux-poweredge at
Subject: Re: NFS problems with RH 9

At 07:21 18-8-2003 -0700, Venkatesh Nagarajan wrote:
>I have imported this disk on a solaris 6 machine. When I invoke the copy 
>on the linux disk from solaris the performance is horrible. I get 
>something in the order of 2K/sec.

Make sure that the network cable is OK. NFS uses udp traffic instead of 
tcp. TCP traffic copes well with bad cabling while udp starts behaving 
erratic and will quickly reduce nfs traffic upto around 1Mbit or so.

Make sure with ifconfig (on both hosts) that you are not losing packets. 
Also check if the switch isn't behaving weird.

Furthermore you can increase the networkbuffer  size as listed in the NFS 
Howto. Newer kernels (2.4.19+) do this automatically.

You can experiment with the wsize and rsize settings as well as trying the 
mount with nfsv3. Although I believe this is problematic between a linux 
and a solaris host in some combinations.


It might just be your lucky day, if you only knew.

Linux-PowerEdge mailing list
Linux-PowerEdge at
Please read the FAQ at or search the list
archives at

More information about the Linux-PowerEdge mailing list