Re: kernel: nfs: server not responding,timed out

Mostly all I have are questions, in response to your query.

It sounds like transient network problems or perhaps a
server-side performance issue.  You didn't say, so I'm
assuming that this isn't only for specific files.  If it
is could it be a size issue?  What dies 'ulimit -f' say
on the client side?

Have you tried checking the status of each copy and
retrying the transfer?  You also referred to the seawolf-
list posting but didn't say whether you tried what was
suggested.

Have you checked with netstat?  nfsstat? Can you run
sar so that you could get a better picture of what is
going on on each end of the transfer?  Can you watch
'top' on the server-side while transfers take place?

Is your nfs mount 'soft' or 'hard', with or without 'intr'?

Regards,
Doug Wyatt

Dan Haines wrote:
> Hi
> 
> We are running RedHat 6.2 on several high end machines in New Zealand,
> at the MetService.  The reason for running 6.2 is that the model we use
> to run weather reports, only works (at the moment) on the 6.2 kernel. 
> We have slightly tweaked all our kernels to 2.2.17-14smp.
> 
> We get errors occasionally, which occer when a machine called "ajax"
> tries to read a file on a readonly nfs share on a machine called "lisa".
>  The job we have running, reads multiple files from this read only nfs
> share.  Something goes wrong, and when this job reads a file from the
> share, it reports that it cannot find the file, and carries on with the
> next file in the job list, from the same nfs share, without a problem. 
> This would seem that the nfs share temporarily times out, which results
> in the following message appearing in /var/log/messages:
> 
> Sep 16 18:00:58 ajax kernel: nfs: server lisa.met.co.nz not responding,
> timed out 
> Sep 16 19:01:15 ajax kernel: nfs: server lisa.met.co.nz not responding,
> timed out 
> 
> So ajax tries to read from lisa via nfs, cant find the file, our program
> generates an error saying the file doesnt exist (when it clearly does),
> and the above entry is added to the messages file on ajax.
> 
> Any help, or suggestions on extra settings to set on the nfs share to
> avoid this problem would be greatly appreciated.  I have searched around
> the net for similar problems, and found this thread, which matches my
> problem as close as I can find:
> 
> http://www.redhat.com/mailing-lists/seawolf-list/msg32253.html
> 
> Thank you very much in advance,
> 
> Kind regards,
> Dan Haines
> 
> Systems Administrator
> MetService,
> New Zealand.
>
-- 

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Doug Wyatt     <[email protected]>
Sys Admin - Kohlman Systems Research, Inc.
319 Perry St.,  Lawrence,  KS  66044   USA
Phone: 785-843-4099      Fax: 785-843-6459