HELP - NFS Errors

[Emil Henry]

Hello!

I am getting the errors below on a k460 running HP-UX 11.11. A web server (Iplanet Enterprise Server 6.0 sp 5) is running on this machine. The errors seems to go away when the web server is shutdown, or restarted. Therefore pointing to the possible problem location. It is impossible to get a decent answer from Sun support within a reasonable time. Any thoughts??

Thanks in advance,

Emil

syslog.log
Jan 3 12:31:24 omega vmunix: NFS read failed for server nfs1: RPC: Timed out
Jan 3 12:36:16 omega vmunix: NFS read failed for server nfs1: RPC: Timed out
Jan 3 12:42:38 omega vmunix: NFS read failed for server nfs1: RPC: Timed out
Jan 3 12:43:05 omega vmunix: NFS getattr failed for server nfs1: RPC: Timed out
Jan 3 12:43:38 omega vmunix: NFS getattr failed for server nfs1: RPC: Timed out
Jan 3 12:57:32 omega vmunix: NFS read failed for server nfs1: RPC: Timed out
Jan 3 12:59:03 omega vmunix: NFS read failed for server nfs1: RPC: Timed out
Jan 3 13:13:00 omega vmunix: NFS read failed for server nfs1: RPC: Timed out

[T G Manikandan]

check this hp doc

http://www1.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000062912242

[James Murtagh]

Hi Emil,

Is it possible this is just down to an increase in network traffic as more connections are made to the web server? Have NFS and Iplanet been working ok previously? Has anything changed at all recently, i.e. speed or duplex settings of cards, patches installed, kernel reconfig?

Other than than if you run "nfsstat -c" on omega and print the results, it may yield more clues. Other information of interest would be mount options and protocol used for NFS and a description of the network configuration.

Also, what does "sar -b 2 5" show when both are running?

Regards,

James.

[Ravi_8]

Hi,

check the whether you can access the NFS server from the client (ping/telnet). if you could access restart the NFS services on server(#/sbin.init.d/nfs.server stop/start
#/sbin/init.d/nfs.client stop/start)

if you are getting the same error again restart the NFS services on client end also

[T G Manikandan]

I would also suggest you to check whether there is a speed and duplex mismatch between the server and the switch to which is connected.

Just go to lanadmin
do a
'dis'
then
'lan'
check whether you can find any collisions

REvert

[U.SivaKumar_2]

Hi,

The problem may be because of severe network congestion or lack of memory for network buffering.

With webserver and nfs server running.

run this command.

#top

What is the free memory shown ?
what all processes are in running state?
what is the CPU load ?

then

#swapinfo -m

what is the output

Now stop only the webserver

#top

what is the free memory now ?

#swapinfo -m

what is the output ?

run this

#netstat -s

paste the output here.

Use glanceplus to analyse the network resource utilization,

if you have enough memory and cpu load is normal with webserver and nfs try this and get back with results

stop nfs server. then

#nfsd 25

regards,

U.SivaKumar

[Steven E. Protter]

This looks a little like a pfs_mount problem.

Are you or were you mounting a cd from oracle that requires rock ridge format.

Typically, if you have mounted this cd and not unmounted it proberly these errors could occur.

I'm also intersted to know what version of NFS you are running and if you are current on patches.

Steve

[Niraj Kumar Verma]

Hi,

To me it looks to be NFS version problem

can you see that Solaris box is using NFS 3 as HP-UX 11.11
???

you can also try

ndd -set /dev/ip ip_path_mtu_discovery 0

on sun box.

Regards
-Niraj

[Jerry Shen_1]

To Steve:

I have similar problems with pfs mounted cdrom. I stopped the pfs daemons before I unmount the cdrom. Then whenever I try to list or bdf, NFS server error message came back. Do you have a solution other that reboot or restart NFS server daemon?