stale nfs error on linux

[Binu_5]

Hi

I am facinng some problem with stale nfs error on linux systems ..The issue is filesystems(clearcase vobs) are exported from the hpunix system to the clearcase client system system ..whenevee I give a build it thows stale nfs error..amd deamon is running in both server and the client ..Can any help me in this issue ..

Thanks
-Binu

[Peter Godron · Accepted solution]

Binu,
how quickly does the nfs go stale?
Try re-starting the client daemon.
What re-try parameters have you set?

[Binu_5]

Hi

Can you please let me know the configuration files we should look into ..

Thanks
-Binu

[Peter Godron]

Binu,
on the HP box the file to look at is /etc/exports
When you do the nfs mount from the linux box check for options
retry=x (number of re-tries to mount)
timeo=x (timeout in 10th of secs)

[Ivan Ferreira]

Ensure that you don't have connectivity problems between your client and your server. Use udp as the protocol for your connection. You should check your /etc/fstab for mount options.

This error you do get all the time or after some time.

[Steven E. Protter]

Hmmm Clearcase vobs.

We have similar issues as well, but it was when our central storage was flakey.

Check the HP-UX systems /var/log/syslog/syslog.log files for events that might explain the problem.

There is most likely a problem on the server or interuption of networking that could be causing this.

I would install all NFS patches on the HP-UX servers and make sure you are running the latest supported version of NFS there.

Also, clearcase can be the issue, you may need to increase the default configuration of lock manager from default if you have a lot of vobs or very big vobs.

SEP

[Binu_5]

Hi

How to increase the lock manager ....

-Binu