NFS stale file handle errors on network mounts

Hi,

I have a cluster of about 20 macs all auto-mounting home directories (to /home) and also /usr/local via NFS. (Don’t ask me about /usr/local – I didn’t set it up that way.) This all works fine in most situations, but occasionally I get errors like ‘/usr/local/bin/perl : bad interpreter: stale NFS file handle.’ When I run the command again, it works.

An interesting tidbit: the error is almost always when accessing perl. Other binaries don’t seem to cause the problem. However, a few times I have gotten it when accessing files on the /home mount also, but this is much more rare.

I’ve tried turning off automount and manually setting mount parameters – no change.

Suspecting timeout problems, I ran a cron job on three machines to access a script every ten minutes. This produced intermittent NFS errors on all three machines. The interesting thing is that the errors occurred on all three machines at exactly the same time. This leads me to suspect a network problem, where the share is temporarily unavailable.

But before I start looking into that, I thought I’d throw it out there to see if anyone else has seen this kind of issue, or can think of any other explanations.

Thanks,

Duncan