Luke Diamand wrote:
>
> Johannes Nix wrote:
> >
> > I have a running workstation wich is NFS client and also server (SMP
> > Pentium II / 2.2.5 Kernel). Normally everything works.
> >
> > Sometimes, about every few days, I am not able to mount files
> > with nfs anymore. This happens with the local filesystems, with
> > filesystems on IRIX and Linux Hosts.
>
> Strangely, I've just started seeing a similar problem, perhaps related,
> perhaps not.
>
> As far as I can tell, I find that after a while, when automount tries to
> mount a directory for me, I end up with _2_ automounters running on the
> same mount point with both of them stuck. Killing the second one gets
> things back to life.
>
> ps auxl gives me:
>
> 40 0 848 1 1 0 944 532 do_pipe_rea S p0 0:00
> automount /usr/site yp auto_site rw
>
> 140 0 916 848 1 0 1040 644 do_pipe_rea S p0 0:00
> automount /usr/site yp auto_site rw
>
> Killing PID 916 makes it all better.
>
> The only things I've changed recently are upgrading to 2.3.10 (but
> downgrading to 2.2.7 makes no difference) and adding
> 'rsize=1024,wsize=1024' to localoptions (to workaround a broken bridge).
> I'm using autofs 3.1.3.
>
> Any ideas?
>
autofs frequently spawns worker processes as a matter of course; that's
why you're seeing two processes. In this case, it looks like 916 is
going bonkers... furthermore, it's doing it in a way that is paralyzing
the main process.
Please set syslog to log *everything*, and send me any autofs-related
log entries.
Are you using stock autofs?
-hpa
