Re: OSX clients, Irix SGI server, NFS problem, rpcbind cores

From: Derrick MacPherson (dmacpher_at_vfs.com)
Date: 01/20/04 

Date: 20 Jan 2004 14:10:38 -0800

Max Matveev <max.matveev@sgi.com> wrote in message news:<netzncjqypf.fsf@kuku.melbourne.sgi.com>...
> >>>>> "dmacpher" == Derrick MacPherson <dmacpher@vfs.com> writes:
>
> dmacpher> I have a few (4 for now, more to come) OSX 10.3.2 clients,
> dmacpher> NFS mounting a directory from an O200 server, and several
> dmacpher> times a day, the rpcbind dies, and leaves a core file. I
> dmacpher> then have to start/stop the network, and all is fine. The
> dmacpher> way I find out it's died, is the OSX clients bitch about
> dmacpher> it.
>
> dmacpher> Any ideas? thoughts? things to try?
>
> Irix version? Why do you run rpcbind and not portmap? Where exactly
> does it die?

6.5.17f on that machine, i am just starting the big download now of
.22f
rpcbind - not sure why it's running, this box was built a long time
ago, I will investigate more. And where it dies... good question,
here's what pattern i've noticed in the last 3 days, when the osx
clients reboot, or a new one joins the network, they all end up losing
their NFS connection, and I need to start/stop the network and all is
fine. I find a core file, that is usualy time stamped with
approximately the time NFS stopped working, and the command file tells
me it's from rpcbind.