Tjo!

We've got 4 diskless Sparc ELCs booting of a diskfull Sparc-1,
swapping to a linux box, and mounting filesystems both from
Suns running 4.1.x and DEC Vax running Ultrix something or other.
This works quite nice except for one thing.

When I went current (first 10 december, later 10 january)
on these boxes I got a problem with the diskless machines
hanging on mount -a -t nfs in the rc script.

After that, ping works but nothing higher.

The funny thing is that if I reboot to single I can
mount all filesystems, but when going multiuser
the machine irrevocably hangs.

Replacing /sbin/mount_nfs with the old (Apr 26) copy
the machines boot up just fine.

When I looked in doc/CHANGES it said:
        ...
        change default for nfs mounts to not do a connect on the socket.
                Add option -C to force a connected mount. Now option -c
                is still there, but it is the default. Maybe amd should
                be changed similarly. [christos 961027]
        ...

I'm thinking that the problem is with some of our nfs-mounts
to the SunOS or Ultrix boxes and that they use the old default.

Does this mean that I have to edit /etc/rc and add the -C
switch to the mount command, or do I change the line to be
mount_nfs -a -t nfs -C, or is it perhaps something else?

I'd be grateful for a hint, I'd like some moral support
before I try trial and error...:-)

/^JN - Jesper Nilsson
--
              I've heard of UNIseX, but I've never had it.
      Jesper Nilsson -- dat92jni@ludat.lth.se || jesper@df.lth.se