On Sun, Dec 19, 1999 at 12:00:38PM -0500, mcmahill@mtl.mit.edu wrote:
> 
> 
> I have a couple of NetBSD-1.4.1 machines on a network here.  The network
> uses NIS with a solaris NIS master.  Both NIS and AMD seem to work most of
> the time, but yesterday, when some of the solaris machines (an NIS master
> and also NFS server) went down and got rebooted, both NetBSD machines
> became somewhat hung.  I can log in, sort of.  I get the "NetBSD 1.4.1
> ..." kernel version message, and the "Welcome to NetBSD!", but no prompt.
> This is when I log in as root, so it shouldn't be looking for a remote
> home directory.
> 
> Then I just get messages like:
> 
> nfs server pid197@ballast:/u: not responding
> nfs server pid197@ballast:/u: is alive again
> nfs server pid197@ballast:/u: not responding

  Do you have your mail on /u, by any chance?  Depending on your
shell, some of them like to check your mailbox for incoming
mail for you, which can cause the shell to hang, if the mail
mount is flaky, IIRC.  From a quick ktrace experiment, it looks
like bash does a stat on /var/mail/root, but csh, sh, and tcsh
don't.  Of course, if your setup sets the tcsh/csh "mail"
variable, then csh/tcsh _will_ check.

  Also, does anything in your /root/.profile or /root/.cshrc
reference any non-local filesystems (like a shared /usr/local,
or /u/myfavoriteuser/bin)?

  I've found the best way to debug these problems is, get
everything happy again, start a ktrace -i -p <the pid for getty
on the console>, unplug the network from the machine, and log
in as root.  Once it hangs, do a kdump in another window so you
can see how far it has gotten, and then re-plug the network in.
Or something like that.

  Good luck!

  Brian