Path: utzoo!attcan!uunet!samsung!zaphod.mps.ohio-state.edu!ub!uhura.cc.rochester.edu!rochester!rit!cci632!mth
From: mth@cci632.UUCP (Michael Hickman)
Newsgroups: comp.sys.apollo
Subject: Re: SR10.3  (and a problem)
Message-ID: <41371@cci632.UUCP>
Date: 26 Oct 90 18:19:29 GMT
Reply-To: mth@ron.UUCP (Michael Hickman)
Organization: none
Lines: 32


	We received SR10.3 from Mentor Graphics on Monday (10/22).  I have
upgraded 6 of our 11 nodes so far.  

	Last night we had three nodes (one SR10.1, one SR10.2, and one SR10.3!) 
that would lock-up with the DM getting 97% of the cpu time (from dspst)
after displaying the 'Welcome to Domain/OS SR10.3' message in the message
window.  No shell, nothing.  I could 'crp' onto the node prior to logging
in on the display and work fine, but it would lock up after the attempted
DM login.

	I discovered that if our main file server (still at SR10.1) was not running
any daemons, this problem did not occur.  The server is normally running:
tcpd, inetd, rwhod, routed, mountd, portmap, nfsd, cron, sendmail, and lpd.
I tried bringing the system up 10 different times with one more of these installed 
each time, to see which one was causing the other systems to have problems, but 
once I got to the point that all the daemons were running, the lock-up problem
hadn't re-occured!!!

	I am suspecting the nfs daemons.  We have problems with them in the past, and 
have found that if machines that we have cross-mounted go down for backups, our 
server dies.  

	Any ideas???


Michael Hickman  CAE/CAD System Administrator   mth@cci.com
Computer Consoles, Inc.           
Rochester, NY

PS  Thanks to those that gave me advice on copying SR10.2 X11R3 to SR10.1.
	Since I just got SR10.3, I no longer need to try this...