I originally asked:
> I've got an old IPX running Solaris 2.5.1 with the latest Recommended
> Suite. This is a machine at a remote site, so I administer it over
> the network. Unfortunately I am unable to reboot it remotely. If
> I do either "/etc/reboot" or "/usr/sbin/shutdown -g0 -i6" the machine
> shuts down but never comes back up again. Someone must go in to the
> room with this box, do an L1-a, and type "boot". This always works
> without a problem, I might add. I have been told by the folks who
> go in to look at the box that there are no obvious error messages on
> the console.
>
> I don't have this problem with any of my remote SS5 boxes, and this
> problem existed before and after I applied the June 1998 Recommended
> Suite.
And I added shortly thereafter:
> From the responses I see that I need to be more explicit. The
> auto-boot? eeprom value is set to true, but that isn't the problem.
> The problem is that the box never reaches the "ok" prompt. The OS
> appears to shut down (the box is not pingable), but the "ok" prompt
> never appears and the box does not come back up. The person who walks
> in to the room to sit at the console for me must type an L1-a in order
> to get any response from the system.
The suggestions:
make sure auto-boot? is set to true.
Make sure the s6 entry in /etc/inittab looks like this:
s6:6:wait:/sbin/rc6 >/dev/console 2>&1 </dev/console
Make sure you have a corresponding rc0.d/Kxxx script for every custom
rc2.d/Sxxx and rc3.d/Sxxx script.
Add verbose comments to all rc?.d/Kxxx scripts to see how far the system
gets before hanging.
Add a "set -x" to the top of /sbin/rc0 to watch all of the rc?.d/Kxxx
scripts start off. This won't debug the individual scripts, but it
will quickly and easily tell you which script is the culprit.
It looks like I'm going to have to plan on taking a fieldtrip. The next
time I'm down at the site with this box, I'll add the "set -x" to my rc
scripts in /sbin and watch the system shutdown. I got several "me too"
emails, so I'm not alone in this problem. Unfortunately this isn't a
simple known problem, so I'm just going to have to do enable some verbose
logging and watch the system.
Thanks to:
Jim McVey <jmcvey@nhgri.nih.gov>
reynolds@illustra.com (John Reynolds)
Chenthil Kumar K G <chenthil@mtcts1.ho.lucent.com>
"Mark A. Baldwin" <mark.baldwin@makesys.com>
djohnson@nbserv1.dseg.ti.com (Danny Johnson)
Carlo Cosolo <ccosolo@tse.com>
Larry Zins <lzins@lucent.com>
Bismark Espinoza <bismark@alta.jpl.nasa.gov>
David Schiffrin <daves@adnc.com>
"V. Q. Hoang" <vqh@dw.lucent.com>
Dieter Gobbers <gobbers@faw.uni-ulm.de>
+-----------------------------------------------------------------------+
| Christopher L. Barnard O When I was a boy I was told that |
| cbarnard@tsg.cbot.com / \ anybody could become president. |
| (312) 347-4901 O---O Now I'm beginning to believe it. |
| http://www.cs.uchicago.edu/~cbarnard --Clarence Darrow |
+----------PGP public key available via finger or PGP keyserver---------+
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:12:45 CDT