[26753] in Athena Bugs
sunblade 150s crash with corrupt gconf
daemon@ATHENA.MIT.EDU (jhawk@mit.edu)
Tue Feb 28 12:43:17 2006
Date: Tue, 28 Feb 2006 12:42:35 -0500 (EST)
From: jhawk@mit.edu
Message-Id: <200602281742.k1SHgZPw011240@m56-129-7.mit.edu>
To: bugs@mit.edu
X-Spam-Score: 2.178
X-Spam-Level: ** (2.178)
X-Spam-Flag: NO
Errors-To: bugs-bounces@mit.edu
Seen so far on m56-129-8 and m26-100-1. They hang after the afs cache
scan and nothing happens. I moved aside /etc/gconf to /etc/gconf.broken
and then they come back.
Quite painful to debug because the output of the startup scripts no
longer goes to the console. Instead it seems to go...nowhere? What's
the story with that? Is it a new bug, a deliberate behavior, or...?
I ended up having to use "boot kadb" and ::ptree to figure out
where things were hanging, etc. How do you kill a process from kadb/kmdb,
anyhow, oh Solaris mavens?
--jhawk
p.s.: Ironically, people seem to unplug the monitors in these cases and
plug them back in to the wrong video card. I wonder why, but it sure
doesn't ease debugging.
