[2996] in Release_7.7_team
An update-related outage in DSpace
daemon@ATHENA.MIT.EDU (Bill Cattey)
Wed Oct 3 16:57:28 2001
Message-ID: <AvirfyVz0001I3WnAg@mit.edu>
Date: Wed, 3 Oct 2001 20:54:22 +0000 ()
From: Bill Cattey <wdc@MIT.EDU>
To: release-team@mit.edu
One of the HP systems that survived many milestones in Athena land (IE
it installed 8.4 ok, and upgraged itself to 9.0 ok) had a problem after
the 9.0.17 debacle. The XF86Config file was screwed up somehow. (And
stupid me, while I was testing, I did not save the old broken one.)
After I remembered to work-around how we turn off all virtual consoles
by doing 'linux single' at lilo, and after remembering what to do to
inittab to enable some virtual consoles so I'd get enuf of a system to
work with, I ran Xconfigurator by hand, and AFTER REBOOTING (telinit q
did not respawn xdm), I was back in business.
Is there SOME way we can do a better job of leaving a way to log in if
xdm fails to start X?
Perhaps we should re-consider our decision to make Linux like Athena,
and turn off virtual consoles?
Let's do please keep in mind that system updates may be needlessly
running the Xconfigurator program and that that buggy piece of dung may
damage working config files.
-wdc
