[819] in Release_7.7_team
Re: IRIX 8.0J release in dev cell; Solaris announcement
daemon@ATHENA.MIT.EDU (Greg Hudson)
Tue Jan 14 19:33:01 1997
Date: Tue, 14 Jan 1997 19:32:57 -0500
From: Greg Hudson <ghudson@MIT.EDU>
To: brlewis@MIT.EDU (Bruce R. Lewis)
Cc: rel-eng@MIT.EDU, ops@MIT.EDU, release-team@MIT.EDU
In-Reply-To: "[818] in Release_7.7_team"
> Don't anybody propagate this to the athena cell yet.
(It would wait for the release team anyway.)
> Mike's machine (megara) is endlessly rebooting; the release
> shouldn't go out until we know the same won't happen in the
> clusters.
Mike's machine was missing the CellServDB and SuidCells files from
/usr/vice/etc. Nothing in the 8.0J update touched those files.
/usr/athena/lib/update/do_update does do a certain amount of
unnecessary fiddling with /usr/vice/etc, but given that Mike's machine
is not PUBLIC, I can't see how it would delete those files.
Other machines seem to have taken the update with no problems.
Verdict:
* This update is probably about as safe as any other update;
the problem could easily recur on other people's machines,
but I don't think it's any more likely that it will now than
that it would on previous updates.
* We really need to be logging the output of do_update so we
can figure out why these things happen.
To fix the problem:
* Boot the machine single-user.
* Do:
cd /usr/vice/etc
ls
Verify that the CellServDB and SuidCells files are missing.
If they're present, the problem is probably something else.
cp CellServDB.public CellServDB
cp SuidCells.local SuidCells
exit
* Choose runlevel 2. The machine should fix the CellServDB
and SuidCells files when it runs config_afs.
Assuming we go ahead with the update, we should probably make sure
cluster has these instructions and understands them.