[2402] in Release_7.7_team
Minutes of 2000-08-02 release-team meeting
daemon@ATHENA.MIT.EDU (Greg Hudson)
Wed Aug 16 16:43:36 2000
Date: Wed, 16 Aug 2000 16:43:28 -0400 (EDT)
Message-Id: <200008162043.QAA02282@small-gods.mit.edu>
From: Greg Hudson <ghudson@MIT.EDU>
To: release-team@MIT.EDU
Attending: ghudson aurora ajfox rbasch wdc jweiss
Penalty box: tb
1. 8.4.10 patch release
We identified the following problems which needed to be fixed:
* OpenGL support needs finishing
* Linux crontab needs to be fixed (again)
* Linux login options didn't include 8-bit login
* Some Linux RPM updates for security (unfortunately, we can't
take the new kernel without risking AFS incompatibility)
* IRIX version script needs to repeat the backout of the
disabling of the workaround for the processor bug.
2. Monitor resolution for Linux machines
Linux GX110 cluster machines with the new monitors come up at
1600x1200. Most of the release team feels we should change them to
use 1280x1024 because:
* The monitor documentation lists 1280x1024 as the recommended
resolution.
* Consistency per platform may be considered good. (The GX1
machines come up at 1280x1024.)
* The default font is very small at 1600x1200 on a 19" screen.
(We could maybe change this instead of the resolution, but
we're not sure how at the moment.)
Thomas disagreed, arguing that we should migrate to higher resolutions
as they become feasible rather than locking ourselves at 1280x1024,
that a variety of resolutions allows users of varying preferences to
all find machines they like, and that the monitor manual may recommend
1280x1024 on the basis that Windows is not as careful with sync rates
as X.
ghudson declared consensus with one outlier on changing to 1280x1024;
Thomas was given the task of implementing the change.
3. Starting without network
tb suggests that scripts should detect that there is no working
non-local interface and not start services. When network is detected,
the script that runs should start these services. This is a good idea
if it is in line with what Red Hat does.
Independently, we should fix stuff (gettime, zhm, AFS) to fail quickly
instead of slowly or not at all when there is no network. Of course,
we can't fix AFS, but we should probably ask Transarc to consider an
enhancement where they time out servers instantly if they get network
unreachable errors trying to contact them.