[7447] in Release_7.7_team
Minutes from Release-team, April 1, 2pm, N42-286
daemon@ATHENA.MIT.EDU (Jonathan Reed)
Thu Apr 7 15:34:01 2011
Mime-Version: 1.0 (Apple Message framework v1084)
Content-Type: text/plain; charset=us-ascii
From: Jonathan Reed <jdreed@MIT.EDU>
In-Reply-To: <B6E1C0F9-596C-4891-9D05-6E98C7A86232@mit.edu>
Date: Thu, 7 Apr 2011 15:33:53 -0400
Cc: Benjamin Kaduk <kaduk@mit.edu>
Message-Id: <914C39E3-29E3-451B-A532-14B51281BB1D@mit.edu>
To: "release-team@mit.edu Team" <release-team@mit.edu>
Content-Transfer-Encoding: 8bit
> 1. Unity (or lack thereof)
Someone should investigate whether or not we can do something clever with our own gnome-session, though I suspect we still get Unity. Also, support for ~/.xsession is now back, so we should investigate that. Do we want to run a different WM?
At the very least, we should ensure that we don't get the Dock/Launcher/whatever by default.
> 2. Testing needed
Please test stuff in -development. I'm going to start pushing things I've tested on at least 2 separate cluster machines, because there's stuff we're actively losing on (like when someone hard-boots a machine in the middle of an update).
> 3. clean-tmp-areas (alexp)
An (MIT-developed) app writes a cache in /tmp, and then automatically assumes that it will be able to access that cache, and fails if the cache is owned by the previous user. Discussion ensues, and it is believed that no known distribution cleans /tmp after each logout, so the app is broken for assuming that only one user will be using the machine. The app should be patched to use user-specific directories in /tmp. Alex will follow up with the developer, though if we get no response, we can apparetly walk over to 32 and go knock on their door.
> 4. Review of release timeline and adjustment if necessary
Due to staffing changes, we currently have no IS&T developers assigned to Athena, which is ... not good. jdreed will follow up with rsmurphy.
> 5. Moving stuff from the dev cell to a de-luxe apartment in the sky-y-y (amb)
amb will inventory stuff in the dev cell, and send a proposed list of things to punt to release-team.
jweiss would prefer it if we didn't move completely useless stuff (like a local source tree from andrew-8.0, for example), but we shouldn't put too high a bar. At some point, hesiod cluster info will be updated so any remaining 9.4 dev cell machines will point at the Athena cell.
> 6. Other business
jweiss installed debathena-login-graphical from PXE and ended up without a panel both times. We currently blame cosmic radiation bit flipping, since I can't reproduce. jweiss will try a liveUSB stick (using "usb-creator-gtk" to turn an ISO image into a USB stick) and then using the debathena installer.
