[6610] in Release_7.7_team
Release-Team Minutes, 1/29/10
daemon@ATHENA.MIT.EDU (Jonathan Reed)
Fri Jan 29 20:50:06 2010
From: Jonathan Reed <jdreed@MIT.EDU>
Content-Type: text/plain; charset=us-ascii; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit
Date: Fri, 29 Jan 2010 20:49:56 -0500
Message-Id: <E26B6857-26A6-402B-A16A-5AAE2CAC07DB@mit.edu>
To: release-team@mit.edu
Mime-Version: 1.0 (Apple Message framework v1075.2)
Our next meeting is 2/12 at _3:00pm_. The meeting time will be moving
to 3:00 for the duration of the Spring term.
Attending: rbasch, geofft, mmanley, amb, othomas, jdreed, alexp,
aurora, broder
Agenda:
1) Server Status
2) OpenGL is broken
3) Printing update
4) New SIPB devs
5) Documentation
6) Moving AFS to DKMS
7) The Auto-Updater
8) Dialup update
9) Karmic Update & Ticket Review
10) debathena-announce e-mail
1) Server status:
zulu (the build server) is currently paid for by ISDA, and is not
funded past 2/15.
ACTION ITEM: Oliver will send new cost object to ops and CC amb.
demeter (athena10.mit.edu) is really only there for the installer,
bceause it needs a real machine with a real IP. Maybe we should punt
it and SIPB will move debathena.mit.edu to a real IP.
ACTION ITEM: broder/geofft will look into this.
2) OpenGL:
It fails to repaint the screen, at least in Mathematica. The failure
mode varies between hardware types. Alex has a workaround involving a
locker install of Mesa. We will check on the status of fglrx (the
ATI binary drivers) in Karmic and see if they work better than in
Jaunty.
3) Printing update:
We haven't yet moved to printers.mit.edu, because apparently header
pages don't work well (users get 2, with a blank page between them).
We think this is a CUPS 1.3 vs 1.4 issue.
ACTION ITEM: broder will send jdreed the data to reproduce the
failure, and jdreed will test this on Karmic
amb notes that cview doesn't necessarily reflect reality. perhaps
CUPS processes jobs faster than cview refreshes.
4) New SIPB students?
Evan and Geoff will be following up with some interested students and
scheduling some sort of brain dump.
5) Documentation
We have lots of documentation repositories: OLH, the IST site, the
Debathena site, Hermes, the SIPB wiki, etc.
Going forward: debathena.mit.edu should host no docs, only links to
other docs. Most docs should go in Hermes, in the ist-contrib space
(which debathena-dev folks have bits to edit).
ACTION ITEM: jdreed will deal
6) AFS -> DKMS
We want to switch to DKMS, but backports are required. If not, the
auto-updater fails stupidly and silently. So we should hold off a bit
until the auto-updater is friendlier. There's also the possibility
that we can put the necessary packages in the openafs component.
ACTION ITEM: broder will look into the feasibility of the latter
7) The auto-updater
It should offer better feedback to users. The relevant tickets are in
Trac.
8) Dialup update
Mark has deployed the patched sshd on test.dialup, which prompts for a
password if the user does not delegate credentials. However, it
doesn't appear to work in all cases. Both geofft and a consultant
found that they could still get logged in without delegating
credentials.
ACTION ITEM: More debugging and testing.
9) Karmic Update and Ticket Review:
We need to think about desync'ing the re-installation of machines a
bit more than normal values, since there will be considerably more
bandwith used than a normal update.
ACTION ITEM: amb will consider this
We need to update the PXE installer, once we're ready to go live
(until then, we need to preserve the ability for Hotline to access the
Jaunty installer for reinstalls). At the same time, we will ask NIST
to remove the AThena 9 entry. WE'll also ask for an increase of the
"Press F8" timeout value, since it's kind of short.
10) The debathena-announce e-mail should go out early next week.
