in Release_7.7_team
Minutes form 8/4 release team meeting
daemon@ATHENA.MIT.EDU (Jonathan Reed)
Tue Aug 4 15:44:24 2009
Cc: "firstname.lastname@example.org" <email@example.com>
From: Jonathan Reed <jdreed@MIT.EDU>
To: Jonathan Reed <firstname.lastname@example.org>
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Mime-Version: 1.0 (Apple Message framework v919.2)
Date: Tue, 4 Aug 2009 15:44:12 -0400
> 1) Logins are still slow; adding RAM to the HPs does not noticeably
> improve things
> a) Should we or should we not do something clever with the AFS
> cache and snapshotting?
We should investigate aufs, and ramdisk. We should also investigate
the details of the problem. Note that stock Jaunty on an HP is still
slow (due all the GNOME stuff), but -cluster makes it noticeably slower.
> 2) We should start thinking harder about distribution upgrades for
> the clusters.
> a) do-release-upgrade worked for Intrepid->Jaunty, but asked me
> every conceivable debconf question ever, and took an order of
> magnitude longer than a reinstall
jdreed will investigate how long an upgrade takes vs a reinstall.
> 3) Bugme is getting hidden by Nautilus.
There's a ticket open in Trac. Can we perhaps add window manager
hints? This is a blocker for further quickstation deployment.
> 4) Cluster deployment update
> a) Changing the IP address of a cluster machine is harder than it
> should be.
We can create a script to edit /etc/network/interfaces, /etc/hosts,
and /etc/hostname and store it locally.
> b) Gaining access to a broken cluster machine is harder than it
> should be
> I) Can we do something awesome/evil with remctl to easily enable
> debugging without having to reboot into single user mode repeatedly.
No, we can't, and we should just be sad.
> 5) Meeting schedule during term
We will meet on 8/18 and 9/1, and then pick a schedule for term.
> 6) Other business
Bugs were triaged.
IAP is closer than we think it is.