Please provide a method to process urgent updates more quickly

[Jamie Strandboge]

In reviewing the timeline in the recent pam regression incident report (https://wiki.ubuntu.com/IncidentReports/2011-05-31-pam-security-update-breaks-cron), I noticed that there was a problem getting packages to build on hardy/amd64 because of a hung build which resulted in the update being delayed by approximately an hour and a half.

It would be nice if key individuals could have a method to say "this update is so important, if needed please kill other builds (and restart them later) to make room for this one". I'm not sure how this would be achieved, but it seems ACLs could be involved (perhaps the TB, perhaps individuals). Obviously this functionality should only be used in the most urgent situations....

[Martin Pool (mbp)]

As a workaround, I believe that TB members can somewhat do this now by asking Launchpad operators to kill other builds. (But this will be messy, because it won't guarantee the important build gets to the top, and it is not a very clean solution.)

[Julian Edwards (julian-edwards)]

Tricky fix with the current model/code because we don't have a way to do binary-only copies once the killed builds get restarted. This is also a dupe of another bug somewhere but I can't find it.