Define a process for handling of build failures in the LP buildd network
Traditionally, build failures have been handled in a haphazard way; sometimes I will just fix the bug (this will likely still continue), sometimes I will file a bug, sometimes I will poke someone on IRC to fix the bug, and sometimes the failures will just go unnoticed when I'm too busy to check logs, and I won't find them until weeks later.
This spec should be a combined Ubuntu/LP spec (perhaps can be split later) to define A) a method for soyuz notifying the Right People when a failure has occurred, so I don't have to read all the build logs myself and go insane in the process, and B) a process for properly dealing with and clearing build failures off our plate.
Blueprint information
- Status:
- Complete
- Approver:
- Matt Zimmerman
- Priority:
- Essential
- Drafter:
- Adam Conrad
- Direction:
- Needs approval
- Assignee:
- Celso Providelo
- Definition:
- Approved
- Series goal:
- None
- Implementation:
-
Implemented
- Milestone target:
- None
- Started by
- Adam Conrad
- Completed by
- Celso Providelo
Whiteboard
cprov 23/06/06: straighforward specification, ready for review.
iwj 23/06: 12:00: seems okay
cprov 04/08: base system implemented, pending code review in my `build-
cprov 14/08: target to Soyuz-1.0
cprov 10/10: running in production
