AllHands2011/tbplPlanning: Difference between revisions

Latest revision as of 23:27, 7 April 2011

Planning meeting for evolution of tbpl during April 2011 All Hands

Friday 9am in C3PO.

I mainly use tbpl to, at a high level:

@@ Line 14: / Line 14: @@
 # finish up the post-push workflow (updating bugs with commit messages)
-== Can I Push? ==
+http://etherpad.mozilla.org:9000/tbpl-evolution
-* Are there unstarred failures in recent builds?
-* What are they? Should I worry about them?
-** Intermittent oranges
-** A failure in one build might be fixed in a later build
-** Some infrastructure failures are not autodetected as such
-** For the rest, eyeball them briefly:
-*** does it look intermittent?
-*** does it look like it was caused by that same push?
-*** can I identify the culprit to harass them over IRC? (what channel?)
-*** has philor looked at it? is he already agitating over it?
-* Is the build infrastructure operating? (Are the expected set of jobs appearing?)
-* How far behind is the build infrastructure? (tbpl-at-a-glance more convincing than magically derived "expected wait time")
-== Prep for a Push ==
-* Star anything not yet starred (star suggestions rule!)
-* Wait until scary jobs go through
-== Monitoring a Push ==
-* Watch for new failures on preceding pushes
-* Watch for failures on my push
-** Triage the failures
-** Trigger a backout
-** Cancel jobs (currently only via separate self-serve API)
-== Aftermath ==
-* Grab URLs for changesets to paste into bugs
-* Email build directory to people (or use myself -- other options, other platforms)