Closed
Bug 1380796
Opened 7 years ago
Closed 5 years ago
don't block scheduling of balrog rule changes on a human decision task
Categories
(Release Engineering :: Release Automation: Updates, enhancement, P1)
Release Engineering
Release Automation: Updates
Tracking
(Not tracked)
RESOLVED
FIXED
People
(Reporter: bhearsum, Unassigned)
References
Details
(Whiteboard: [releaseduty])
In bug 1347212 we did a quick hack to sometimes replace the "publish in balrog" task with a "schedule publishing in balrog" task. This is causing some confusion because we now essentially have two human breakpoints - the human decision task, and the signoffs on the scheduled change itself. We should change the release promotion graph to not block scheduling on a human decision task. Probably the best thing to do is to run it after everything prior to the human decision task has succeeded, but I could be convinced that we could do it a bit earlier (eg: maybe at the same time as test channel rule changes)
Reporter | ||
Comment 1•7 years ago
|
||
Something to be aware of when we do this is that scheduled changes will stick around if we abandon a build attempt. This means that we should do one of two things: 1) Have humans delete the scheduled changes when a build attempt is abandonded 2) Make automation smart enough to look for and update an existing scheduled change instead of trying to create a new one (which will fail).
Updated•7 years ago
|
Priority: -- → P2
Whiteboard: [releaseduty]
Updated•7 years ago
|
Priority: P2 → P1
Updated•6 years ago
|
Component: Release Automation: Other → Release Automation: Updates
Comment 3•5 years ago
|
||
This is no longer needed since we switched to Ship-it for production. RelMan holds the key now to push the release at their convenience.
Status: NEW → RESOLVED
Closed: 5 years ago
Resolution: --- → FIXED
You need to log in
before you can comment on or make changes to this bug.
Description
•