Closed Bug 1256070 Opened 4 years ago Closed 3 years ago

Re-enable release-sanity.py in release-runner.py

Categories

(Release Engineering :: Release Automation: Other, defect)

defect
Not set

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: rail, Assigned: mtabara)

References

Details

We disabled release sanity at some point, but it would be better to enable it back to catch possible mistakes.

http://hg.mozilla.org/build/tools/file/old-release-runner/buildfarm/release/release-runner.py#l423 is how it is invoked in old world.
Assignee: nobody → mtabara
We had some more release sanity related issues in 48.0b3 - https://github.com/mozilla/releasewarrior/blob/master/releases/firefox-beta-48.0b3.md

* we should block the build if one of the ship-it partials build did not ship successfully to prevent situations like this 
(48.0b2build1 failed on one particular win64 chunk, we had a follow-up build2, but for 48.0b3 we used 48.0b2build1 which ended up serbing complete mars for users, instead of partial mars
* maybe consider blocking the build if there is follow-up for one of the specified partial builds (if used 48.0b2build1, fail and suggest there is a follow-up build2 or build3 to be used)
* ensure all previous mar files exist for each platform based on partial list submitted in ship-it
See Also: → 1282959
Would be also great to check if RC has both beta and release versions in partials, see bug 1265579
Blocks: 1265579
No longer blocks: 1265579
Depends on: 1282959
See Also: 1282959
This is done now as per bug 1282959. Release sanity in the non-relpro world from http://hg.mozilla.org/build/tools/file/old-release-runner/buildbot-helpers/release_sanity.py has been migrated to http://hg.mozilla.org/build/tools/file/tip/lib/python/kickoff/sanity.py.

Closing this for now all old sanity + some other checks have been migrated and turned on.
Status: NEW → RESOLVED
Closed: 3 years ago
Resolution: --- → FIXED
Awww: :)

Traceback (most recent call last):
  File "release-runner.py", line 413, in main
    validate_graph_kwargs(queue, gpg_key_path, **kwargs)
  File "release-runner.py", line 225, in validate_graph_kwargs
    sanitize_en_US_binary(queue, task_id, gpg_key_path)
  File "release-runner.py", line 190, in sanitize_en_US_binary
    validate_signatures(checksums, signature, tempdir, gpg_key_path)
UnboundLocalError: local variable 'signature' referenced before assignment
2016-07-18 18:54:23,077 - DEBUG - Releasing lock: /builds/releaserunner/tools/buildfarm/release/.release-runner.lock
Status: RESOLVED → REOPENED
Resolution: FIXED → ---
Aactually the issue wasn't that code, we found that add-on-devel builds were overriding win*-opt build indexes and those builds have not checksums, see bug 1287665.

however, the message is misleading. Probably we should rearrange the code to get rid of UnboundLocalError.

Feel free to close this bug (enough!) and address it in a separate bug.
I'll close this bug and add a TODO for bug 1287343 which is more suitable to address the aforementioned issue.
Thanks rail for brining this up - will take care of it as part of moving the en-US binary sanity into the new shiny main one.
Status: REOPENED → RESOLVED
Closed: 3 years ago3 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.