Releases/Post-mortems/Firefox 16.0

From MozillaWiki
Jump to: navigation, search

Schedule / Location / Call Information

  • Tuesday, 10/30/2012 @ 9:00 am PT
  • Mozilla Mountain View: Very Good Very Mighty, 3V
  • Vidyo Room: Release Coordination

Other communication channels

  • irc.mozilla.org #planning for back channel
  • This etherpad for real-time notes (we'll copy the wiki page in at the beginning of the meeting, and out at the end)

Talking Points

Timeline

16.0

  • Went to build 10/5
  • Pushed to mirrors 10/8
  • Released, throttled due to Patch Tuesday, 10/9
  • Pulled off the wire and from download pages on the morning of 10/10

16.0.1

  • Found out about the necessity for a 16.0.1 morning of 10/10 (bug 720619)
  • Went to build afternoon of 10/10
  • QA sign-off provided morning of 10/11
  • Released morning of 10/11, unthrottled
  • Throttled for feedback on 10/13

16.0.2

  • Starting discussing the necessity for a 16.0.2 on 10/18
  • Made a final decision on the necessary changeset 10/18
  • Landed the changes on branches and watched for regressions until 10/23 (17.0b3 go)
  • Went to build on 10/24
  • Released on 10/26, unthrottled

Release Issues

  • bug 799348 – [Firefox 16] load event is not always triggered by XUL windows
    • wontfix'd based upon Firebug workaround and lack of other majorly affected add-ons
  • bug 786740 – CSS transition with rounded borders cause rendering issues
    • web regression, but didn't have major user impact

QA

  • [Action from 15.0] - QA (ashughes) to add checks for no what's new page to Beta 1 update testing and to release update testing
    • This was done as of 16.0b1 -- manual spotcheck in place until we have an automated test
  • QA sign-off wasn't clear for 16.0.2, prevented a push to mirrors the evening before pushing out

RelEng

  • went well: hotfix signing, no fallout from patch tuesday, vanilla+esr+stub
  • room to improve: timing of "go to builds", colo outages, coord w/TB releases
    • need action before FF17.0 with double/overlapping ESR releases
  • "push to releasetest"... change since CDN. still checking, but expect change in process
  • [akeybl] PT (or later) time zone coverage for week 6 builds?
    • instead do "go to build" ~10am PDT, by not doing migration on friday
    • migration mozilla-beta -> mozilla-release earlier?
    • migration to happen as soon as b6 shipped (wed?)
    • [communicate the above at channel meeting]
    • 10AM Friday feedback deadline, to go to build
    • ship wed morning instead of tues morning? also avoids patch tuesday
  • not able to (reasonably) pull broken build from android marketplace
  • [akeybl] launch delayed significantly by late filing of bug 799473 – please update AUS to AUS2_PRODUCTION tag
    • 40mins-1.5hrs

Security

  • Providing context about how we're trying to prevent future 16.0.1/16.0.2 type releases
  • May want a security repo w/ private tests

Marketing & Press

  • Differing feedback for 16.0.1 and 16.0.2
  • From Product Marketing (unable to attend) nothing to note here other than the slight delay in launch time
  • Would be good to record guidelines for when and how we communicate future chemspills, security risks, updates, etc. [PR]

Other

Notes

  • [akeybl] how can we keep users installing the previous version of Firefox if we have to pull the product down? (mobile)
  • [akeybl/Pike] how to catch build failures in l10n early? See also bug 800432

Actions

  • RelMgmt - what disabling an Android product means, and a proposal for future instances (re-building the prior version with a higher buildid) - QA & RelEng scheduling impacted
  • All - before pulling the trigger on any releases, we need to make sure to have one final conversation
  • johnath - review of release-drivers for people who may not be appropriate to see