Mobile/Testing/02 12 14

Previous Action Items

• snorp will talk to jmaher about talos improvements to find gradual, long-term regressions
• gbrown will investigate slow Android 2.3 reftests
• gbrown will follow-up on startup crashes with :glandium
• jake will ask around about bug 967452

Status reports

Rel Eng

Thanks to gbrown:

• deployed bug 962743 - Report complete logcats for Android 4.0 panda jobs
  • Android panda jobs will report full logcat to blobber at end of test run!

A Team

• tegra failure rate: [1.5%]
  • last week 6.46
• panda failure rate: [14.19%]
  • last week 16.42
  • excluding retries 0.75%

Android 2.3 emulator

• See "Other Opt" at https://tbpl.mozilla.org/?tree=Ash&jobname=Android.*&rev=7da8241436d0

• [dminor] looking at slow reftests
  • plan to profile emulator to see where the time is going - likely software renderer is slow, but maybe something else will show up
  • will also look at reftest harness itself to see if anything can be tweaked there
  • open to suggestions

Autophone

• Turned off ipfw on 2014-02-08. It appears that I had previously configured ipfw incorrectly on 2014-02-05 which had already disabled traffic shaping.
• bug 968905 in review. Will allow partitioning devices across repositories and should help with keeping up with the daily builds.
• Production Autophone down due to an OSX only regression from {bug|965557}}. Investigating and hope to have back up today.

Eideticker

• Lots of use of the eideticker machines by the Android team to track startup perf regressions
• Various bugs fixed (or about to be fixed) in automation that were causing unreliable results: bug 970151, bug 970110