Mobile/Testing/04 10 13

From MozillaWiki
Jump to: navigation, search

Previous Action Items

  • wlach - Report back on regression
    • Done: see status
  • jmaher - test new power supplies

Status reports

Dev team

  • gbrown may not be able to attend this week / don't wait for me!!
  • Reported bug 858781 Investigate limited ts improvement with android pgo
  • Looked into xpcshell failure; thanks :Callek and :jmaher
  • Note that bug 856621 is playing havoc with robocop stability, but gradually being resolved
  • Fixed bug 833765 Intermittent DMError: Automation Error: DeviceManager isdir returned null -- part of the panda puzzle
  • Still investigating panda robocop failures

Rel Eng

  • (Callek) - fixed bug 858865 - Android xpcshell failing with Automation Error: error pushing file: Automation Error: Agent Error processing command 'push /mnt/sdcard/tests/xpcshell//netwerk/test/httpserver/test/data/sjs/sjs 72'; err='Push failed!'
  • bug 828069 enabled xpchell tests on other branches, previously was just cedar
  • next: bug 725703 tests should figure out the app name for fennec instead of hardcoding everywhere
  • (kmoir) setup foopies in rack 4 for jmaher and dividehex to conduct more power supply testing

IT

  • Still working on a higher density chassis, parts coming in to build a prototype.
  • Working on Panda stability issues. bug 860028 Replacing 5v supply wire and adj power supply output in panda chassis in scl1

A Team

  • total failure rate:
    • tegra [4.44%]
      • m1 is biggest offender
    • panda [23.85%]
    • panda/hidden [5.96%]
      • ignoring rc1/rc2 and m1
  • panda reboot investigations with Jake
    • 6 chassis running smoketests at full speed and just changing out the wiring to the fuse box
    • foopy102: pass: 711, fail: 31, %fail: 4.2%
    • foopy60: pass: 564, fail: 18, %fail: 3.1%
    • foopy61: pass: 542, fail: 18, %fail: 3.2%
    • foopy62: pass: 465, fail: 24, %fail: 4.9%
    • foopy63: pass: 516, fail: 27, %fail: 5.0%
    • foopy64: pass: 425, fail: 17, %fail: 3.8%
    • aggregate: pass: 3223, fail: 135, %fail: 4.0%
    • ignoring 5 pandas with a 10-12% failure rate, we have a 3.15% total failure rate
  • panda running Android 2.3.x
    • Have Linaro 11.10 (Android 2.3.5) running on panda, able to run mochitests and robotium stably using SUTAgent

x86 automation

  • able to run mochitests on android x86 emulator using SUTAgent

Autophone

  • Data is going into phonedash
  • Putting cloverview+ phone from intel into phonedash this week and next
  • We have lots of data going back to 2012 - there seems to be a general regression around april 2012
  • Some data is spotty due to crashes on different phones.

Eideticker

  • Cross browser page load test added to dashboard (bug 852744)
    • Our perf is quite bad, filed a bug (bug 857359) but no action. Anything that can be done?
  • Found source of regression in LG G2X CNN test (bug 855019)

Round Table

Action Items