CrashKill/2013-09-30

From MozillaWiki
Jump to: navigation, search

General

Socorro

  • RabbitMQ is very fast, but we have had a few teething problems
    • On Thursday we discovered that the config was fat fingered on Tuesday during the release, meaning many crashes were processed without symbols (all plugin, Fennec, and B2G crashes). bug 921042
    • We reprocessed 2 days of crashes in just over an hour.
    • On Sunday processors blipped and when oncall checked in they restarted the monitor, leading to alerts - no cause for concern, but a couple of actions to clean up. bug 921831
    • Another issue is about:crashes links don't insert priority crashes in the right place. There's a PR for this which should land and ship today. See bug 921443
  • As a result of those issues, Supersearch (61) did not ship last week. We plan to ship today
    • Stage DB refresh is happening now, because it failed yesterday, due to a config change by DBOps last Thursday. bug 922156
    • Once that's done we can ship.
  • No ADIs yet for today: Metrics is working on it.

Flash

  • Now that the major crash fix in 11.8 has been delivered to users, this version is lower in both crashes and hangs compared to 11.7 - still ~50% higher than 11.2 in crashes and ~100% higher in hangs but this is the best Flash release since they released sandboxing and hardware acceleration with 11.3 in June of 2012.
  • 11.9 beta regressions have been fixed, current beta builds should be more or less final, release coming in October (patch Tuesday).

Desktop

  • bug 772330 (Radeon): Discussing with AMD. They think there is potential in the theory that it's a CPU bug and will investigate with their internal tools.

Beta

Aurora

Trunk

Mobile

B2G

1.3

  • still seeing some instances of bug 914823: mozilla::gl::SharedSurface_Gralloc::~SharedSurface_Gralloc
  • need to wait to see if we end up getting some more crashes from : mozilla::docshell::POfflineCacheUpdateChild::Unregister(int)
    • it may be a dup of bug 914823
  • Note : bug 915223 ended up WFM; not sure what caused the fix for activesync connections

1.2

  • bug 922196 - crash in nsBaseAppShell::DoProcessNextNativeEvent(bool, unsigned int)
    • could be a crash on purpose.
  • Note : bug 915869 should have landed in aurora.
  • Note : several crashes were fixed and we are waiting for them to drop off.
    • ex bug 917594 - crash in js::ShapeTable::search(int, bool)

Release

  • Camera crashes:
    • bug 878542 - crash in liboemcamera.so@0x12240 on Keon / [@ liboemcamera.so@0x12494 ] on inari
    • bug 901226 - crash in liboemcamera.so@0x116bc
  • bug 878366 - crash on abort in mozilla::dom::ContentChild::ProcessingError
    • Partial fix is in : bug 867025 - [unagi][tara][weekly build 13.04.17]monkey test crash in mozilla::dom::ContentChild::ProcessingError
  • BT Crashes: (Note: this fell down in the numbers, people probably gave up)
    • bug 891160 - [inari][bluetooth] Bluetooth crash in mozilla::dom::bluetooth::PBluetoothRequest::Transition with abort message: "__delete__()d actor" after trying to send to the Mac and then switching to camera app

Stability Week Actions

Notes

  • No meeting on October 7 due to Summit travels

Upcoming PTO

  • KaiRo probably off most of the two weeks after the Summit.
  • Tracy PTO Sept. 30th.