Mobile/Testing/09 07 16: Difference between revisions

From MozillaWiki
Jump to navigation Jump to search
m (→‎Perfherder: - alerts update)
(→‎Autophone: webrtc)
Line 13: Line 13:
* [gbrown] I was talking with KevinChen on the weekend. He is trying to resolve the jemalloc_crash failures, but is having trouble reproducing the crashes: On a taskcluster one-click loaner, the loan expires before he hits a crash. He borrowed a slave from releng, but cannot reproduce the crash there, even though he is running full jobs from mozharness. He has been trying to repro with a debug build; I suggested he try an opt build too. I also suggested he might try a different aws type (tst-emulator64 vs tst-linux64). If you have more suggestions, please contact him!
* [gbrown] I was talking with KevinChen on the weekend. He is trying to resolve the jemalloc_crash failures, but is having trouble reproducing the crashes: On a taskcluster one-click loaner, the loan expires before he hits a crash. He borrowed a slave from releng, but cannot reproduce the crash there, even though he is running full jobs from mozharness. He has been trying to repro with a debug build; I suggested he try an opt build too. I also suggested he might try a different aws type (tst-emulator64 vs tst-linux64). If you have more suggestions, please contact him!
== Autophone ==
== Autophone ==
* [dminor] running WebRTC tests on inbound, fxteam, autoland, etc.
** Maybe start with the crashtest job on inbound and verify that we're filtering jobs properly. It runs fast, so it should have a small impact if things go wrong.


== Perfherder ==
== Perfherder ==

Revision as of 17:11, 7 September 2016

Previous Action Items

Status reports

Dev team

Rel Eng

A Team

  • gbrown is on PTO - back next week.
  • [gbrown] I was talking with KevinChen on the weekend. He is trying to resolve the jemalloc_crash failures, but is having trouble reproducing the crashes: On a taskcluster one-click loaner, the loan expires before he hits a crash. He borrowed a slave from releng, but cannot reproduce the crash there, even though he is running full jobs from mozharness. He has been trying to repro with a debug build; I suggested he try an opt build too. I also suggested he might try a different aws type (tst-emulator64 vs tst-linux64). If you have more suggestions, please contact him!

Autophone

  • [dminor] running WebRTC tests on inbound, fxteam, autoland, etc.
    • Maybe start with the crashtest job on inbound and verify that we're filtering jobs properly. It runs fast, so it should have a small impact if things go wrong.

Perfherder

  • bug 1300490 - perf regression on mytimes local/remote pages - assigned
  • no installer size alerts!

Round Table

Action Items