oops, I forgot to use the right template.

There are actually several issues here, I believe. remote_run_result turns purple when anything in the build throws an exception. I'm working to resolve this bug in the recipe engine separately, but for now things will just be confusing.

See  bug 650322 , bug 634418, and bug 650353 for bugs tracking what appear to be 3 separate issues currently affecting the trybots.

 bug 650355  also seems to be affecting the tryserver.

is there an update on this? I'm still seeing most of the tryjobs on that CQ builder failing.

bug 634418 is still unsolved. I believe it is still causing most of the tryjobs to fail. Help is needed on the chromium side; something is generating a 300MB json file, which causes python on windows to OOM.

 bug 650355  is not as big of a culprit, I believe.

We seem to be recovering.

Still looks busted.

here's a recent failure: https://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_x64_rel_ng/builds/286926

it says
browser_tests browser_tests
TEST RESULTS WERE INVALID

I looked at the swarming output of all the browser_tests shards. all are fine except one:
https://chromium-swarm.appspot.com/user/task/3182d83d80c0ab10
It has 5000 hits for
[1296:1076:0926/220553:ERROR:crashpad_client_win.cc(490)] haven't called UseHandler()

(although it says "<truncated>" before the first one so there could very well be more).


scottmg: I see that crashpad was updated 4 days ago (from https://chromium.googlesource.com/chromium/src/+log/master/third_party/crashpad), to https://chromium.googlesource.com/chromium/src/+/ca7b105dc952bd2d9b2fd5aaf9563f25f15aa752. could that be the cause?

Going to merge this into 634418 as it looks like the same root cause.