StartInstrumentation is the test execution. One or more tests are flakily taking a long time.

I don't think we have any results to upload to the flakiness dashboard in the event of a swarming timeout.

Do you mean, in this case, we don't upload anything to the flakiness dashboard?

Can you identify which tests are flakily taking a log time from the logs? It's not really clear to me.

#3: correct.

And no, not in this case. We lose too much of the log w/ the timeout kill, and we don't get the logcat at all.

er, to clarify: we do upload a JSON to the dashboard, but it's empty: http://test-results.appspot.com/testfile?builder=linux_android_rel_ng&name=full_results.json&master=tryserver.chromium.android&testtype=content_browsertests%20%28with%20patch%29&buildnumber=153539

I think we're *supposed* to be getting a SIGTERM from swarming when we hit the timeout, but I don't see evidence of that in the log (which I would expect to see, even if it's just a log message). I'm not sure if that's because our SIGTERM handling is broken or because swarming's not sending it correctly.

Swarming sends the SIGTERM to the process it started. If you are a few process down the tree, it's possible the propagation is not done properly (?)

Yeah, that's exactly what's going on - SIGTERM is getting delivered to the wrapper script, but that's not propagating it down to the test runner.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/626bdde804d615c4582b5c1a78bd8929f83b1ab9

commit 626bdde804d615c4582b5c1a78bd8929f83b1ab9
Author: jbudorick <jbudorick@chromium.org>
Date: Wed Oct 05 18:12:05 2016

[Android] Use os.exec in the test wrapper scripts.

BUG= 652787 

Review-Url: https://codereview.chromium.org/2397523003
Cr-Commit-Position: refs/heads/master@{#423220}

[modify] https://crrev.com/626bdde804d615c4582b5c1a78bd8929f83b1ab9/build/android/gyp/create_test_runner_script.py

What effect do we expect this change to have with respect to the original issue? With the test runner receiving the SIGTERM, will we be able to see more info in the logs? Will the results be uploaded to the flakiness dashboard? 

#11: we should see:
 1) 'Received SIGTERM. Stopping test execution.' in the log (from https://codesearch.chromium.org/chromium/src/build/android/pylib/local/device/local_device_test_run.py?rcl=0&l=85)
 2) the test runner complete with an exit code other than -15. I expect it to be nonzero, but I'm not sure what it'll be.
 3) Some results reported in the log and uploaded to the flakiness dashboard in the subsequent step.

 Issue 649831  should have fixed execution timeout errors being marked as infra failures.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/626bdde804d615c4582b5c1a78bd8929f83b1ab9

commit 626bdde804d615c4582b5c1a78bd8929f83b1ab9
Author: jbudorick <jbudorick@chromium.org>
Date: Wed Oct 05 18:12:05 2016

[Android] Use os.exec in the test wrapper scripts.

BUG= 652787 

Review-Url: https://codereview.chromium.org/2397523003
Cr-Commit-Position: refs/heads/master@{#423220}

[modify] https://crrev.com/626bdde804d615c4582b5c1a78bd8929f83b1ab9/build/android/gyp/create_test_runner_script.py

[Automated comment] removing mislabelled merge-merged-2840