1755966 - telemetry client tests that crash firefox on Windows don't have crash dump output

Status: RESOLVED FIXED

(Toolkit :: Telemetry, defect, P3)

Description

[Will Kahn-Greene [:willkg] ET needinfo? me]

When I was working on bug #1594515, I noticed that I could get crash processing output from the stackwalker on Linux and macOS, but not on Windows.

https://bugzilla.mozilla.org/show_bug.cgi?id=1594515#c25

The output_timeout value we're talking about is set here:

https://searchfox.org/mozilla-central/rev/9bed2623831804ac086bbb80cb666e5c3b00416a/testing/mozharness/scripts/telemetry/telemetry_client.py#233

It's set to 300 (seconds), but raising it to 1000 (seconds) gives Windows machinery enough time to recognize that it crashed and run the minidump through minidump-stackwalk and get crash processing output.

Between bug #1594515 and bug #1746747, we've fixed symbolication of crash stacks, but we still have to increase the output_timeout so we get stacks on Windows.

Comment 1

[Benjamin Forehand Jr[:b4hand]]

Is there a way to test this change on try or locally? I am assuming no.

Comment 2

[Will Kahn-Greene [:willkg] ET needinfo? me]

I tested it out by forcing a crash in the telemetry test client code:

diff --git a/toolkit/components/telemetry/tests/marionette/harness/telemetry_harness/testcase.py b/toolkit/components/telemetry/tests/
marionette/harness/telemetry_harness/testcase.py
--- a/toolkit/components/telemetry/tests/marionette/harness/telemetry_harness/testcase.py
+++ b/toolkit/components/telemetry/tests/marionette/harness/telemetry_harness/testcase.py
@@ -72,6 +72,13 @@ class TelemetryTestCase(WindowManagerMix
             self.marionette.close()
             self.marionette.switch_to_window(start_tab)
 
+    def crash_browser(self):
+        """Crash the browser."""
+        # FIXME(willkg): remove this
+
+        with self.marionette.using_context(self.marionette.CONTEXT_CONTENT):
+            self.marionette.navigate("about:crashparent")
+
     def search(self, text):
         """Perform a search via the browser's URL bar."""
 
diff --git a/toolkit/components/telemetry/tests/marionette/tests/client/test_main_tab_scalars.py b/toolkit/components/telemetry/tests/
marionette/tests/client/test_main_tab_scalars.py
--- a/toolkit/components/telemetry/tests/marionette/tests/client/test_main_tab_scalars.py
+++ b/toolkit/components/telemetry/tests/marionette/tests/client/test_main_tab_scalars.py
@@ -12,6 +12,9 @@ class TestMainTabScalars(TelemetryTestCa
     def test_main_tab_scalars(self):
         """Test for Telemetry Scalars."""
 
+        # Crash the browser
+        self.crash_browser()
+
         with self.marionette.using_context(self.marionette.CONTEXT_CHROME):
             start_tab = self.marionette.current_window_handle

Comment 3

[Benjamin Forehand Jr[:b4hand]]

Attachment: Bug 1755966 - Increase timeout to gather logs during a crash.; r?willkg,chutten

Comment 4

[BugBot [:suhaib / :marco/ :calixte]]

The severity field is not set for this bug.
:chutten, could you have a look please?

For more information, please visit auto_nag documentation.

Comment 5

[BugBot [:suhaib / :marco/ :calixte]]

There's a r+ patch which didn't land and no activity in this bug for 2 weeks.
:b4hand, could you have a look please?
If you still have some work to do, you can add an action "Plan Changes" in Phabricator.
For more information, please visit auto_nag documentation.

Comment 6

[Will Kahn-Greene [:willkg] ET needinfo? me]

I don't have level 3 commit access, so I asked chutten to kindly press the button to land this.

Comment 7

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/2033129d4715
Increase timeout to gather logs during a crash.; r=willkg,chutten

Comment 8

[Atila Butkovits]

https://hg.mozilla.org/mozilla-central/rev/2033129d4715