Open Bug 1760134 Opened 2 years ago Updated 1 month ago

Intermittent toolkit/components/backgroundhangmonitor/tests/test_BHRObserver.js | single tracking bug

Categories

(Core :: Performance Engineering, defect, P3)

defect

Tracking

()

People

(Reporter: jmaher, Unassigned, NeedInfo)

References

Details

(Keywords: intermittent-failure, intermittent-testcase, leave-open, Whiteboard: [stockwell disabled])

Attachments

(1 file, 1 obsolete file)

No description provided.
Severity: -- → S4

There have been 37 total failures in the last 7 days, recent failure log.
Affected platforms are:

  • windows7-32-qr
  • windows7-32-shippable-qr
[task 2022-10-09T03:33:13.181Z] 03:33:13     INFO -  TEST-START | toolkit/components/backgroundhangmonitor/tests/test_BHRObserver.js
[task 2022-10-09T03:38:13.222Z] 03:38:13  WARNING -  TEST-UNEXPECTED-TIMEOUT | toolkit/components/backgroundhangmonitor/tests/test_BHRObserver.js | Test timed out
[task 2022-10-09T03:38:13.222Z] 03:38:13     INFO -  TEST-INFO took 300040ms
[task 2022-10-09T03:38:13.222Z] 03:38:13     INFO -  >>>>>>>
[task 2022-10-09T03:38:13.222Z] 03:38:13     INFO -  (xpcshell/head.js) | test MAIN run_test pending (1)
[task 2022-10-09T03:38:13.223Z] 03:38:13     INFO -  (xpcshell/head.js) | test run_next_test 0 pending (2)
[task 2022-10-09T03:38:13.223Z] 03:38:13     INFO -  (xpcshell/head.js) | test MAIN run_test finished (2)
[task 2022-10-09T03:38:13.223Z] 03:38:13     INFO -  running event loop
[task 2022-10-09T03:38:13.224Z] 03:38:13     INFO -  toolkit/components/backgroundhangmonitor/tests/test_BHRObserver.js | Starting test_BHRObserver
[task 2022-10-09T03:38:13.224Z] 03:38:13     INFO -  (xpcshell/head.js) | test test_BHRObserver pending (2)
[task 2022-10-09T03:38:13.224Z] 03:38:13     INFO -  TEST-PASS | toolkit/components/backgroundhangmonitor/tests/test_BHRObserver.js | test_BHRObserver - [test_BHRObserver : 27] Should have BHRTelemetryService - [object Object] == true
Flags: needinfo?(dtownsend)
Whiteboard: [stockwell needswork:owner]

There have been 43 total failures in the last 7 days, recent failure log.
Affected platforms are:

  • windows7-32-qr
  • windows7-32-shippable-qr

Redirecting NI to Doug as I believe he's more familiar with this area.

Flags: needinfo?(dtownsend) → needinfo?(doug.turner)

Sorry, moving to correct Doug.

Flags: needinfo?(doug.turner) → needinfo?(dothayer)

Update:

There have been 33 failures within the last 7 days:

  • 15 failures on Windows 7 WebRender opt
  • 18 failures on Windows 7 WebRender Shippable opt

Recent failure log: https://treeherder.mozilla.org/logviewer?job_id=394878575&repo=autoland&lineNumber=4433

[task 2022-10-30T13:07:13.914Z] 13:07:13     INFO -  TEST-START | toolkit/components/backgroundhangmonitor/tests/test_BHRObserver.js
[task 2022-10-30T13:12:13.944Z] 13:12:13  WARNING -  TEST-UNEXPECTED-TIMEOUT | toolkit/components/backgroundhangmonitor/tests/test_BHRObserver.js | Test timed out
[task 2022-10-30T13:12:13.944Z] 13:12:13     INFO -  TEST-INFO took 300030ms
See Also: → 1781295
Duplicate of this bug: 1781295
Whiteboard: [stockwell disable-recommended]

There have been 44 total failures in the last 7 days, recent failure log.
Affected platforms are:

  • windows7-32-qr
  • windows7-32-shippable-qr
Whiteboard: [stockwell disable-recommended] → [stockwell needswork:owner]
Assignee: nobody → csabou
Status: NEW → ASSIGNED
Pushed by csabou@mozilla.com:
https://hg.mozilla.org/integration/autoland/rev/7824c1902b6b
disable test_BHRObserver.js on win32 because of frequent failures. r=intermittent-reviewers,jmaher DONTBUILD
Pushed by csabou@mozilla.com:
https://hg.mozilla.org/integration/autoland/rev/9f4fd5a62c72
Fix identation and syntax for disabling condition. a=test-only
Assignee: csabou → nobody
Status: ASSIGNED → NEW
Keywords: leave-open
Whiteboard: [stockwell disable-recommended] → [stockwell disabled]

Clearing needinfo so this can be auto closed

Flags: needinfo?(dothayer)

Sorry this was disabled needinfo is probably valid

Flags: needinfo?(dothayer)

Hey Nika, you worked on the original version test. Considering that we seem to not run it in most conditions, and the test is still periodically failing intermittently, do you think it's worth removing entirely?

Flags: needinfo?(nika)

It'd be unfortunate to lose coverage for the BHR telemetry probes completely, but I suppose we'd notice if BHR hangs from content stopped coming in, so perhaps it's not an issue, and the code is disabled outside of nightly anyway.

I think ideally we'd change the code to test as much of the hang reporter as possible without it being flakey. I think that would involve adding a test-only method to induce a fake hang in the BHR monitor without actually hanging. I think this would probably involve dispatching a runnable to the background hang manager thread which simulates the behaviour from BackgroundHangThread::Notify when a hang is detected (https://searchfox.org/mozilla-central/rev/6444ed17e9f5e0d4e7dabc13c930d46b529fca15/toolkit/components/backgroundhangmonitor/BackgroundHangMonitor.cpp#506-521). This would then be notified at the start of the next runnable to be processed fairly reliably (https://searchfox.org/mozilla-central/rev/6444ed17e9f5e0d4e7dabc13c930d46b529fca15/toolkit/components/backgroundhangmonitor/BackgroundHangMonitor.cpp#267).

Flags: needinfo?(nika)
Component: General → Performance
Product: Toolkit → Core
Component: Performance → Performance Engineering
Attachment #9383141 - Attachment is obsolete: true
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: