Closed Bug 1518746 Opened 3 years ago Closed 3 years ago

Intermittent netwerk/test/mochitests/test_accept_header.html | Test timed out.

Categories

(Core :: Networking, defect)

defect
Not set
normal

Tracking

()

RESOLVED FIXED
mozilla66
Tracking Status
firefox66 --- fixed

People

(Reporter: intermittent-bug-filer, Assigned: mossop)

References

Details

(Keywords: intermittent-failure, Whiteboard: [necko-triaged][stockwell unknown])

Attachments

(2 files)

Filed by: nerli [at] mozilla.com

https://treeherder.mozilla.org/logviewer.html#?job_id=220765340&repo=mozilla-inbound

https://queue.taskcluster.net/v1/task/cDUtI-u4SIS8zyuzvh067A/runs/0/artifacts/public/logs/live_backing.log

[task 2019-01-09T07:50:54.357Z] 07:50:54 INFO - 125 INFO TEST-START | netwerk/test/mochitests/test_accept_header.html
[task 2019-01-09T07:56:12.928Z] 07:56:12 INFO - 126 INFO TEST-UNEXPECTED-FAIL | netwerk/test/mochitests/test_accept_header.html | Test timed out.
[task 2019-01-09T07:56:12.928Z] 07:56:12 INFO - SimpleTest.ok@SimpleTest/SimpleTest.js:275:7
[task 2019-01-09T07:56:12.928Z] 07:56:12 INFO - reportError@SimpleTest/TestRunner.js:121:7
[task 2019-01-09T07:56:12.928Z] 07:56:12 INFO - TestRunner._checkForHangs@SimpleTest/TestRunner.js:142:7
[task 2019-01-09T07:56:12.928Z] 07:56:12 INFO - 127 INFO TEST-OK | netwerk/test/mochitests/test_accept_header.html | took 311547ms

Whiteboard: [necko-triaged]

Looking at pushes around the date of the first failure this one is the most likely to affect the test in any way:
https://hg.mozilla.org/mozilla-central/rev/43c43fa35d2a

I assume the test just isn't expecting to get the favicon resource, but I haven't looked at it too closely.

(In reply to Valentin Gosu [:valentin] from comment #2)

Looking at pushes around the date of the first failure this one is the most likely to affect the test in any way:
https://hg.mozilla.org/mozilla-central/rev/43c43fa35d2a

I assume the test just isn't expecting to get the favicon resource, but I haven't looked at it too closely.

Yeah, this looks like a trivial issue in the test, in that:

https://searchfox.org/mozilla-central/rev/c43240cef5829b8a2dec118faff8a5e1fec6ae1b/netwerk/test/mochitests/test_accept_header.html#26,30

disconnects the observer twice, which is already suspicious, but I suppose that maybe the favicon request trips the observer before we get the iframe request in there, and then the test just sits there doing nothing... though I don't see why the accept header would have made a difference for this test.

I suspect that bug 1516658 isn't at fault here. As Gijs says it should not have affected the timing of favicon requests, and checks of the accept header in this test are well guarded by uri checks (and additionally a bad accept header would have been logged as such it wouldn't cause a timeout). Also, the test doesn't include a favicon link anywhere, and attempts to load a site's root favicon are disabled in tests. I ran this locally with some modifications and saw no signs that favicons were loading.

The failure is pretty infrequent so catching it locally is pretty unlikely so lets fix some possible signs of problems and add some logging to see if we can figure out what is going on.

Pushed by dtownsend@mozilla.com:
https://hg.mozilla.org/integration/autoland/rev/837ae1da7e4d
Add additional logging to test_accept_header.html. r=valentin
Status: NEW → RESOLVED
Closed: 3 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla66
Status: RESOLVED → REOPENED
Resolution: FIXED → ---

So the logging shows that the test consistently fails when testing the iframe or style cases (slightly more often in the style case). Those are the two tests that use the PerformanceObserver API to catch the load. Let's try switching to an alternate method.

Assignee: nobody → dtownsend
Pushed by dtownsend@mozilla.com:
https://hg.mozilla.org/integration/autoland/rev/a9ff8ccd4e3d
Switch to load events for all tested resources. r=valentin
Status: REOPENED → RESOLVED
Closed: 3 years ago3 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.