Open Bug 1534590 Opened 5 years ago Updated 9 days ago

Intermittent ShutdownLeaks | process() called before end of test suite [only use for classification if there is no test failure directly mentioned after it]

Categories

(Core :: General, defect, P5)

defect

Tracking

()

People

(Reporter: intermittent-bug-filer, Unassigned)

References

Details

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

#[markdown(off)]
Filed by: rmaries [at] mozilla.com

https://treeherder.mozilla.org/logviewer.html#?job_id=233336005&repo=mozilla-inbound

https://queue.taskcluster.net/v1/task/HbbXfR5yTsCPjdPCKF9YGA/runs/0/artifacts/public/logs/live_backing.log

10:50:22 INFO - TEST-START | browser/components/preferences/in-content/tests/browser_security-2.js

10:50:24 INFO - GECKO(2123) | Exiting due to channel error.
10:50:24 INFO - TEST-INFO | Main app process: exit 1
10:50:24 ERROR - TEST-UNEXPECTED-FAIL | ShutdownLeaks | process() called before end of test suite
10:50:24 INFO - Buffered messages logged at 10:50:22
10:50:24 INFO - Entering test bound setup
10:50:24 INFO - Buffered messages finished
10:50:24 ERROR - TEST-UNEXPECTED-FAIL | browser/components/preferences/in-content/tests/browser_security-2.js | application terminated with exit code 1
10:50:24 INFO - runtests.py | Application ran for: 0:11:22.955978
10:50:24 INFO - zombiecheck | Reading PID log: /var/folders/kx/vjxn9xvj62q_mj4t8xyh3nm400000w/T/tmpAqDdCqpidlog
10:50:24 INFO - ==> process 2123 launched child process 2124
10:50:24 INFO - ==> process 2123 launched child process 2125
10:50:24 INFO - ==> process 2123 launched child process 2126
10:50:24 INFO - ==> process 2123 launched child process 2127
10:50:24 INFO - ==> process 2123 launched child process 2128
10:50:24 INFO - ==> process 2123 launched child process 2129
10:50:24 INFO - ==> process 2123 launched child process 2130
10:50:24 INFO - ==> process 2123 launched child process 2131
10:50:24 INFO - ==> process 2123 launched child process 2132
10:50:24 INFO - ==> process 2123 launched child process 2133
10:50:24 INFO - zombiecheck | Checking for orphan process with PID: 2124
10:50:24 INFO - zombiecheck | Checking for orphan process with PID: 2125
10:50:24 INFO - zombiecheck | Checking for orphan process with PID: 2126
10:50:24 INFO - zombiecheck | Checking for orphan process with PID: 2127
10:50:24 INFO - zombiecheck | Checking for orphan process with PID: 2128
10:50:24 INFO - zombiecheck | Checking for orphan process with PID: 2129
10:50:24 INFO - zombiecheck | Checking for orphan process with PID: 2130
10:50:24 INFO - zombiecheck | Checking for orphan process with PID: 2131
10:50:24 INFO - zombiecheck | Checking for orphan process with PID: 2132
10:50:24 INFO - zombiecheck | Checking for orphan process with PID: 2133

Some of the failures reported here have an assertion https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=233884608&repo=mozilla-inbound&lineNumber=16151 and seems to fail the most netwerk/cookie/test/browser/browser_sharedWorker.js

Should that get a separate bug? What with the remaining failures which represent a sudden increase in shutdown leak reports? Any idea what started those?

Flags: needinfo?(gbrown)

(In reply to Sebastian Hengst [:aryx] (needinfo on intermittent or backout) from comment #1)

Many of the netwerk/cookie/test/browser/browser_sharedWorker.js failures are crashes. I think the leak reports are a consequence of the crash. I would like to see a bug filed for browser_sharedWorker.js crashes. I see SharedWorkerService in those crash reports, and that was recently modified in bug 1511749 -- worth an ni to :baku.

I don't know about the remaining failures. :mccr8 might have ideas?

Flags: needinfo?(gbrown)

(In reply to Geoff Brown [:gbrown] from comment #3)

I see SharedWorkerService in those crash reports, and that was recently modified in bug 1511749 -- worth an ni to :baku.

Thanks, that fixed it.

Over the last 7 days there are 54 failures on this bug. These happen on linux32, linux64, linux64-ccov, osx-10-10, windows10-64-ccov, windows10-64.

Here is the most recent failure log: https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=237250665&repo=mozilla-inbound&lineNumber=14001

Flags: needinfo?(jaws)

The most recent failure logs are coming from /dom/webauthn. Nothing has changed recently in browser_security-2.js that would have triggered this in Preferences.

Component: Preferences → DOM: Web Authentication
Flags: needinfo?(jaws)
Product: Firefox → Core

(Redirecting my needinfo to JC Jones, triage lead for the new component)

Flags: needinfo?(jjones)

See https://bugzilla.mozilla.org/show_bug.cgi?id=1540378#c6 for at least my part in this. I don't know about the earlier issues, but the current rash are explained a bit there in Bug 1540378.

Flags: needinfo?(jjones)

So this should be resolved now, or at least back to whatever it was before my WebAuthn change lit it up.

Failure rate has decreased significantly since the 3rd of April.

Whiteboard: [stockwell disable-recommended]
Whiteboard: [stockwell disable-recommended]
Whiteboard: [stockwell disable-recommended]
Flags: needinfo?(jjones)
Whiteboard: [stockwell needswork:owner]

I checked the first 15 of the failures, none of their logs have the WebAuthn tests in them. The webauthn team is just me, and I'm afraid I don't know even where to move this bug from WebAuthn. :(

Flags: needinfo?(jjones) → needinfo?(rmaries)

I checked again the most recent failures and some of them are misclassified.

Flags: needinfo?(rmaries)

I have no idea where this goes, so going to put it in Core General for now.

Component: DOM: Web Authentication → General
Summary: Intermittent ShutdownLeaks | process() called before end of test suite → Intermittent ShutdownLeaks | process() called before end of test suite [only use for classification if there is no test failure directly mentioned after it]
Severity: normal → S3
You need to log in before you can comment on or make changes to this bug.