Open Bug 1141544 Opened 9 years ago Updated 2 years ago

Intermittent test_oob_cert_auth.js | xpcshell return code: 0 | :0: error: 127.0.0.1:49413 uses an invalid security certificate.

Categories

(Core :: Security: PSM, defect, P2)

Product:
Core
Component:
Security: PSM
Platform:
x86
Gonk (Firefox OS)
Type:
defect

Tracking

()

People

(Reporter: cbook, Unassigned)

References

(
URL
)

Details

(Keywords: intermittent-failure, Whiteboard: [psm-intermittent][psm-backlog]) 

b2g_emulator_vm mozilla-inbound debug test xpcshell-2

https://treeherder.mozilla.org/logviewer.html#?job_id=7401399&repo=mozilla-inbound

01:40:07 WARNING - TEST-UNEXPECTED-FAIL | toolkit/devtools/security/tests/unit/test_oob_cert_auth.js | xpcshell return code: 0 

01:40:07 INFO - PROCESS | toolkit/devtools/security/tests/unit/test_oob_cert_auth.js | :0: error: 127.0.0.1:49413 uses an invalid security certificate.
Any ideas on this, David?
Component: Security → Security: PSM
Flags: needinfo?(dkeeler)
I think the certificate aspect of this may be a red herring. In any case, I don't know what's going on and can't reproduce locally. Maybe someone from devtools and/or the original author has some insight?
Flags: needinfo?(dkeeler) → needinfo?(jryans)
Will investigate, keeping ni? until I know more.
Flags: needinfo?(jryans)
See Also: → 1166032, 1163052
I'm making some progress investigating this.  I'll assign for now, hopefully have more to say soon.
Assignee: nobody → jryans
Status: NEW → ASSIGNED
As far as I know, nothing has changed in the related DevTools that would cause this to occur.

It looks like a latent race condition in either networking code or the TLSServerSocket code (which I added a while ago).  See also bug 1166032 and bug 1163052 for similar bugs intermittent failures on this test, but those have crash dumps in TLSServerSecurityObserverProxy::OnHandshakeDoneRunnable or DispatchCertVerificationResult.

I can continue investigating, but I don't expect to find a quick answer at the moment.  Also, I'll be on PTO next week.

If we need a short term fix to stop hassling the sheriffs, let's consider disabling this test on b2g emulators for now, which is where nearly all the intermittents appear to be from.
I don't have time to investigate this further at the moment, so for now, I've disabled this test on B2G emulator debug and Android opt, which is where it appears to fail according to Orange Factor.
Assignee: jryans → nobody
Status: ASSIGNED → NEW
Mass whiteboard change to annotate PSM intermittent test failures as [psm-intermittent]. Filter on 31b932bd-1aad-4e29-9f4b-4cd864a3ffdc if that's important to you.
Whiteboard: [psm-intermittent]
Bulk assigning P3 to all open intermittent bugs without a priority set in Firefox components per bug 1298978.
Priority: -- → P3
Stumbled upon this recently while reviewing patches to clean up B2G test annotations. Wonder if we need to disable this on Android opt at this point.
Flags: needinfo?(ryanvm)
Still hits on buildbot-based Android jobs, but not on Taskcluster ones. Guess I'll add a note that this should be re-enabled once we fully decommission buildbot-based Android tests.
Flags: needinfo?(ryanvm)
Pushed by ryanvm@gmail.com:
https://hg.mozilla.org/integration/mozilla-inbound/rev/70b63da571e6
Clean up the devtools security xpcshell manifest.
The leave-open keyword is there and there is no activity for 6 months.
:keeler, maybe it's time to close this bug?
Flags: needinfo?(dkeeler)
The comment in the patch says we can re-enable the test when we don't run android tests on buildbot. Judging by https://wiki.mozilla.org/Buildbot being blank and the last change comment being "deleting obsolete page" (oh and I guess this was b2g-specific), I imagine we can go ahead and do this.
Flags: needinfo?(dkeeler)
Keywords: leave-open
Priority: P3 → P2
Whiteboard: [psm-intermittent] → [psm-intermittent][psm-backlog]
Severity: normal → S3
You need to log in before you can comment on or make changes to this bug.