Intermittent test_ev_certificate.py TestEVCertificate.test_ev_certificate | AssertionError: u'verifiedDomain' != 'verifiedIdentity'

RESOLVED WONTFIX

Status

defect
RESOLVED WONTFIX
3 years ago
7 days ago

People

(Reporter: intermittent-bug-filer, Unassigned)

Tracking

({intermittent-failure})

Version 3
Points:
---
Dependency tree / graph

Firefox Tracking Flags

(Not tracked)

Details

Fixed by the proxy changes on bug 1289697.
Status: NEW → RESOLVED
Closed: 3 years ago
Resolution: --- → FIXED
Issues like those can still happen. So reopening.
Status: RESOLVED → REOPENED
Resolution: FIXED → ---
The problem here is that sometimes the certificate cannot be verified via the OCSP URL. I noticed that also on bug 1322199 and maybe the proxy changes there will get this fixed.
Status: REOPENED → NEW
Depends on: 1322199
https://wiki.mozilla.org/Bugmasters#Intermittent_Test_Failure_Cleanup
Status: NEW → RESOLVED
Closed: 3 years ago2 years ago
Resolution: --- → INCOMPLETE
Status: RESOLVED → REOPENED
Resolution: INCOMPLETE → ---
Why has this bug been reopened? There are no such failures.
Status: REOPENED → RESOLVED
Closed: 2 years ago2 years ago
Flags: needinfo?(csabou)
Resolution: --- → INCOMPLETE
Sorry, it was wrongly mistaken by me with this bug: https://bugzilla.mozilla.org/show_bug.cgi?id=1407663 from this fail: https://treeherder.mozilla.org/#/jobs?repo=mozilla-central&revision=3ea5af83c0c681bbe4a81ba3fbb6ae71059dd465&selectedJob=136227292 
I know how I judged that. Sorry.
Flags: needinfo?(csabou)
Recent log failure: https://treeherder.mozilla.org/logviewer.html#?job_id=187528044&repo=mozilla-inbound&lineNumber=8324
Status: RESOLVED → REOPENED
Resolution: INCOMPLETE → ---
What I really wonder is why this failure only occurs on trunk but not beta nor 60esr! Are there special network settings set for builds from trunk, which we don't enable on beta or releases yet? April, are you aware of something?
Flags: needinfo?(april)
I know that there are some NSS log debugging options you can enable, but I don't think they're enabled when building from trunk.

The best person to ask is maybe... mgoodwin or keeler? I've NI'd them to see if they have any ideas.
Flags: needinfo?(mgoodwin)
Flags: needinfo?(dkeeler)
Flags: needinfo?(april)
One thing you could do is enable the certverifier log (set the environment variable MOZ_LOG to "certverifier:5") and look for any OCSP failures (see the MOZ_LOG statements in https://searchfox.org/mozilla-central/rev/71ef4447db179639be9eff4471f32a95423962d7/security/certverifier/NSSCertDBTrustDomain.cpp#355 ).
Flags: needinfo?(dkeeler)
Flags: needinfo?(mgoodwin)
In the last month it was only happening once. So I'm going to defer any work on this bug until the intermittent failure rate has been substantially increased.
(In reply to Andreea Pavel [:apavel] from comment #32)
> This seems to have started with bug 1493000, Aaron can you please take a
> look?

No, this is just a single failure, and as you can see this bug has been created already a while ago. Maybe there is some flakiness with the network lately. gtest is also something totally different.
Flags: needinfo?(aklotz)
(In reply to Henrik Skupin (:whimboo) from comment #33)
> (In reply to Andreea Pavel [:apavel] from comment #32)
> > This seems to have started with bug 1493000, Aaron can you please take a
> > look?
> 
> No, this is just a single failure, and as you can see this bug has been
> created already a while ago. Maybe there is some flakiness with the network
> lately. gtest is also something totally different.

My assessment was based on the retriggers. I know it's an old bug, i was referring to the most recent occurrences. Thanks for looking into this Henrik. Should we close this as worksforme since you mention maybe there's something with the network and it's just a single failure?
Flags: needinfo?(hskupin)
It still fails so we need an open bug. If the number of failures increases we will have to further investigate.
Flags: needinfo?(hskupin)
okay Henrik, thank you!
There have been 40 failures in the last 7 days.

Failures per platform:

osx-10-10 / debug: 1
windows10-32-msvc / debug: 1
windows10-64 / all build types: 23
windows7-32 / all builds but !asan: 15

Summary: Intermittent test_ev_certificate.py TestEVCertificate.test_ev_certificate | AssertionError: u'verifiedDomain' != 'verifiedIdentity'

Recent relevant log file: https://treeherder.mozilla.org/logviewer.html#?job_id=205220000&repo=autoland&lineNumber=10622
Flags: needinfo?(hskupin)
Whiteboard: [stockwell needswork]
Last week we have seen an increase of those failures:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?startday=2018-09-15&endday=2018-10-15&tree=all&bug=1289344

When this continues with such a high rate we might want to find a way to setup the certverifier log as stated in comment 27. I will keep the needinfo request to have a look again later this week.
The problem here is an empty browser window as already reported on bug 1312632. This is something I will actually have to investigate ASAP.
Depends on: 1312632
Flags: needinfo?(hskupin)
The failures stopped on Oct 18th. I will revisit this bug in a couple of days.
Flags: needinfo?(hskupin)
(In reply to Narcis Beleuzu [:NarcisB] from comment #47)
> This started to almost perma fail starting with 
> https://treeherder.mozilla.org/#/
> jobs?repo=autoland&resultStatus=testfailed%2Cbusted%2Cexception%2Crunnable&re
> vision=3c7165875616d4b9250815f1a2d324469d5d2a36&searchStr=fxfn-r-e10s

This is based on which assumption? I absolutely can't see why this should have been started at least almost permafailing with this commit. I did some more retriggers of older commits and the failure rate is close to be equal.
Flags: needinfo?(alwu)
This failure spiked on October 26th and 27th only. After that I can only see 2 or so failures per day. There is nothing I can do to further investigate the issue.
Flags: needinfo?(hskupin)
(In reply to Henrik Skupin (:whimboo) from comment #52)
> This failure spiked on October 26th and 27th only. After that I can only see
> 2 or so failures per day. There is nothing I can do to further investigate
> the issue.

Should we close this as WONTFIX or WORKSFORME and reopen when there are new occurrences?
Status: REOPENED → RESOLVED
Closed: Last year10 months ago
Flags: needinfo?(hskupin)
Resolution: --- → FIXED
This is not fixed, and the problem might come up again. I would leave the bug open for now. In case nothing comes up it will be closed as incomplete.
Status: RESOLVED → REOPENED
Flags: needinfo?(hskupin)
Resolution: FIXED → ---
(In reply to Henrik Skupin (:whimboo) from comment #54)
> This is not fixed, and the problem might come up again. I would leave the
> bug open for now. In case nothing comes up it will be closed as incomplete.

Sorry, i changed the resolution by mistake, else i wouldn't have asked about how to close it :D thanks for clarifying though, Henrik.
There are 3 failures in the last 7 days, removing disable recommended.
Whiteboard: [stockwell disable-recommended]
Status: REOPENED → RESOLVED
Closed: 10 months ago5 months ago
Resolution: --- → INCOMPLETE
Status: RESOLVED → REOPENED
Resolution: INCOMPLETE → ---

The test will be removed via bug 1573393 because it is no longer valid due to EV isn't handled differently than DV.

Status: REOPENED → RESOLVED
Closed: 5 months ago7 days ago
Resolution: --- → WONTFIX
You need to log in before you can comment on or make changes to this bug.