1904967 - Permafailing multiple condprof mochitest-plain jobs on windows 11

Status: NEW

(Testing :: Condprofile, defect, P2)

Description

[Iulian Moraru]

• Push where this was first noticed
  All failure logs for the failing tests:
• Failure log for failures on test_file_upload.html
• Failure log for failures on test_bug370098.html
• Failure log for failures on test_bug338583.html
• Failure log for failures on test_bug1728171.html

Comment 1

[Joel Maher ( :jmaher ) (UTC -8)]

this is odd because on try this passed: https://treeherder.mozilla.org/jobs?repo=try&tier=1%2C2%2C3&searchStr=condprof&revision=8bfda803712dfa42aef10dae4eccf8bca3976f10&selectedTaskRun=NxYd1nJ5TK69Gtj1z7N8BA.0

so this could be 1 of 2 things:

1. windows updated image that took ~24 hours to cycle out
2. the conditioned profile (that comes from mozilla-central) was regenerated and is now causing failures

I am thinking this is #2.

Comment 2

[Joel Maher ( :jmaher ) (UTC -8)]

I am almost certain it is related to the conditioned profile (i.e. ^ #2)

Comment 3

[Joel Maher ( :jmaher ) (UTC -8)]

the message I see in the logs is:

[task 2024-06-26T19:38:11.114Z] 19:38:11     INFO - GECKO(2740) | JavaScript error: , line 0: TypeError: NetworkError when attempting to fetch resource.
[task 2024-06-26T19:38:13.634Z] 19:38:13     INFO -  console.trace: Places - FrecencyRecalculator: "Recalculate 50 frecency values"
[task 2024-06-26T19:38:13.634Z] 19:38:13     INFO -  resource://gre/modules/PlacesFrecencyRecalculator.sys.mjs 240 recalculateSomeFrecencies
[task 2024-06-26T19:38:13.635Z] 19:38:13     INFO -  resource://gre/modules/PlacesFrecencyRecalculator.sys.mjs 176 #taskFn
[task 2024-06-26T19:38:13.635Z] 19:38:13     INFO -  resource://gre/modules/DeferredTask.sys.mjs 341 _runTask
[task 2024-06-26T19:38:13.635Z] 19:38:13     INFO -  resource://gre/modules/DeferredTask.sys.mjs 312 _timerCallback/<
[task 2024-06-26T19:38:13.635Z] 19:38:13     INFO -  resource://gre/modules/DeferredTask.sys.mjs 331 _timerCallback
[task 2024-06-26T19:38:13.636Z] 19:38:13     INFO -  resource://gre/modules/DeferredTask.sys.mjs 179 _startTimer/callback/<
[task 2024-06-26T19:38:13.636Z] 19:38:13     INFO -  D:\task_171942871688720\build\tests\mochitest\server.js 178 runServer
[task 2024-06-26T19:38:13.636Z] 19:38:13     INFO -  D:\task_171942871688720\build\tests\mochitest\server.js 71
[task 2024-06-26T19:38:13.637Z] 19:38:13     INFO -  console.error: (new TypeError("can't access property \"executeTransaction\", db is undefined", "resource://gre/modules/PlacesFrecencyRecalculator.sys.mjs", 246))
[task 2024-06-26T19:38:13.637Z] 19:38:13     INFO -  console.error: Places - FrecencyRecalculator: (new TypeError("can't access property \"executeTransaction\", db is undefined", "resource://gre/modules/PlacesFrecencyRecalculator.sys.mjs", 246))
[task 2024-06-26T19:41:46.563Z] 19:41:46     INFO - TEST-INFO | started process screenshot

I wonder if this is the case because we have some pre-existing profile

Comment 4

[Joel Maher ( :jmaher ) (UTC -8)]

we have a new profile generated as of ~2.5 hours ago and this seems to work:
https://treeherder.mozilla.org/jobs?repo=autoland&resultStatus=success%2Cpending%2Crunning%2Ctestfailed%2Cbusted%2Cexception%2Crunnable&searchStr=condprof%2Cwin%2Cmochitest%2Cship&revision=eb8e54ddbc044ecab12d09d7a7417a9d7a9bec8e&selectedTaskRun=NwmqYeFsRf2Ah_AgLcRARA.0

so this is an issue with a profile- no idea why that specific error would be called- maybe the profile was corrupt.

I see :mak touched the code last to define the db which is undefined here- maybe he would know what the edge case is?

Comment 5

[Marco Bonardo [:mak]]

Having undefined there means gAsyncDBWrapperPromised is undefined, that's not impossible if we failed getting a connection to the database, though I would expect to see a console.error if we failed that.

In a different log I see console.error: (new TypeError("connection not specified or invalid.", "resource://gre/modules/Sqlite.sys.mjs", 1566)) that means PlacesUtils.history.DBConnection is also invalid, that pretty much means opening the main connection to places.sqlite failed.

If you have a log from a DEBUG build that would probably tell us more, with the available info it could be many things, maybe the db file copy was wrong, or there was a filesystem access problem, or the test tried to initialize Places very late during shutdown (though the code seems to handle this case, so I'd rather think it's one of the other two).

I think the tests are not failing due to this anyway, so it's just console error "noise"?

Comment 6

[Intermittent Failures Robot]

14 failures in 4116 pushes (0.003 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• autoland: 10
• mozilla-central: 4

Platform and build breakdown:

• windows11-64-2009-qr: 5
  • opt: 5
• windows11-64-2009-shippable-qr: 9
  • opt: 9

For more details, see:
https://treeherder.mozilla.org/intermittent-failures/bugdetails?bug=1904967&startday=2024-06-24&endday=2024-06-30&tree=all