1077348 - B2G contacts test is timing out waiting for content to be ready

Status: RESOLVED WONTFIX

(Testing Graveyard :: Eideticker, defect)

Description

[Dave Hunt [:davehunt] [he/him] ⌚BST]

It looks like the contacts app is timing out while waiting for the content to be ready:

Traceback (most recent call last):
  File "./bin/update-dashboard.py", line 279, in <module>
    main()
  File "./bin/update-dashboard.py", line 256, in main
    eideticker.prepare_test(testkey, options)
  File "/var/jenkins/workspace/b2g.flame-512.b2g-inbound.tests/src/eideticker/eideticker/runtest.py", line 58, in prepare_test
    test.prepare_app()
  File "/var/jenkins/workspace/b2g.flame-512.b2g-inbound.tests/src/eideticker/eideticker/b2gtestmixins.py", line 26, in prepare_app
    self.launch_app()
  File "/var/jenkins/workspace/b2g.flame-512.b2g-inbound.tests/src/eideticker/eideticker/b2gtestmixins.py", line 18, in launch_app
    self.wait_for_content_ready()
  File "/var/jenkins/workspace/b2g.flame-512.b2g-inbound.tests/src/eideticker/eideticker/b2gtestmixins.py", line 31, in wait_for_content_ready
    expected.element_present(*app._contact_locator))
  File "/var/jenkins/workspace/b2g.flame-512.b2g-inbound.tests/.env/local/lib/python2.7/site-packages/marionette/wait.py", line 143, in until
    cause=last_exc)
marionette.errors.TimeoutException: TimeoutException: Timed out after 244.6 seconds

There are no obvious changes in gaiatest that would address this. I'll try to replicate it locally.

Comment 1

[Dave Hunt [:davehunt] [he/him] ⌚BST]

I can replicate this, and the contacts are not appearing in the app. I wonder if perhaps the reference workload needs to be updated. John: do you have any ideas?

Comment 2

[Dave Hunt [:davehunt] [he/him] ⌚BST]

I've temporarily disabled this test so that we can regain coverage for other tests:
https://github.com/mozilla/eideticker-ci/commit/29f3055cc5e467793fced4f468848ab09daaeb78

Comment 3

[Jon Hylands [:jhylands]]

Typically when stuff like this happens, it means there is an issue with the database migration code. Reference workloads haven't changed in over a year. I would do a manual test (make reference-workload-xxx) with whatever build the automation system is using, and see what happens.

Comment 4

[Dave Hunt [:davehunt] [he/him] ⌚BST]

I can't replicate this with the latest b2g-inbound build. I've re-enabled the test in CI to see if it's fixed there too. If it is, I'll revert the change to disable it.

Comment 5

[Dave Hunt [:davehunt] [he/him] ⌚BST]

It turns out that this happens when we exclude the contact pictures from the reference workload. This was needed for bug 937730. Unfortunately I no longer have access to the failed Eideticker jobs to get a regression range. Will: Do we still need to exclude the contact pictures?

I see the following in logcat:
W/GeckoConsole(11754): [JavaScript Error: "IndexedDB UnknownErr: ActorsParent.cpp:5515"]
E/Communications(12126): Content JS ERROR: Error getting first contacts 
E/Communications(12126):     at c_getFirstContacts/onerror (app://communications.gaiamobile.org/contacts/gaia_build_defer_index.js:804:1349)

I suspect that if a picture associated with a contact cannot be found that we no longer silently ignore it.

Comment 6

[William Lachance (:wlach)]

Eugh, that's annoying. Unfortunately the contacts pictures made the results unreliable, since they were loaded asynchronously. Is there no way we could account for this case in gaia itself? "no picture for contact" seems like the sort of case where we should log a warning and ignore, not break entirely.

needinfo'ing contacts owners for more information

Comment 7

[Jose Manuel Cantera]

(In reply to William Lachance (:wlach) from comment #6)
> Eugh, that's annoying. Unfortunately the contacts pictures made the results
> unreliable, since they were loaded asynchronously. Is there no way we could
> account for this case in gaia itself? "no picture for contact" seems like
> the sort of case where we should log a warning and ignore, not break
> entirely.
> 
> needinfo'ing contacts owners for more information

Could you be more explicit on what information do you need?

thanks

Comment 8

[Francisco Jordano [:arcturus] [:francisco]]

Hei Dave and William,

I think we have a problem with recent indexeddb changes. When I try to add a workload and run the contacts app there is a problem in the migration process.

Same is happening in the dialer, the call log migration process gets stuck cause of the same error.

We are tracking that error here:
https://bugzilla.mozilla.org/show_bug.cgi?id=1082505

Comment 9

[Dave Hunt [:davehunt] [he/him] ⌚BST]

We're no longer running Eideticker tests against B2G.