1456022 - Intermittent /webdriver/tests/actions/sequence.py | test_release_no_actions_sends_no_events - setup error

Status: RESOLVED DUPLICATE of bug 1243108

(Core :: IPC, defect, P1)

Description

[Treeherder Bug Filer]

Filed by: ncsoregi [at] mozilla.com

https://treeherder.mozilla.org/logviewer.html#?job_id=175054505&repo=autoland

https://queue.taskcluster.net/v1/task/Es6T7gTCSnq97DQuTy-a6Q/runs/0/artifacts/public/logs/live_backing.log

task 2018-04-22T23:34:31.017Z] 23:34:31     INFO - STDOUT: :438: NoSuchWindowException
[task 2018-04-22T23:34:31.018Z] 23:34:31     INFO - STDOUT: =============================== warnings summary ===============================
[task 2018-04-22T23:34:31.018Z] 23:34:31     INFO - STDOUT: None
[task 2018-04-22T23:34:31.018Z] 23:34:31     INFO - STDOUT:   Module already imported so cannot be rewritten: mozlog
[task 2018-04-22T23:34:31.019Z] 23:34:31     INFO - STDOUT: -- Docs: http://doc.pytest.org/en/latest/warnings.html
[task 2018-04-22T23:34:31.019Z] 23:34:31     INFO - STDOUT: ================ 2 passed, 1 warnings, 2 error in 7.27 seconds =================
[task 2018-04-22T23:34:31.020Z] 23:34:31     INFO - 
[task 2018-04-22T23:34:31.020Z] 23:34:31     INFO - TEST-PASS | /webdriver/tests/actions/sequence.py | test_no_actions_send_no_events 
[task 2018-04-22T23:34:31.021Z] 23:34:31     INFO - TEST-PASS | /webdriver/tests/actions/sequence.py | test_release_char_sequence_sends_keyup_events_in_reverse 
[task 2018-04-22T23:34:31.021Z] 23:34:31     INFO - TEST-UNEXPECTED-ERROR | /webdriver/tests/actions/sequence.py | test_release_no_actions_sends_no_events - setup error
[task 2018-04-22T23:34:31.021Z] 23:34:31     INFO - session = <Session 35c081a3-3ce3-4ccb-8813-8160d17d10f0>
[task 2018-04-22T23:34:31.022Z] 23:34:31     INFO - test_actions_page = None
[task 2018-04-22T23:34:31.022Z] 23:34:31     INFO - request = <SubRequest 'key_reporter' for <Function 'test_release_no_actions_sends_no_events'>>
[task 2018-04-22T23:34:31.022Z] 23:34:31     INFO - 
[task 2018-04-22T23:34:31.023Z] 23:34:31     INFO -     @pytest.fixture
[task 2018-04-22T23:34:31.023Z] 23:34:31     INFO -     def key_reporter(session, test_actions_page, request):
[task 2018-04-22T23:34:31.023Z] 23:34:31     INFO -         """Represents focused input element from `test_keys_page` fixture."""
[task 2018-04-22T23:34:31.023Z] 23:34:31     INFO - >       input_el = session.find.css("#keys", all=False)
[task 2018-04-22T23:34:31.024Z] 23:34:31     INFO - 
[task 2018-04-22T23:34:31.024Z] 23:34:31     INFO - request    = <SubRequest 'key_reporter' for <Function 'test_release_no_actions_sends_no_events'>>
[task 2018-04-22T23:34:31.024Z] 23:34:31     INFO - session    = <Session 35c081a3-3ce3-4ccb-8813-8160d17d10f0>
[task 2018-04-22T23:34:31.025Z] 23:34:31     INFO - test_actions_page = None
[task 2018-04-22T23:34:31.025Z] 23:34:31     INFO - 
[task 2018-04-22T23:34:31.025Z] 23:34:31     INFO - tests/web-platform/tests/webdriver/tests/actions/conftest.py:27: 
[task 2018-04-22T23:34:31.026Z] 23:34:31     INFO - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Comment 1

[Henrik Skupin [:whimboo][⌚️UTC+2]]

This most likely failed due to the following assertion which caused a content crash:

[task 2018-04-22T23:34:25.690Z] 23:34:25     INFO - PID 856 | [Child 919, Main Thread] ###!!! ABORT: failed to unlink shm: Input/output error: file /builds/worker/workspace/build/src/ipc/chromium/src/base/shared_memory_posix.cc, line 121

I haven't seen such a failure before so I think it could most likely be related to the landing of the patch on bug 1439057. Jed, can you have a look?

Comment 2

[Jed Davis [:jld] ⟨⏰|UTC-7⟩ ⟦he/him⟧]

I already need to back out bug 1447867 (and bug 1439057) because of bug 1455828, once the trees reopen, but:

[task 2018-04-22T23:34:25.637Z] 23:34:25     INFO - PID 856 | Sandbox: Unexpected EOF, op 9 flags 00 path /dev/shm/org.mozilla.ipc.919.1076
[task 2018-04-22T23:34:25.639Z] 23:34:25     INFO - PID 856 | [Child 919, Main Thread] ###!!! ABORT: failed to unlink shm: Input/output error: file /builds/worker/workspace/build/src/ipc/chromium/src/base/shared_memory_posix.cc, line 121

There's no other error message from the file broker before the "Unexpected EOF", the parent process hasn't crashed, and we shouldn't be deleting the ContentParent while the process is still running, so… bug 1243108, maybe?

Also, without the patch in bug 1447867, we would've still been failing to delete the file… and completely ignoring it.  So we'd allocate and dirty a bunch of memory in the shm file and then leak it.

Comment 3

[Jed Davis [:jld] ⟨⏰|UTC-7⟩ ⟦he/him⟧]

(In reply to Jed Davis [:jld] (⏰UTC-6) from comment #2)
> Also, without the patch in bug 1447867, we would've still been failing to
> delete the file… and completely ignoring it.  So we'd allocate and dirty a
> bunch of memory in the shm file and then leak it.

Actually, no.  Further investigation shows that the requests are received and acted upon; the response (and anything else sent on the response socket) is what's lost.  So ignoring the error happens to be the “right” thing here.