A flaky webkit layout test is reported to be in step "gfx_unittests (with patch)" |
|||
Issue descriptionFindit received a Chromium-Try-Flakes report for flaky test 'virtual/mojo-loading/http/tests/serviceworker/chromium/register-error-messages.html' which is interesting. The step name is "gfx_unittests (with patch)" instead of "webkit_tests". This doesn't lead to error on Findit side, but resulted in a confusing analysis: https://findit-for-me.appspot.com/waterfall/check-flake?master_name=chromium.linux&builder_name=Linux%20Tests&build_number=49797&step_name=gfx_unittests&test_name=virtual/mojo-loading/http/tests/serviceworker/chromium/register-error-messages.html The filed bug by chromium-try-flakes is: 672013 The suspicious failed CQ try is: https://luci-milo.appspot.com/buildbot/tryserver.chromium.linux/linux_chromium_rel_ng/351032 In this CQ try, the test failed in step "webkit_tests (with patch)", and "gfx_unittests (with patch)" failed as well. The CQ try was purple. The report received on Findit side is: build_number=351032, builder_name=u'linux_chromium_rel_ng', master_name=u'tryserver.chromium.linux', reported_time=datetime.datetime(2016, 12, 7, 11, 51, 33, 850040), step_name=u'gfx_unittests (with patch)' At the moment, I will add a filtering on Findit side to workaround this.
,
Dec 14 2016
Indeed, I have nailed it down to the FlakyRun entity in datastore: https://pantheon.corp.google.com/datastore/entities/edit?key=0%2F%7C8%2FFlakyRun%7C19%2Fid:5636156676898816&project=chromium-try-flakes&ns=&kind=FlakyRun&filter=7%2F__key__%7CKEY%7CEQ%7C46%2FKey(Namespace(%27%27),%20FlakyRun,%205636156676898816). We do store information about all flaky steps in a FlakyRun and then add matching flaky runs to the Flake entity. We should filter runs to match the name of the Flake when sending the data to FindIt.
,
Dec 14 2016
,
Dec 14 2016
I'd like to test this on staging instance before deploying to prod, but currently staging instance still can't talk to monorail-staging, so this is blocked on issue monorail:1840 .
,
Dec 14 2016
The following revision refers to this bug: https://chromium.googlesource.com/infra/infra.git/+/fefe35bed90185d8ca6c0cc8ef72a43c38a97fca commit fefe35bed90185d8ca6c0cc8ef72a43c38a97fca Author: Sergiy Byelozyorov <sergiyb@chromium.org> Date: Wed Dec 14 15:34:02 2016 Only report occurrences matching the Flake's name to FindIt BUG= 673898 Change-Id: Id56b77d099202e2cb0a55fcddb66d217206704af Reviewed-on: https://chromium-review.googlesource.com/419516 Reviewed-by: Shuotao Gao <stgao@chromium.org> Reviewed-by: Andrii Shyshkalov <tandrii@chromium.org> Commit-Queue: Sergiy Byelozyorov <sergiyb@chromium.org> [modify] https://crrev.com/fefe35bed90185d8ca6c0cc8ef72a43c38a97fca/appengine/chromium_try_flakes/findit/findit.py [modify] https://crrev.com/fefe35bed90185d8ca6c0cc8ef72a43c38a97fca/appengine/chromium_try_flakes/findit/test/findit_test.py
,
Dec 14 2016
I've deployed this to prod directly, since I seem to get no traction on the dependent bug.
,
Dec 14 2016
Many thanks! We will monitor on our end. If we see new cases, we will update this bug :) |
|||
►
Sign in to add a comment |
|||
Comment 1 by st...@chromium.org
, Dec 13 2016