Page URL: https://findit-for-me.appspot.com/waterfall/flake?redirect=1&key=ag9zfmZpbmRpdC1mb3ItbWVy1AELEhdNYXN0ZXJGbGFrZUFuYWx5c2lzUm9vdCKdAWNocm9taXVtLndpbi9XaW43IFRlc3RzIChkYmcpKDEpLzY5Nzc0L25ldHdvcmtfc2VydmljZV9icm93c2VyX3Rlc3RzL1JHVmpiR0Z5WVhScGRtVk9aWFJTWlhGMVpYTjBRbkp2ZDNObGNsUmxjM1F1VUdGblpVRnNiRzkzYVc1blFWQkpYMEp5YjNkelpYSlNaWEYxWlhOMGN5OHcMCxITTWFzdGVyRmxha2VBbmFseXNpcxgBDA Description: I kicked off an analysis yesterday and it's still running.
There is a regression in Findit that it now assumes that the test always exists in an older revision: https://chromium.googlesource.com/infra/infra/+/ef313bedc91693adc0db825dde62ac45e1cde1a1/appengine/findit/services/flake_failure/flake_test_results.py#34 For this particular flaky test "DeclarativeNetRequestBrowserTest.PageAllowingAPI_BrowserRequests/0", it doesn't exist in the browser_tests binary at r567856 while it exists at r567866. So the culprit adding the test should be in https://crrev.com/567856..567866?pretty=fuller i.e. https://chromium.googlesource.com/chromium/src/+/2860433068296b76190c08acdf857a0a5d4cabc3
The following revision refers to this bug: https://chromium.googlesource.com/infra/infra/+/6aac02d1195e9b108a8caabedbb964cec0cf0efe commit 6aac02d1195e9b108a8caabedbb964cec0cf0efe Author: Shuotao Gao <stgao@chromium.org> Date: Tue Jun 19 16:53:36 2018 [Findit] Handle newly-added tests properly for flake analysis. In older revisions, newly-added tests might not exist. Bug: 854014 Change-Id: If1dab4af0715b9a3b54503ec1571609abf7aeb12 Reviewed-on: https://chromium-review.googlesource.com/1106422 Reviewed-by: Chan Li <chanli@chromium.org> Commit-Queue: Shuotao Gao <stgao@chromium.org> [modify] https://crrev.com/6aac02d1195e9b108a8caabedbb964cec0cf0efe/appengine/findit/services/flake_failure/test/flake_test_results_test.py [modify] https://crrev.com/6aac02d1195e9b108a8caabedbb964cec0cf0efe/appengine/findit/services/flake_failure/flake_test_results.py [modify] https://crrev.com/6aac02d1195e9b108a8caabedbb964cec0cf0efe/appengine/findit/services/flake_failure/flake_swarming.py
The bug is fixed as verified by another analysis https://findit-for-me.appspot.com/waterfall/flake?key=ag9zfmZpbmRpdC1mb3ItbWVyowELEhdNYXN0ZXJGbGFrZUFuYWx5c2lzUm9vdCJtY2hyb21pdW0ubGludXgvQ2FzdCBBdWRpbyBMaW51eC8xMDM3My9iYXNlX3VuaXR0ZXN0cy9WR0Z6YTFOamFHVmtkV3hsY2tsdGNHeFVaWE4wTGtsa1pXNTBhV1pwWVdKc1pWTjBZV05yY3c9PQwLEhNNYXN0ZXJGbGFrZUFuYWx5c2lzGA0M However, I still need to clean up those analyses left in running status before we could rerun them.
Comment 1 by st...@chromium.org
, Jun 19 2018Status: Assigned (was: Unconfirmed)