StorageMonitorLinuxTests flakily timing out since 2017-11-24 |
||||
Issue description^^ This affects several tests in the suite, e.g.: - StorageMonitorLinuxTest.Removable: https://chromium-swarm.appspot.com/task?id=3a2c805e6017fc10 - StorageMonitorLinuxTest.MultipleMountPointsWithNonDCIMDevices: https://chromium-swarm.appspot.com/task?id=3a2c7ad3ff237510 - StorageMonitorLinuxTest.DevicePartitionSize: https://chromium-swarm.appspot.com/task?id=3a2c6cac76864210 - StorageMonitorLinuxTest.MultiDevicesMultiMountPoints: https://chromium-swarm.appspot.com/task?id=3a2c6d9e5997ea10 etc. You can see more of them if you look at any 8+minute shard in https://chromium-swarm.appspot.com/tasklist?f=name%3Acomponents_unittests&f=buildername%3Alinux_chromium_rel_ng&l=50&n=true&s=created_ts%3Adesc The net result of this has been to increase the 90th percentile shard execution time of components_unittests swarming shards (see graph attached to https://bugs.chromium.org/p/chromium/issues/detail?id=788875#c8). This appears to have started shortly before 2017-11-24T01:00:00Z (see attached). Given that, I suspect https://chromium.googlesource.com/chromium/src/+/d4723a3dfd0215a72507d10ce0bd0e75f5f76bb6 is responsible.
,
Dec 1 2017
also +cc //components/storage_monitor owners
,
Dec 1 2017
r487141 was the last major change to StorageMonitorLinuxTests, so all I'm going to do is point fingers at the suspects. :) (Also, Tommy is OOO)
,
Dec 6 2017
Since base::test::ScopedTaskEnvironment is a test-only class, this isn't a performance regression that affects users. gab@/robliao@: We should avoid introducing delays in base::test::ScopedTaskEnvironment since it's used in performance tests. |
||||
►
Sign in to add a comment |
||||
Comment 1 by jbudorick@chromium.org
, Dec 1 2017