Issue metadata
Sign in to add a comment
|
webkit_layout_tests timing out (flake) on chromium.webkit/WebKit Mac10.11 |
||||||||||||||||||||||||
Issue descriptionwebkit_layout_tests failing on chromium.webkit/WebKit Mac10.11 Builders failed on: - WebKit Mac10.11: https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Mac10.11 Builders are periodically timing out (on one of the shards). A normal build seems to take around 10 min. These periodically take 20 min and then time out. Failures are difficult to notice due to Issue 787548 which was consistently failing from r518269 to r518449 (plugins/webview-plugin-type-change.html). Timeout failures (unrelated to that bug) have occurred on these recent builds: - https://luci-milo.appspot.com/buildbot/chromium.webkit/WebKit%20Mac10.11/25761 - https://luci-milo.appspot.com/buildbot/chromium.webkit/WebKit%20Mac10.11/25762 - https://luci-milo.appspot.com/buildbot/chromium.webkit/WebKit%20Mac10.11/25777 - https://luci-milo.appspot.com/buildbot/chromium.webkit/WebKit%20Mac10.11/25778
,
Nov 22 2017
Looking... all the shards that timed out ran on different VMs (except for vm137-m1 that shows up twice) so I think we can rule out that it's not a VM-specific problem.
,
Nov 22 2017
cc'ing maruel@
,
Nov 22 2017
Looking at the actual Swarming tasks; https://chromium-swarm.appspot.com/task?id=39f817d03171d510 https://chromium-swarm.appspot.com/task?id=39f83a53c8a0a910 https://chromium-swarm.appspot.com/task?id=39fa14bd5132a810 https://chromium-swarm.appspot.com/task?id=39fa375b81e12d10 They all timed out due to I/O timeout; that is, no output to stdout for too long; currently 20 minutes. Each task has a help section to show how to reproduce, and there's a "Debug" button to help you reproduce the task. Either either (or both) functionality to help diagnose the problem.
,
Nov 22 2017
that's strange; I have no idea how there would be *no* output at all. Maruel, any theories?
,
Nov 22 2017
There's no output because when SIGTERM is sent to the task's process, it doesn't react and just dies before it had time to write anything to the output directory.
,
Nov 22 2017
,
Jan 24 2018
This is happening somewhat frequently for the non-Webkit Mac bots too. For example, on Mac10.12 Tests, 3 of the last 10 builds have experienced this failure: https://uberchromegw.corp.google.com/i/chromium.mac/builders/Mac10.12%20Tests -> https://uberchromegw.corp.google.com/i/chromium.mac/builders/Mac10.12%20Tests/builds/9466 https://uberchromegw.corp.google.com/i/chromium.mac/builders/Mac10.12%20Tests/builds/9464 https://uberchromegw.corp.google.com/i/chromium.mac/builders/Mac10.12%20Tests/builds/9470 I notice that all three of these failures are on "m4" but I don't know if that's significant. Tagging for trooper attention.
,
Jan 24 2018
,
Jan 24 2018
I don't think -m4 is relevant here.
,
Jan 24 2018
Believe we've seen this before -- shard timeout after 20 minutes but, according the log, shortly after RWT starts. |
|||||||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||||||
Comment 1 by olka@chromium.org
, Nov 22 2017