Win (10?) swarming bots sometimes fail to import "win32pipe" during webkit_layout_tests |
|||||||||
Issue descriptionI've noticed Windows 10 (not sure if 7 is affected as well) swarming jobs sometimes fail because the win32pipe Python module cannot be found. Two recent examples: https://ci.chromium.org/buildbot/tryserver.blink/win10_blink_rel/5958 https://chromium-swarm.appspot.com/task?id=3b280f6dc3005710&refresh=10&show_raw=1 https://ci.chromium.org/buildbot/tryserver.blink/win10_blink_rel/5968 https://chromium-swarm.appspot.com/task?id=3b280f6a83e29810&refresh=10&show_raw=1 The failures seem quite random. Only a portion of builds had issues, and other shards in the broken build passed. The problem is quite concerning to webkit_layout_tests because it would cause incomplete results and confuse the rebaseline tool. On one hand, this issue needs to be investigated; on the other hand, I think it's perhaps better to make the whole build (step) fail if some Swarming shard(s) fail without isolated output.
,
Jan 22 2018
Saw this on waterfall as well: https://ci.chromium.org/buildbot/chromium.webkit/WebKit%20Win10/29968 https://chromium-swarm.appspot.com/task?id=3b376d8c2129c710
,
Jan 22 2018
,
Jan 22 2018
,
Jan 22 2018
maruel, can you take a look and reassign if needed? This might belong to Infra>Client>Chrome but I'm not sure.
,
Jan 22 2018
I think we need a "base image" component, which should preferably under Infra>Labs.
,
Jan 22 2018
I think probably the thing to do here is to ensure that win32pipe (and whatever other modules we need) are listed in src's //.vpython ?
,
Jan 23 2018
Unassigning so this shows up in CCI triage.
,
Jan 23 2018
Issue 804657 has been merged into this issue.
,
Jan 23 2018
,
Jan 23 2018
This appears to be flaky because it only fails on swarmXXX-c4 bots: http://shortn/_sNpfdgadkq
,
Jan 23 2018
https://chrome-internal-review.googlesource.com/c/infradata/config/+/550509 added those bots to pool:Chrome
,
Jan 23 2018
... or at least, made them explicitly listed as pool:Chrome?
,
Jan 23 2018
,
Jan 23 2018
The following revision refers to this bug: https://chrome-internal.googlesource.com/infradata/config/+/ca7eea09e642a5b7a62c20f240b3aee50e492dfc commit ca7eea09e642a5b7a62c20f240b3aee50e492dfc Author: John Budorick <jbudorick@google.com> Date: Tue Jan 23 16:30:12 2018
,
Jan 23 2018
CL in #15 has mitigated this. The error should no longer appear, but we should likely still do what Dirk suggested in #7.
,
Jan 27 2018
The following revision refers to this bug: https://chrome-internal.googlesource.com/infradata/config/+/4d73d917d307740f9293fad0a66a993292de707c commit 4d73d917d307740f9293fad0a66a993292de707c Author: Vadim Shtayura <vadimsh@chromium.org> Date: Sat Jan 27 00:43:34 2018 |
|||||||||
►
Sign in to add a comment |
|||||||||
Comment 1 by robertma@chromium.org
, Jan 22 2018