gtest perf tests failing "TEST RESULTS WERE INVALID", "shard #0 expired, not enough capacity" |
||||||
Issue descriptionFiled by sheriff-o-matic@appspot.gserviceaccount.com on behalf of crouleau@chromium.org e9442cd7-160c-498c-8bff-6c06ca33540e failures include media_perftests, load_library_perf_tests, and angle_perftests, among others. Builders failed on: - Win 7 Nvidia GPU Perf: https://ci.chromium.org/buildbot/chromium.perf/Win%207%20Nvidia%20GPU%20Perf - Win 7 Perf: https://ci.chromium.org/buildbot/chromium.perf/Win%207%20Perf - Win 7 Perf: https://ci.chromium.org/buildbot/chromium.perf/Win%207%20Perf - win-10-perf: https://ci.chromium.org/buildbot/chromium.perf/win-10-perf
,
Aug 28
Only test results-related change I could find was crrev.com/585905. Brian, do you think your change could have caused this?
,
Aug 28
Ah nah, that's a java-only change.
,
Aug 28
For https://ci.chromium.org/buildbot/chromium.perf/Win%207%20Nvidia%20GPU%20Perf/ first failure had a possible culprit range of 586036 to 586122. but https://ci.chromium.org/buildbot/chromium.perf/win-10-perf/354 has a revision range of 585882 to 585945. Since those ranges are completely distinct, I wonder if this is not caused by a change to Chromium codebase.
,
Aug 28
The root cause seems to be "shard #0 expired, not enough capacity" https://chrome-swarming.appspot.com/task?id=3f9877d91302f010&refresh=10&show_raw=1 Looks like tests are expiring after sitting in PENDING state for 2 hours.
,
Aug 28
,
Aug 28
Could this be a problem with soft device affinity? I just don't understand why the shards are expiring. Ned, can you take a look or re assign to someone who can help?
,
Aug 28
|
||||||
►
Sign in to add a comment |
||||||
Comment 1 by crouleau@chromium.org
, Aug 28