New issue
Advanced search Search tips

Issue 688180 link

Starred by 1 user

Issue metadata

Status: Fixed
Owner:
Closed: Feb 2017
Cc:
EstimatedDays: ----
NextAction: ----
OS: Android
Pri: 1
Type: Bug-Regression



Sign in to add a comment

Lots of benchmarks failing on Android

Project Member Reported by dtu@chromium.org, Feb 3 2017

Issue description

Cc: waff...@chromium.org
Owner: waff...@chromium.org

=== Auto-CCing suspected CL author waffles@chromium.org ===

Hi waffles@chromium.org, the bisect results pointed to your CL, please take a look at the
results.


=== BISECT JOB RESULTS ===
Test failure found with culprit

Suspected Commit
  Author : waffles
  Commit : 9fefd20253ceb6af4fc3604ca29fcafc4c8454cd
  Date   : Wed Feb 01 18:56:45 2017
  Subject: Update component updater URLs.

Bisect Details
  Configuration: android_nexus5_perf_bisect
  Benchmark    : blink_perf.layout
  Metric       : ArabicLineLayout/ArabicLineLayout

Revision             Exit Code      N
chromium@447553      0 +- N/A       1      good
chromium@447559      0 +- N/A       1      good
chromium@447562      0 +- N/A       1      good
chromium@447563      0 +- N/A       1      good
chromium@447564      1 +- N/A       1      bad       <--
chromium@447565      1 +- N/A       1      bad
chromium@447577      1 +- N/A       1      bad
chromium@447600      1 +- N/A       1      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=attach-inlines-2.html blink_perf.layout

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8988720566678374048

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5000228190552064


| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Tests>AutoBisect.  Thank you!

Comment 4 by dtu@chromium.org, Feb 3 2017

Cc: dtu@chromium.org
Summary: Lots of benchmarks failing on Android (was: blink_perf failing on 8 builders)
Bisect result looks pretty clear. I'm reverting.
https://codereview.chromium.org/2675903002/

=== BISECT JOB RESULTS ===
Test failure found with culprit

Suspected Commit
  Author : waffles
  Commit : 9fefd20253ceb6af4fc3604ca29fcafc4c8454cd
  Date   : Wed Feb 01 18:56:45 2017
  Subject: Update component updater URLs.

Bisect Details
  Configuration: android_nexus5X_perf_bisect
  Benchmark    : blink_perf.events
  Metric       : EventsDispatching/EventsDispatching

Revision             Exit Code      N
chromium@447537      0 +- N/A       20      good
chromium@447557      0 +- N/A       20      good
chromium@447562      0 +- N/A       20      good
chromium@447563      0 +- N/A       20      good
chromium@447564      1 +- N/A       20      bad       <--
chromium@447565      1 +- N/A       20      bad
chromium@447567      1 +- N/A       20      bad
chromium@447577      1 +- N/A       20      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests blink_perf.events

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8988724288236728352

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5837319128481792


| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Tests>AutoBisect.  Thank you!
The results are hard to argue with but I don't understand at all how this change could have caused timeouts on perf tests, especially for Android only.

Is there Android source that is not visible in cs.chromium.org?
I can't see the errors for the linked bots. I suspect that this "regression" is a dup of  Issue 688230 

We are going over the arbirtary limit for the size of command lines on Android

Sign in to add a comment