New issue
Advanced search Search tips

Issue 729088 link

Starred by 1 user

Issue metadata

Status: WontFix
Owner: ----
Closed: Oct 25
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 1
Type: Bug-Regression



Sign in to add a comment

system_health.common_desktop failure on chromium.perf/win-high-dpi since 5/23

Project Member Reported by charliea@chromium.org, Jun 2 2017

Issue description

Link to example log: https://luci-logdog.appspot.com/v/?s=chrome%2Fbb%2Fchromium.perf%2FWin_10_High-DPI_Perf%2F663%2F%2B%2Frecipes%2Fsteps%2Fsystem_health.common_desktop_on_Intel_GPU_on_Windows_on_Windows-10-10240%2F0%2Fstdout

I think this might be due to running out of memory based on the following excerpt from the log:

Traceback (most recent call last):
  File "c:\b\s\w\ir\third_party\catapult\tracing\tracing\trace_data\..\..\bin\trace2html", line 14, in <module>
    sys.exit(trace2html.Main(sys.argv))
  File "c:\b\s\w\ir\third_party\catapult\tracing\tracing_build\trace2html.py", line 51, in Main
    config_name=args.config_name)
  File "c:\b\s\w\ir\third_party\catapult\tracing\tracing_build\trace2html.py", line 120, in WriteHTMLForTracesToFile
    WriteHTMLForTraceDataToFile(trace_data_list, title, output_file, config_name)
  File "c:\b\s\w\ir\third_party\catapult\tracing\tracing_build\trace2html.py", line 103, in WriteHTMLForTraceDataToFile
    output_file, load_sequence, title, extra_scripts=scripts)
  File "c:\b\s\w\ir\third_party\catapult\third_party\py_vulcanize\py_vulcanize\generate.py", line 266, in GenerateStandaloneHTMLToFile
    extra_script.WriteToFile(output_file)
  File "c:\b\s\w\ir\third_party\catapult\tracing\tracing_build\trace2html.py", line 71, in WriteToFile
    output_file.write(b64_content)
  File "c:\b\depot_tools\python276_bin\lib\codecs.py", line 691, in write
    return self.writer.write(data)
  File "c:\b\depot_tools\python276_bin\lib\codecs.py", line 351, in write
    data, consumed = self.encode(object, self.errors)
MemoryError
 
Telemetry's stated trace sizes for the various trace parts:

Trace sizes in bytes: {
  'cpuSnapshots': 801675L, (801kB)
  'traceEvents': 1202255205L, (1.2GB)
  'telemetry': 398759L, (398kB)
  'tabIds': 40L, (40B)
  'powerTraceAsString': 14760615L (14MB)
}
Note that this has caused no data to be uploaded for this benchmark at all since 5/23. This is bad.
It seems the story causing it to fail is: browse:media:flickr_infinite_scroll. I'm not confident that other stories won't fail, though, if this one passes.
I think we might be using the 32 bit version of Python on all Windows bots, which could explain why we're running into problems trying to use O(1GB) memory.
Hmmm, strange. It looks like it hasn't actually been failing since 5/23, but it looks like the about_blank story was load:chrome:blank, which was last run on 6/2. The system_health.common stories are still having problems, but I think those problems are adequately covered by https://bugs.chromium.org/p/chromium/issues/detail?id=729089
Components: Speed>Metrics>SystemHealthRegressions
Components: -Speed>Metrics>SystemHealthRegressions
Components: Speed>Metrics>SystemHealthRegressions
Status: WontFix (was: Untriaged)
This builder has been decommissioned.

Sign in to add a comment