New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 809063 link

Starred by 1 user

Issue metadata

Status: Fixed
Owner:
Closed: Feb 2018
Cc:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 1
Type: ----



Sign in to add a comment

v8.detached_context_age_in_gc failing on all platforms.

Project Member Reported by sheriff-...@appspot.gserviceaccount.com, Feb 5 2018

Issue description

Filed by sheriff-o-matic@appspot.gserviceaccount.com on behalf of sullivan@google.com

v8.detached_context_age_in_gc/Docs_(1_open_document_tab) and 1 other(s) in v8.detached_context_age_in_gc failing on multiple builders

Builders failed on: 
- Android Nexus5X Perf: 
  https://build.chromium.org/p/chromium.perf/builders/Android%20Nexus5X%20Perf
- Android Nexus6 Perf: 
  https://build.chromium.org/p/chromium.perf/builders/Android%20Nexus6%20Perf
- Linux Perf: 
  https://build.chromium.org/p/chromium.perf/builders/Linux%20Perf
- Mac 10.11 Perf: 
  https://build.chromium.org/p/chromium.perf/builders/Mac%2010.11%20Perf
- Mac 10.12 Perf: 
  https://build.chromium.org/p/chromium.perf/builders/Mac%2010.12%20Perf
- Mac Air 10.11 Perf: 
  https://build.chromium.org/p/chromium.perf/builders/Mac%20Air%2010.11%20Perf
- Mac Pro 10.11 Perf: 
  https://build.chromium.org/p/chromium.perf/builders/Mac%20Pro%2010.11%20Perf
- Mac Retina Perf: 
  https://build.chromium.org/p/chromium.perf/builders/Mac%20Retina%20Perf

Example logs:
https://logs.chromium.org/v/?s=chrome%2Fbb%2Fchromium.perf%2FMac_10.11_Perf%2F1791%2F%2B%2Frecipes%2Fsteps%2Fv8.detached_context_age_in_gc_on_Intel_GPU_on_Mac_on_Mac-10.11%2F0%2Fstdout

All I see in the stack is "TimeoutException: Timed out while waiting 60s for IsJavaScriptExpressionTrue." stack, but when I look at the screenshots I see 404 error messages trying to display the page: https://console.developers.google.com/m/cloudstorage/b/chrome-telemetry-output/o/444ac85c-0a44-11e8-92d9-10ddb199a53a

Will kick off a bisect.
 
Project Member

Comment 2 by bugdroid1@chromium.org, Feb 5 2018

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/8b1aeba0f679d4fdb237bdedd33bf066db3a0213

commit 8b1aeba0f679d4fdb237bdedd33bf066db3a0213
Author: Annie Sullivan <sullivan@chromium.org>
Date: Mon Feb 05 19:09:53 2018

Skip v8.detached_context_age_in_gc on all platforms.

It is failing all test cases on all platforms.

Bug:  809063 
Change-Id: I64f09cc842edb6e1541508067b9b9babcc82a3e4
Tbr: nednguyen@google.com
Reviewed-on: https://chromium-review.googlesource.com/901903
Reviewed-by: Annie Sullivan <sullivan@chromium.org>
Reviewed-by: Charlie Andrews <charliea@chromium.org>
Commit-Queue: Annie Sullivan <sullivan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#534444}
[modify] https://crrev.com/8b1aeba0f679d4fdb237bdedd33bf066db3a0213/tools/perf/expectations.config

Cc: nedngu...@google.com sunn...@chromium.org chiniforooshan@chromium.org
Owner: chiniforooshan@chromium.org
Status: Assigned (was: Available)
📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/1299a789840000

Add UI stories to top_25_smooth by chiniforooshan@chromium.org
chromium @ 2b63134791c3910cb5f3805ffbb65b073ae57816

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions
I don't understand how crrev/c/895411 can cause v8.detached_context_age_in_gc to fail?

I looked at one of the logs:

https://chromium-swarm.appspot.com/task?id=3b83e1d1d7d8e110&refresh=10&show_raw=1

The suspecious error to me is:
	E/chromium( 4537): [ERROR:devtools_http_handler.cc(249)] Cannot start http server for devtools. Stop devtools.

Owner: dtu@chromium.org
I think this is a bad bisect :-/

Dave: can you look into this?
Cc: -nedngu...@google.com nednguyen@chromium.org

Comment 7 by dtu@chromium.org, Feb 7 2018

Cc: -chiniforooshan@chromium.org
Oh, yep, looks like it failed because the bots were overloaded (go/catabug/4227)
I'll re-run the job.
Cc: nedngu...@google.com chiniforooshan@chromium.org
Owner: chiniforooshan@chromium.org
📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/15deada1840000

Add UI stories to top_25_smooth by chiniforooshan@chromium.org
chromium @ 2b63134791c3910cb5f3805ffbb65b073ae57816

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions

Comment 10 by dtu@chromium.org, Feb 7 2018

Ah, it's because v8.detached_context_age_in_gc is failing on "Docs_(1_open_document_tab)", which is in `top_25_012.wprgo`, and the above CL contains a change to that wprgo.
Labels: -Pri-2 Pri-1
Ooops, thanks for catching that, Dave!

I bump this up to P1 because we are losing coverage
Thanks! So, is this because file names in //tools/perf/page_sets/data/top_25.json and //tools/perf/page_sets/data/top_25_smooth.json overlap? This looks strange, since, as far as I remember, the name top_25_012.wprgo was automatically generated by some script (record_wpr?).

Should I regenerate the overwritten wprgo file (how)?
Ehsan: you can record the file with:

./tools/perf/record_wpr top25_smooth_page_set --browser=system --story-filter=Docs

./tools/perf/record_wpr top25_page_set --browser=system --story-filter=Docs

Lemme know if you have any problem with running those
> ./tools/perf/record_wpr top25_smooth_page_set --browser=system --story-filter=Docs

Should I do this one too? Docs_(1_open_document_tab) uses top_25_009.wprgo in top_25_smooth.json, not top_25_012.wprgo (the overwritten file).
Cc: -nedngu...@google.com
If it's not broken, I would just leave it as-is :P
Project Member

Comment 16 by bugdroid1@chromium.org, Feb 9 2018

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/1865f3e0ebeaca1598f3f83057d4575609fbefbb

commit 1865f3e0ebeaca1598f3f83057d4575609fbefbb
Author: Ehsan Chiniforooshan <chiniforooshan@chromium.org>
Date: Fri Feb 09 18:39:05 2018

Telemetry: fix v8.detached_context_age_in_gc

It started to fail because top_25_012.wprgo was accidentally overwritten in
crrev/c/895411.

Regenerated the file by
./tools/perf/record_wpr top25_page_set --browser=system --story-filter=Docs
and then renaming the output to avoid overwriting top_25_013.

Bug:  809063 
Change-Id: I9bdfbf2248641f710979d02605af4d277e4408b1
Reviewed-on: https://chromium-review.googlesource.com/911808
Commit-Queue: Ehsan Chiniforooshan <chiniforooshan@chromium.org>
Reviewed-by: Ned Nguyen <nednguyen@google.com>
Cr-Commit-Position: refs/heads/master@{#535761}
[modify] https://crrev.com/1865f3e0ebeaca1598f3f83057d4575609fbefbb/tools/perf/page_sets/data/top_25.json
[add] https://crrev.com/1865f3e0ebeaca1598f3f83057d4575609fbefbb/tools/perf/page_sets/data/top_25_014.wprgo.sha1
[add] https://crrev.com/1865f3e0ebeaca1598f3f83057d4575609fbefbb/tools/perf/page_sets/data/top_25_015.wprgo.sha1

Project Member

Comment 17 by bugdroid1@chromium.org, Feb 12 2018

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/7017c63a904048a93b3d6f6532bfc6e0bc3d7882

commit 7017c63a904048a93b3d6f6532bfc6e0bc3d7882
Author: Ehsan Chiniforooshan <chiniforooshan@chromium.org>
Date: Mon Feb 12 16:16:59 2018

Re-enable v8.detached_context_age_in_gc

The fix, crrev/c/911808, is landed.

NOTRY=true

Bug:  809063 
Change-Id: I180e93f1528d3a90e90e154a8a8c3f9543456a97
Reviewed-on: https://chromium-review.googlesource.com/912151
Commit-Queue: Ehsan Chiniforooshan <chiniforooshan@chromium.org>
Reviewed-by: Ned Nguyen <nednguyen@google.com>
Reviewed-by: Ehsan Chiniforooshan <chiniforooshan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#536092}
[modify] https://crrev.com/7017c63a904048a93b3d6f6532bfc6e0bc3d7882/tools/perf/expectations.config

Status: Fixed (was: Assigned)

Sign in to add a comment