Run raptor-youtube-playback jobs for Chrome in CI
Categories
(Testing :: Raptor, task, P2)
Tracking
(firefox73 fixed)
Tracking | Status | |
---|---|---|
firefox73 | --- | fixed |
People
(Reporter: whimboo, Assigned: marauder)
References
(Blocks 1 open bug)
Details
Attachments
(1 file)
It would be nice to have the playback performance tests running against Chromium to get comparison numbers with Firefox on desktop and mobile.
Given that also Chrome/Chromium has blocking auto-play turned on, we will have to run those tests muted.
We could already get the tests enabled by specifying the same test_url
as for android, or we wait until bug 1554966 has been done, so we know if we have to run with audio enabled.
Comment 1•5 years ago
|
||
Bugbug thinks this bug is a enhancement, but please change it back in case of error.
Updated•5 years ago
|
Updated•5 years ago
|
Comment 2•5 years ago
|
||
:jmaher do you have any concerns over turning on additional tests for Chromium? We could consider running these once/week.
Comment 3•5 years ago
|
||
a few jobs once/week isn't going to kill things- especially on a weekend :)
Assignee | ||
Updated•5 years ago
|
Assignee | ||
Comment 5•5 years ago
|
||
Created a cron job for this task and added changes to yml and ini files.
Assignee | ||
Comment 6•5 years ago
|
||
This cron job will generate the following target_tasks.json:
[
"test-windows10-64-shippable/opt-raptor-youtube-playback-chromium-e10s",
"test-macosx1014-64-shippable/opt-raptor-youtube-playback-chromium-e10s",
"test-linux64-shippable/opt-raptor-youtube-playback-chromium-e10s",
"test-windows7-32-shippable/opt-raptor-youtube-playback-chromium-e10s"
]
Assignee | ||
Comment 7•5 years ago
|
||
Comment 8•5 years ago
•
|
||
Can you please provide an update here :marauder?
Assignee | ||
Comment 9•5 years ago
|
||
Hi Dave,
I did several tests locally using a local server and the repository from https://github.com/mozilla/perf-youtube-playback
I tested on both branches master and raptor.
The tests are running ok on Chrome but is a perma fail on Chromium.
A few more details for Chrome:
https://phabricator.services.mozilla.com/D41553#1385409
More details for Chromium:
https://phabricator.services.mozilla.com/D41553#1396207
Comment 10•5 years ago
|
||
Let's switch this to Chrome then as that's our preferred target anyway. We'll just need to depend on Greg's work in bug 1563708.
Updated•5 years ago
|
Assignee | ||
Comment 11•5 years ago
|
||
Finished the transition from Chromium to Chrome.
Push to try:
https://treeherder.mozilla.org/#/jobs?repo=try&revision=cb87698cdbfe3b1e93e5b23dcab2046b9ab317b9
Comment 12•5 years ago
|
||
Marian: Can you provide an update on this?
Assignee | ||
Comment 13•5 years ago
|
||
Hi Dave,
I did a git bisection locally for the youtube playback repo: https://phabricator.services.mozilla.com/D41553#1499340
to find what is causing the permanent failures on Chrome, but i didn't find any issues with the tests.
And it seems that what we have running on the production server (https://yttest.prod.mozaws.net) has some issues.
https://phabricator.services.mozilla.com/D41553#1523519
Henrik asked Tarek to bring up the server that ran a while back at
http://yttest.dev.mozaws.net/2019/main.html?test_type=playbackperf-test
to be able to do that bisection and run on the server not on localhost.
Comment 14•5 years ago
|
||
From experience, mentions in Phabricator rarely get the attention of the person. It's best to use needinfo in bugzilla. Tarek: could you take a look over the comment from Henrik in https://phabricator.services.mozilla.com/D41553#1523519
Comment 15•5 years ago
|
||
Thanks Dave, sorry for not answering in phab. I have restarted the youtube dev instance I had (luckily I did not erase it!)
http://yttest.dev.mozaws.net/2019/main.html?test_type=playbackperf-test is back online
@marauder let me know if I can be of further assistance - I can give you access to that dev box if you need.
If you can't reach the URL (dns just changed) replace the domain by ec2-34-244-207-63.eu-west-1.compute.amazonaws.com
Assignee | ||
Comment 16•5 years ago
|
||
Hei Tarek,
I can access that URL but i will also need that dev box credentials to jump on raptor branch and check each commit to see which one caused the permanent failures on Chrome.
Thanks!
Reporter | ||
Updated•5 years ago
|
Assignee | ||
Comment 18•5 years ago
|
||
I rebased the patch and did a push to try:
https://treeherder.mozilla.org/#/jobs?repo=try&revision=8fb25c17dfbcf8f75c1078413b701a851ddcec32
Comment 19•5 years ago
|
||
Comment 20•5 years ago
|
||
bugherder |
Assignee | ||
Comment 21•5 years ago
|
||
The cron job is running, the results for Dec 11 will be available here:
Reporter | ||
Comment 22•5 years ago
|
||
Results are totally worse for Chrome on Mac:
I assume that this might be the same problem with the Mac Minis and Firefox. I don't have that other bug handy. Do you have? Might be good to add this info over there.
How do the other platforms compare with Firefox?
Assignee | ||
Comment 23•5 years ago
|
||
Hey Henrik,
Here are the results for various platforms:
https://treeherder.mozilla.org/perf.html#/graphs?highlightAlerts=1&series=mozilla-central,2054686,1,10&series=mozilla-central,2044136,1,10&series=mozilla-central,2044425,1,10&series=mozilla-central,2045003,1,10&series=mozilla-central,2043847,1,10&series=mozilla-central,2044714,1,10&series=mozilla-central,2043558,1,10&timerange=2592000
Mac OS 10 has the highest values (if we exclude the results for Windows 10 Aarch64).
Can you give more details about Mac Minis ?
Thanks!
Reporter | ||
Comment 24•5 years ago
|
||
I'm not working on this project anymore for a couple of months. So please work with stakeholders of your team or developers to continue. Also why don't you follow-up on my comment from bug 1604207?
Description
•