search

teerapap / grunt-protractor-runner

A Grunt plugin for running protractor runner.
MIT License
149 stars 123 forks source link

breaks Protractor sharding #79

Open dborin opened 10 years ago

dborin commented 10 years ago

Does the runner support Protractor sharding? I'm seeing a semi-frequent (50%+) instances where even with one "shard", the Protractor tests fail to ever exit because they're detecting a remaining webdriver instance, even after logging that it was shutting down the remaining instance

Shutting down selenium standalone server. [launcher] 1 instance(s) of WebDriver still running Shutting down selenium standalone server. [launcher] 1 instance(s) of WebDriver still running

Please let me know what additional information you'd need

zarroba commented 10 years ago

I'm using protractor 1.0.0 with grunt-protractor-runner 1.1.0 with sharding enabled with 4 instances on our pipeline and haven't had any issue with it.

Could it be that you have a test in a infinite loop or something?

dborin commented 10 years ago

This doesn't happen every time we run the tests. Also, without sharding enabled we never see this issue when using the runner. I'll make sure we're using the versions you mentioned and if it continues I will let you know (should take a day or two at most)

On Tue, Aug 5, 2014 at 3:32 AM, José Tavares notifications@github.com wrote:

I'm using protractor 1.0.0 with grunt-protractor-runner 1.1.0 with sharding enabled with 4 instances on our pipeline and haven't had any issue with it.

Could it be that you have a test in a infinite loop or something?

— Reply to this email directly or view it on GitHub https://github.com/teerapap/grunt-protractor-runner/issues/79#issuecomment-51179925 .

ttaghavi commented 10 years ago

We have a similar behaviour which started appearing during the week-end, though we use a different setup (grunt-concurrent to parallelize our tests against saucelabs).

It is a problem that seems to happen randomly, more than 50% in our case. One of the tests "hangs up" forever while on saucelabs side it is timing out (no command received for 90s).

We are still in the process of debugging (downgrading grunt-protractor-runner and other libs, rebuilding the the node environment, disabling concurrency, etc...), but as of yet, no luck.

I am currently not even sure if the problem is grunt-protractor-runner, but since the description of the problem is so similar, I thought to comment here.

dborin commented 10 years ago

@hankduan says that a recent fix (#1167 for angular/protractor) could/should fix the sharding issue. I'll have to wait until they merge it to master before I can renable sharding and validate it.