-
Notifications
You must be signed in to change notification settings - Fork 30
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
When browser disconnects, remaining tests are skipped #42
Comments
Hi - I have setup karma-parallel on our test suite as well. Approx 2700 tests. On the first run the sharding is working great and it has massively decreased our suite run times. However as above we are getting false positives when the runner disconnects from chrome. It will run again but not run the tests in the shard that disconnected. Unfortunately this means we can't use this day to day at the moment. |
We are running 500 tests and it has reduced our time as well but the browser disconnect issue is preventing us from running this within our CI/CD Pipeline. :( |
We are running 16k tests and got the same problem as well (concurrency 2-3, chunks 200-500). |
There's a PR for this problem for this at #54 |
Is it abandoned? |
I'm submitting a ...
What is the current behavior?
When a test (shard) disconnects, that shard is restarted but it only runs 1 more test in that set of tests giving it a false SUCCESS.
More info
We had tests that had been failing on our CI but we didn't know it because the build was being reported as successful. Checking in new tests caused the shards to change revealing failures for tests that were previously being skipped
The logs below are at the time when one of the browsers disconnects and it's restarted with a single test.
This is the end of a not so successful run
One shard that seem to be absent is the following (with 490 tests), which had been logged earlier
That shard had not connected back to the karma server for over 7 minutes. Here's the log for that disconnect:
And it looks like it was replaced with a duplicate of the shard with 543 (the above logs show the duplicate entry for that shard.
We have not been able to create a minimal demo
When a shard disconnects, it either runs the entire, or restarts from the test that had a problem.
I can provide a full log of different runs if it helps. I can also try any suggestions. I understand this is not the best bug submission but it does look like there's a problem when restarting a shard after the browser disconnects.
The text was updated successfully, but these errors were encountered: