Cypress 6.3.0 Fails to exit if connecting to Chrome falls back to TCP
See original GitHub issueCurrent behavior
Cypress 6.3.0 Running in Headless Mode, set to Chrome Browser does not EXIT after the tests are done, if the connection falls back to TCP, showing this in any part of the test.
Timed out waiting for the browser to connect. Retrying...
Warning: Cypress failed to connect to Chrome via stdio after 1 minute, 0 seconds. Falling back to TCP...
Connecting to Chrome via TCP was successful, continuing with tests
Additional Information:
-
We have to run “killall -9 chrome” for the Cypress to exit, if the tests all passed, it does return a success status when doing this. So it is likely an issue with Cypress during termination.
-
If there are retries in the test, which should have otherwise passed due to retry, the retry is not honored, and the test is reported as failed
-
This only happens in our AWS EC2 Linux Ubuntu 20.04, it DOES NOT happen in our Mac Mini environment, maybe because the Chrome TCP Fallback does not happen at all in our Mac-Mini Server.
Desired behavior
Cypress 6.3.0 Running in Headless Mode, set to Chrome Browser SHOULD EXIT after the tests are done even if the connection falls back to TCP
Test code to reproduce
This happens on any occasion that this notice appears:
Timed out waiting for the browser to connect. Retrying...
Warning: Cypress failed to connect to Chrome via stdio after 1 minute, 0 seconds. Falling back to TCP...
Connecting to Chrome via TCP was successful, continuing with tests
Versions
6.3.0
Issue Analytics
- State:
- Created 3 years ago
- Reactions:49
- Comments:8 (1 by maintainers)
It seems that this was fixed on version 6.4.0: https://github.com/cypress-io/cypress/releases/tag/v6.4.0
Tested it out in our repo and upgrading to 6.4.0 fixed the problem for us
I can confirm the same issue with GitHub Actions, using
ubuntu-18.04
. Our tests are also running in headless mode against Chrome. We also were usingnode@12.13.1
.I can’t link to the run because it is in a private repository. Hope this helps and gets resolved soon 🙂. Our internal systems unfortunately hit our usage limit very quickly because of this problem. Our normal 5 minutes runs turned into 4-6 hours runs 🙃.