-
Notifications
You must be signed in to change notification settings - Fork 151
ci: require more interop tests to pass #1919
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
Conversation
… HandshakeLoss/Corruption tests succeeding
| - "msquic": { | ||
| - "image": "ghcr.io/microsoft/msquic/qns:main", | ||
| - "url": "https://github.com/microsoft/msquic", | ||
| - "role": "both" | ||
| - }, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'm assuming this was intentional since they're always failing here. Would it be worth removing chrome and xquic as well? They haven't been working for some time.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yeah, I removed msquic because it was always failing but also it was taking a really long time to run (it was the longest of all the impls). chrome and xquic at least run relatively quickly. chrome was succeeding a while back, I'm not sure what happened. xquic does succeed on some tests. I think I would just leave those for now
Description of changes:
This change enforces that many more interop tests for CI to pass. This includes requiring tests for the s2n-quic client and s2n-quic-rustls to pass, which previously was not required for any testcase.
I fixed the
check.pyscript to correctly determine success for the s2n-quic client.Call-outs:
The ECN tests has not shown stability for the previous 50 runs (which is what the rest of the required checks are based on), but recent changes should hopefully resolve their instability
HandshakeLoss and HandshakeCorruption still don't seem stable enough to enable enforcement.Increasing timeouts to 5 minutes and disabling MTU probing to try to stabilize these
Rebind Address also has a timing issue that sometimes causes failures as well, so I haven't enabled that either.The rebind address test starts rebinding after 1 second, which in some cases is before the handshake has completed. Since connection migration is not permitted before the handshake has completed, this can cause the test to fail. To address this I have updated the quic-interop-runner patch to start the rebinding after 2 seconds.
Testing:
CI
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.