Skip to content
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

TestFunctional/Tunnel on none: svc should have ingress after tunnel is created, but it was empty! #3873

Closed
tstromberg opened this issue Mar 13, 2019 · 4 comments · Fixed by #4105
Assignees
Labels
area/tunnel Support for the tunnel command co/none-driver kind/flake Categorizes issue or PR as related to a flaky test. priority/important-soon Must be staffed and worked on either currently, or very soon, ideally in time for the next release.
Milestone

Comments

@tstromberg
Copy link
Contributor

This is regularly occurring in our CI system:

https://storage.googleapis.com/minikube-builds/logs/3767/Linux-None.txt

    --- PASS: TestFunctional/Status (0.36s)
        cluster_status_test.go:35: Checking if cluster is healthy.
    --- FAIL: TestFunctional/Tunnel (3.97s)
        tunnel_test.go:45: starting tunnel test...
        tunnel_test.go:62: deploying nginx...
        tunnel_test.go:83: getting nginx ingress...
        tunnel_test.go:98: svc should have ingress after tunnel is created, but it was empty!
@tstromberg tstromberg added co/none-driver kind/flake Categorizes issue or PR as related to a flaky test. area/tunnel Support for the tunnel command labels Mar 13, 2019
@tstromberg tstromberg added the priority/important-soon Must be staffed and worked on either currently, or very soon, ideally in time for the next release. label Mar 13, 2019
@tstromberg tstromberg added this to the v1.0.1-candidate milestone Apr 8, 2019
@tstromberg
Copy link
Contributor Author

I did some searching through to see when this issue started becoming prevalent:

find /var/lib/jenkins/jobs/Linux_Integration_Tests_none -name log | xargs grep "svc should have ingress after tunnel is created" > /tmp/hits

Sorting this list, you can see that the problem becomes prevalent starting at build 3085 (Mar 1, built from #3774), and consistent by build 3089 (Mar 2, built from upstream:

NOTE: 3074 was also built from #3774

Linux_Integration_Tests_none/builds/1289/log
Linux_Integration_Tests_none/builds/1356/log
Linux_Integration_Tests_none/builds/1581/log
Linux_Integration_Tests_none/builds/1591/log
Linux_Integration_Tests_none/builds/1598/log
Linux_Integration_Tests_none/builds/3074/log
Linux_Integration_Tests_none/builds/3085/log
Linux_Integration_Tests_none/builds/3086/log
Linux_Integration_Tests_none/builds/3089/log
Linux_Integration_Tests_none/builds/3090/log
Linux_Integration_Tests_none/builds/3091/log
...

Based on the initial evidence, I wouldn't be surprised if #3774 (2f1f52e) had some unintentional impact here. If nothing else, it gives a good place to begin with git bisect.

@tstromberg
Copy link
Contributor Author

Any luck?

@balopat
Copy link
Contributor

balopat commented Apr 25, 2019

Should be working now, fixed in #4105

@balopat balopat closed this as completed Apr 25, 2019
@balopat
Copy link
Contributor

balopat commented Apr 25, 2019

In the end everything was working just fine with tunnel in None driver as well. We just needed more wait for kubernetes to bring up the ingress than the 2 seconds that we waited. I fixed it by setting up a better polling/error handling in the integration test.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
area/tunnel Support for the tunnel command co/none-driver kind/flake Categorizes issue or PR as related to a flaky test. priority/important-soon Must be staffed and worked on either currently, or very soon, ideally in time for the next release.
Projects
None yet
Development

Successfully merging a pull request may close this issue.

2 participants