I had initially planned to add more tests that cover installation of the two control planes after each other, but it turned out to be too complex and fragile with the integration test framework. So here's just a basic "who reconciles what" test that verifies that GatewayControllerMode does indeed ignore Istio resources outside its scope, but watches Kubernetes resources in the whole cluster.
While at it, I found out that the change of the config map name was not properly propagated in the charts, so I fixed that. I also had to adjust the values_types.proto because the integration test framework uses istioctl to apply config and it would fail without this change.
I had initially planned to add more tests that cover installation of the two control planes after each other, but it turned out to be too complex and fragile with the integration test framework. So here's just a basic "who reconciles what" test that verifies that GatewayControllerMode does indeed ignore Istio resources outside its scope, but watches Kubernetes resources in the whole cluster.
While at it, I found out that the change of the config map name was not properly propagated in the charts, so I fixed that. I also had to adjust the
values_types.proto
because the integration test framework uses istioctl to apply config and it would fail without this change.