Closed lazybetrayer closed 2 years ago
@MarkKharitonov did you follow these steps already? if so, and they resolve the issue, it would be great if you could PR some docs or at least open a docs issue
@afirth - Are these steps relevant if the ingresses live in different namespaces, as is my case? It was my understanding from reading that comment that having ingresses in different namespaces eliminates all the issues naturally.
Have I missed anything?
This is false alarm. I have found the root cause. It is actually educational.
We copied the nginx images to our internal image repo in Azure and never refreshed it while we did refresh the HELM chart deployed.
This is a lesson for us to improve this rather broken process.
NGINX Ingress controller version: v1.0.0
Kubernetes version (use
kubectl version
): v1.20.9Environment: Bare Metal
What happened:
Before v1.0.0, there's a check to skip validating ingress with a different ingressclass: https://github.com/kubernetes/ingress-nginx/blob/f3c50698d98299b1a61f83cb6c4bb7de0b71fb4b/internal/ingress/controller/controller.go#L224
In v1.0.0, this code is removed. With multiple ingress controllers, both will validate the same ingress. If we create two ingresses with same host and path but different ingressclasses, the second one will be rejected.
What you expected to happen:
skip validating ingress with a different ingressclass
/kind bug