Corner Case: upstream name duplication causing ingress pointing to wrong service

[Revolution1]

Version: v1.10.1 (but this bug exists in the latest main branch)

Let's say we have 2 ingress-paths:

1. points to backend: service.name: example, service.port.name: service-rpc
2. points to backend: service.name: example-service, serivce.port.name: rpc

In the ingress controller, they will be converted to upstream name by func upstreamName

ingress-nginx/internal/ingress/controller/util.go

Lines 49 to 60 in a647bc1

	// upstreamName returns a formatted upstream name based on namespace, service, and port
	func upstreamName(namespace string, service *networking.IngressServiceBackend) string {
	if service != nil {
	if service.Port.Number > 0 {
	return fmt.Sprintf("%s-%s-%d", namespace, service.Name, service.Port.Number)
	}
	if service.Port.Name != "" {
	return fmt.Sprintf("%s-%s-%s", namespace, service.Name, service.Port.Name)
	}
	}
	return fmt.Sprintf("%s-INVALID", namespace)
	}

The output for path 1 and 2 with be the same example-service-rpc

Which makes one of the path points to wrong backend.

[k8s-ci-robot]

This issue is currently awaiting triage.

If Ingress contributors determines this is a relevant issue, they will accept it by applying the triage/accepted label and provide further guidance.

The triage/accepted label can be added by org members by writing /triage accepted in a comment.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

[longwuyuan]

/remove-kind bug

• A short text description is not a bug so removing bug label
• See the template of a new bug report and edit your issue description to answer the questions that are asked in a new bug report, in markdown format
• Show a step-by-step detailed guide to reproduce the problem
  • Create a kind cluster
  • Install ingress-nginx controller in it
  • Create a deployment using image nginx:alpine
  • Create a deployment using image httpd:alpine
  • Create the services and ingresses
  • Show the curl command outputs
  • show the kubectl describe of all related resources
  • show the logs
  • show all data and explain the bug in this practical environment

/kind support
/triage needs-information

[Revolution1]

@longwuyuan Okay, will change.
Btw, Is this report not clear enough or you're actually a bot that can only read text scrictly obey the template?

[Revolution1]

the template also got some bugs in it

[Revolution1]

link to: #11938

[longwuyuan]

/close

Duplicate of #11938

[k8s-ci-robot]

@longwuyuan: Closing this issue.

In response to this:

/close

Duplicate of #11938

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.