-
Notifications
You must be signed in to change notification settings - Fork 1.5k
Closed
Description
Describe the bug
I'm trying to reuse a NLB for multiple backends with the same targetPort
the controller prints this error
{"level":"info","ts":"2024-10-28T14:34:35Z","msg":"the pod in endpoint is not found in pod cache yet, will keep retrying","podKey":"lb4/pod1-6545f5f746-tccvn"}
{"level":"info","ts":"2024-10-28T14:34:35Z","msg":"Requeue for potentially ready endpoints","tgb":{"name":"k8s-lb4-mux-b239dd8c8d","namespace":"lb4"}}
{"level":"info","ts":"2024-10-28T14:34:37Z","msg":"the pod in endpoint is not found in pod cache yet, will keep retrying","podKey":"lb4/pod2-6c5c748d49-m25sb"}
{"level":"info","ts":"2024-10-28T14:34:37Z","msg":"Requeue for potentially ready endpoints","tgb":{"name":"k8s-lb4-mux-89437c4588","namespace":"lb4"}}
Steps to reproduce
- create a service with:
- no selector
- 2 ports with different name and no target port
- create Endpoints with 2 subsets pointing to 2 different pods in a different namespace from the service
apiVersion: v1
kind: Service
metadata:
annotations:
service.beta.kubernetes.io/aws-load-balancer-nlb-target-type: ip
service.beta.kubernetes.io/aws-load-balancer-scheme: internet-facing
service.beta.kubernetes.io/aws-load-balancer-type: nlb
name: mux
namespace: multiplexer
spec:
loadBalancerClass: service.k8s.aws/nlb
ports:
- name: 1dcf5c8
nodePort: 31718
port: 31040
- name: d90d676
nodePort: 32450
port: 31815
type: LoadBalancer
---
apiVersion: v1
kind: Endpoints
metadata:
name: mux
namespace: multiplexer
subsets:
- addresses:
- ip: 10.0.6.192
nodeName: ip-10-0-5-150.us-west-2.compute.internal
targetRef:
kind: Pod
name: pod1-6545f5f746-tccvn
namespace: default
uid: c273178d-3c6e-45b4-b410-3927d137e780
ports:
- name: d90d676
port: 80
protocol: TCP
- addresses:
- ip: 10.0.6.121
nodeName: ip-10-0-5-150.us-west-2.compute.internal
targetRef:
kind: Pod
name: pod2-6c5c748d49-m25sb
namespace: default
uid: 5ba62e6f-4b3f-4520-94bb-8dca662483c7
ports:
- name: 1dcf5c8
port: 80
protocol: TCP
Expected outcome
It's expected that the controller will create
- an NLB with 2 listeners:
31040
and31815
- 2 target groups
- 1 target in each target group pointing to pod1 and pod2
Environment
- AWS Load Balancer controller version: v2.9.2
- Kubernetes version: 1.29.8
- Using EKS (yes), version: 1.29
Metadata
Metadata
Assignees
Labels
No labels