Cannot access without adding clusterid

[trumpzhang]

What happened:
I set up three clusters: broker, c1, and c2, all of them has only master node. After that, i use subctl join --kubeconfig ./config/c2.yaml broker-info.subm --clusterid c2 --natt=false make c2 a member .

In c1, I use subctl export service --namespace sample whereami-cs exported the service whereami-cs which in the namespace sample. In c2, I couldn't access it use curl whereami-cs.sample.svc.clusterset.local, but when i use curl c1.whereami-cs.sample.svc.clusterset.local, it worked.

Client pod info:
kubectl run client --image=cr7258/nettool:v1
all test curl command exec in pod client.

info:
use c1.whereami-cs.sample.svc.clusterset.local

use whereami-cs.sample.svc.clusterset.local

test-submariner.yaml
`
apiVersion: apps/v1
kind: Deployment
metadata:
name: whereami
namespace: sample
spec:
replicas: 1
selector:
matchLabels:
app: whereami
template:
metadata:
labels:
app: whereami
spec:
containers:
- name: whereami
image: cr7258/whereami:v1
imagePullPolicy: IfNotPresent
ports:
- containerPort: 80
env:
- name: NAMESPACE
valueFrom:
fieldRef:
fieldPath: metadata.namespace
- name: NODE_NAME
valueFrom:
fieldRef:
fieldPath: spec.nodeName
- name: POD_NAME
valueFrom:
fieldRef:
fieldPath: metadata.name
- name: POD_IP
valueFrom:
fieldRef:
fieldPath: status.podIP

apiVersion: v1
kind: Service
metadata:
name: whereami-cs
namespace: sample
spec:
selector:
app: whereami
ports:

`

What you expected to happen:
use curl whereami-cs.sample.svc.clusterset.local successful.

How to reproduce it (as minimally and precisely as possible):
Install broker first(will as broker role)
subctl join --kubeconfig ./config/c1.yaml broker-info.subm --clusterid c1 --natt=false
subctl join --kubeconfig ./config/c2.yaml broker-info.subm --clusterid c2 --natt=false

Anything else we need to know?:
what should i do to access "whereami-cs.sample.svc.clusterset.local" successfully in cluster c2.
when use nslookup:

complement:
i change test client from cr7258/nettool:v1 to alpine3.19 , use nslookup successful, but curl failed.
like follow:

I'm new to k8s and submariner, need some help.

Environment:

• k3s: v1.27.2+k3s1
• subctl: v0.13.0
• docker: 20.10.9

[yboaron]

Thanks for trying Submariner @trumpzhang ,

Please share the output of [1] from both clusters

[1]
`subctl show all

subctl diagnose all
`

[trumpzhang]

Thanks for trying Submariner @trumpzhang ,

Please share the output of [1] from both clusters

[1] `subctl show all

subctl diagnose all `

thanks your reply. Query result like below:

• broker

subctl show all

subctl diagnose all

• c1

subctl --kubeconfig ./config/c1.yaml show all

subctl --kubeconfig ./config/c1.yaml diagnose all

• c2

subctl --kubeconfig ./config/c2.yaml show all

subctl --kubeconfig ./config/c2.yaml diagnose all

[yboaron]

First of all, seems that you are using an older version (0.13) of Submariner, please reinstall Submariner using latest subctl (0.18), check [1] for more details.

Post deployment, you can use subctl automatic verification for testing inter-cluster service discovery functionality.

[1]
https://submariner.io/operations/deployment/subctl/#installation

[trumpzhang]

@yboaron
Thanks for your reply. I find the problem and fix it up.
A few days ago, I installed app named "Clash Verge" which is a proxy software. For convenience, I turned on its Tun mode, which caused requests from pods in my virtual machine to not be resolved correctly. Because I found that the curl public domain name in the pod was unresponsive, but the curl public domain name on the host was good. When I turned off Tun mode, everything worked normally.

[dfarrell07]

@trumpzhang Glad things are working. Closing this for now, feel free to open something else if you have more questions.