Edge node's pod unable to access outside

[chunfungintel]

What happened:
Edge node's pod unable to access outside

What you expected to happen:
Edge node's pod able to access outside

How to reproduce it (as minimally and precisely as possible):
Cloud:

sudo -E keadm init \
--advertise-address ${KEADM_THE_EXPOSED_IP} \
--profile version=v1.15.0 \
--kube-config=${KUBECONFIG} \
--set cloudCore.service.enable=true \
--set cloudCore.modules.dynamicController.enable=true

EdgeMesh:

helm install edgemesh --namespace kubeedge \
--set agent.psk=`openssl rand -base64 32` \
--set agent.relayNodes[0].nodeName=${KEADM_THE_CLOUD},agent.relayNodes[0].advertiseAddress={${KEADM_THE_EXPOSED_IP}} \
https://raw.githubusercontent.com/kubeedge/edgemesh/main/build/helm/edgemesh.tgz

Edge node join:

sudo keadm join --cloudcore-ipport="${KEADM_THE_EXPOSED_IP}":10000 \
--token=${KEADM_TOKEN} \
--kubeedge-version=v1.15.0 \
--remote-runtime-endpoint="unix:///var/run/containerd/containerd.sock" \
--runtimetype=remote \
--cgroupdriver=systemd \
--edgenode-name=${KUBEEDGE_NODE_NAME}

Please refer here for edgecore configuration
edgecore.txt

I tested with test case here: https://edgemesh.netlify.app/guide/test-case.html
And it is working correctly

$ kubectl exec -it alpine-test -- sh
(in the container environment)
/ # curl hostname-svc:12345
hostname-edge-5c75d56dc4-rq57t

However, if I doing doing outwards access from edge node's pod, I can't. Any idea why?

Anything else we need to know?:

Environment:

• EdgeMesh version: latest

• Kubernetes version (use kubectl version): 1.23.17

• KubeEdge version(e.g. cloudcore --version and edgecore --version): 1.15

• Cloud nodes Environment:

  • Hardware configuration (e.g. lscpu):
  • OS (e.g. cat /etc/os-release): Ubuntu 22.04.2 LTS
  • Kernel (e.g. uname -a): 6.1
  • Go version (e.g. go version):
  • Others:
• Edge nodes Environment:

  • edgecore version (e.g. edgecore --version):
  • Hardware configuration (e.g. lscpu):
  • OS (e.g. cat /etc/os-release): Ubuntu 22.04.2 LTS
  • Kernel (e.g. uname -a): 6.1
  • Go version (e.g. go version):
  • Others:

[chunfungintel]

I found the root cause, iptables forward is not setting correctly:
sudo iptables -P FORWARD ACCEPT

[victorming666]

@chunfungintel I have the same issue as you have. I setup a k8s cluster inside the internel ethernet. And I set a public Internet IP as the advertiseAddress which could be accessed both from the internal cloud nodes and some edge nodes on the Internet, but the p2p communications seems not work. I run the code on the advertiseAddress node to set iptables as you did, but the issue remains. Is there any other possible root cause for this?