I have created an HA kubernetes cluster with kubeadm version 1.11.2 and installed calico CNI plugin which is up and running. I am trying to create a deployment with a docker image. It successfully created the deployment and created a container on the node but the container is failing to communicate to the outside world other than the node on which it's hosted (with the IP).
I have logged into the container and tried to ping the masters and other nodes it's failing.
Can anyone help me in resolving this issue?
Hard to tell, but has to be an issue with Calico/CNI. Are all your Calico pods ready on all your nodes, like this:
$ kubectl get pods -n kube-system
NAME READY STATUS RESTARTS AGE
calico-node-xxxxx 2/2 Running 0 15hYou can check the CNI configs under /etc/cni/net.d Maybe your install-cni.sh container in your Calico pod didn't initialize the configs? For example:
{
"name": "k8s-pod-network",
"cniVersion": "0.3.0",
"plugins": [
{
"type": "calico",
"log_level": "info",
"datastore_type": "kubernetes",
"nodename": "<node-name>",
"mtu": 1500,
"ipam": {
"type": "host-local",
"subnet": "usePodCidr"
},
"policy": {
"type": "k8s"
},
"kubernetes": {
"kubeconfig": "/etc/cni/net.d/calico-kubeconfig"
}
},
{
"type": "portmap",
"snat": true,
"capabilities": {"portMappings": true}
}
]
}
Normally, your container will have an interface and podCidr IP assigned to it, so after you shell into the pod/containers you can check with $ ifconfig