Calico Pod CrashLoopBackOff on Worker Nodes

[Muhammad Faizan]

Hi,
I am following this guide to deploy a kubernetes cluster on CoreOS with 1 master and 3 worker nodes. I have 2 issues:

1. I wanted to assign host names to my k8s Nodes instead on IPs as name so I assigned the --hostname-override to string names like "k8s-master-1", "k8s-worker1" etc.
   Everything worked fine but when I deployed the calico.yaml, the calico-node containers showed the error of not resolve adress for https://k8s-master-1:10250/containerLogs/kube-system/calico-node-jsfj6/calico-node , because it started to use the hostname instead of IP. What could be the solution for this, I mean can I specify it to use IP instead of hostname or does it only uses the hostname?

2. The calico-node containers on worker nodes were in CrashLoopBackOff state so I thought that this is the issue and I changed back the --hostname-override to IP address, and restarted. But the calico-node containers on worker nodes are stll in CrashLoopBackOff state. The calico-node container on master node is working fine.
   The logs of calico-node containers on worker nodes:

$ kubectl logs calico-node-bvzf7 -n kube-system calico-node
Waiting for etcd connection...
No IP provided. Using detected IP: 10.0.2.15
ERROR: IP address 10.0.2.15 is already in use by host w1k8s. Calico requires each compute host to have a unique IP. If this is your first time running the Calico node on this host, ensure that another host is not already using the same IP address.
Calico node failed to start

$ kubectl logs calico-node-fxxsz -n kube-system calico-node
Waiting for etcd connection...
No IP provided. Using detected IP: 10.0.2.15
ERROR: IP address 10.0.2.15 is already in use by host w1k8s. Calico requires each compute host to have a unique IP. If this is your first time running the Calico node on this host, ensure that another host is not already using the same IP address.
Calico node failed to start

Kindly tell me what could be the issue or how can I debug it to find the root cause?