...
Kubernetes Addons - weave second/slaves pods are not starting up after executing kubeadm join command in the slave machines. weave pods status is CrashLoopBackOff
Cause
Enhanced network feature to use ethernet network interface (eth0) is not configured. "ifconfig -a" shown the result with "ens5" for the private ip network instead of eth0.weave second/slaves pods are not starting up after executing kubeadm join command in the slave machines.network interface issue (OR) port conflicts in slave machines. Please check the weave pod log to check the port bind issue. In this case, 6783 port conflicts between slave machine kube weave and gitlab alertmanager.
Fix
Provision the aws instances with "High" network interface configuration (eth0 should be shown in In Aws console, specific instance configuration network interface should be eth0). Kubernetes cni (weave) network will use the ethernet to start the cluster(Result: slave node status will be "Ready").
(OR)
stopped gitlab. Then, did "kubeadm join" command in slave machine.