I am trying to setup kubeadm cluster with the following components on AWS EC2 O . . .

Kubernetes-Slack Discussions
#1

Bhaskar Sarma:
I am trying to setup kubeadm cluster with the following components on AWS EC2

OS : Ubuntu version : 5.4.0-1060-aws #63~18.04.1-Ubuntu
docker-ce=18.06.1~ce~3-0~ubuntu kubelet=1.15.7-00 kubeadm=1.15.7-00 kubectl=1.15.7-00

joining worker nodes are hanging on kubeadm join at preflight… Any help on how to debug it ?
[preflight] Running pre-flight checks

#2

Alistair Mackay:
If it’s hanging, then possibly network issue.
Check security groups. As initial test ensure all nodes of the cluster can reach each other on all ports (can be tightened up later).

Check also kernel modules. May need to enable overlay and br_netfilter
Also sysctls. May need

net.bridge.bridge-nf-call-iptables
net.bridge.bridge-nf-call-ipv6tables
net.ipv4.ip_forward

You may also need to disable source/dest check on the EC2 instances

#3

Bhaskar Sarma:
Perfect … looks to be SG issue. I updated SG inbound rules to allow all traffic with in SG. Assigned the same SG to both master, worker1, worker2 nodes

#4

Bhaskar Sarma:
this resolved the issue

#5

Bhaskar Sarma:
Thank you @Alistair Mackay you are on the dot

#6

Alistair Mackay:
You should be able to tighten up your SGs from this list. Remember to also allow any ports on workers (e.g. for ingress) that applications running in the cluster will need.
https://kubernetes.io/docs/reference/ports-and-protocols/