Adding worker nodes instructions

[hbjcr]

Expected Behavior

Section with instructions explaining how to add new worker nodes

User should be able to follow the instructions to add a new node without having to rebuild the cluster from scratch

Current Behavior

No instructions are currently provided

Steps to Reproduce

NTR

Context (variables)

Operating system:

Hardware: x64, arm64, armhf

Variables Used

Hosts

Possible Solution

• I've checked the General Troubleshooting Guide

[billsioros]

Could anyone who has already come across this issue please provide some instructions?

[emcniece]

#360 (comment)

[fellnase]

This would be a nice feature, if I only have to add a new server or worker in the host.ini file and run the script again.
I also had the problem, that a node died and I have to setup it again. Because of the "special" configuration it is not so easy to know, how to do this manually.

[pbolduc]

If you wanted to do it manually, you could check out this rpi4cluster. The author still uses ansible, but not this playbook.

[ruslanguns]

@pbolduc the url provided is not working, could you check it please?

[pbolduc]

It is not my site. Seems the owner reorganized it. Just remove the path from the URL.

[emcniece]

This worked for me, Ubuntu 22.04:

0. Clone existing node
1. Edit /etc/netplan/50-cloud-init.yaml
   • Update IP address
   • Uptake new address: netplan try
2. Test host google.ca - if there are DNS errors, reboot
3. Edit K8s node annotations:
   • k3s.io/internal-ip: [new ip]
   • k3s.io/node-args: ... --node-ip=[new ip] flag
4. Edit /etc/systemd/system/k3s-node.service
   • Update --node-ip=[new ip] flag
   • systemctl daemon-reload
   • systemctl restart k3s-node

Step 3 may not be needed - I suspect these annotations might be generated during k3s-node service startup. Regardless it is useful to check those annotations after step 4 is complete.