[release-1.21] Cluster fail to provision when using AWS cloud provider

[galal-hussein]

Environmental Info:
RKE2 Version: v1.21.3+rke2r1

Node(s) CPU architecture, OS, and Version:
Ubuntu 20.04

Cluster Configuration:
1 Server

Describe the bug:
When not passing --node-name to rke2 server and using --cloud-provider-name=aws , kubelet fail to register itself with kubernetes api because of the NodeRestriction Admission plugin rancher/rancher#34105 (comment)

Steps To Reproduce:

• Create ec2 instance on AWS
• Tag instance, subnet, vpc, and security group with kubernetes tag
• Start rke2 server with option --cloud-provider-name=aws

Expected behavior:
The cluster should start normally

Actual behavior:
Kubelet fails to register with the following error:

Attempting to register node ip-172-28-68-60.eu-west-1.compute.internal
Unable to register node "ip-172-28-68-60.eu-west-1.compute.internal" with API server: nodes "ip-172-28-68-60.eu-west-1.compute.internal" is forbidden: node "ip-172-28-68-60" cannot modify node "ip-172-28-68-60.eu-west-1.compute.internal"

[cjellick]

I think we want the 1.21.3+rke2r2 milestone

[brandond]

@cjellick I thought we were not actually planning on doing a 1.21.3+rke2r2 release - the current RCs are just for QA. The next actual release will be 1.21.4+rke2r1.

[galal-hussein]

@cjellick I thought we were not actually planning on doing a 1.21.3+rke2r2 release - the current RCs are just for QA. The next actual release will be 1.21.4+rke2r1.

Thats what I thought too

[brandond]

This has been merged to master; I have updated the milestone to 1.22 and created a backport issue for 1.21.

CORRECTION: this was merged to release-1.21 first, will convert the other issue into a forwardport

[rancher-max]

Validated using v1.21.3-rc6+rke2r2

No longer need to supply node-name as hostname when using aws cloud provider. Can still supply it if desired. It will automatically use an equivalent to hostname -f when starting rke2 with aws cloud provider now though.