Systemd does not get notified with READY state in etcd only nodes

[galal-hussein]

Environmental Info:
RKE2 Version:

rke2 version v1.21.3-rc1+rke2r1 (cd52869.dirty)
go version go1.16.3

Node(s) CPU architecture, OS, and Version:
Linux ip-172-31-14-237 5.4.0-1045-aws #47-Ubuntu SMP Tue Apr 13 07:02:25 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux

Cluster Configuration:
1 node (etcd only node)

Describe the bug:
ETCD only node gets stuck with systemd since the recent changes of changing the notify for systemd in agent logic and not in server logic

Steps To Reproduce:
Install rke2 server with disabling api, controller, and scheduler
restart rke2

Expected behavior:
systemctl restart rke2-server restarts rke2 normally

Actual behavior:
systemctl restart command gets stuck

Additional context / logs:

[Jono-SUSE-Rancher]

Just putting a note here that this issue is currently a blocker for rancher/rancher#32362 on the Rancher side.

[rancher-max]

This is fixed in v1.21.3-rc2+rke2r2. I am able to bring up individual role clusters in rancher: