Спасибо всем!
нашел в чем причина была.
Terraform (OpenTofu) создавал hosts.yaml по шаблону
all:
hosts:%{ for idx, master in masters }
master:
ansible_host: ${master.network_interface[0].nat_ip_address}
ip: ${master.network_interface[0].ip_address}
access_ip: ${master.network_interface[0].nat_ip_address}%{ endfor }%{ for idx, worker in workers }
worker-${idx + 1}:
ansible_host: ${worker.network_interface[0].nat_ip_address}
ip: ${worker.network_interface[0].ip_address}
access_ip: ${worker.network_interface[0].nat_ip_address}%{ endfor }
children:
kube_control_plane:
hosts:%{ for idx, master in masters }
${master.name}:%{ endfor }
kube_node:
hosts:%{ for idx, worker in workers }
${worker.name}:%{ endfor }
etcd:
hosts:%{ for idx, master in masters }
${master.name}:%{ endfor }
k8s_cluster:
children:
kube_control_plane:
kube_node:
calico_rr:
hosts: {}
убрал строчки
access_ip: ${master.network_interface[0].nat_ip_address}
access_ip: ${worker.network_interface[0].nat_ip_address}
и все без ошибок поднялось