ansible-role-k3s/tasks/build/configure-k3s-cluster.yml
Xan Manning f684f6d907 A retry has been added to task controlling secondary master startup.
Fixes #17

There appeared to be a race condition where starting all secondary
masters all at once would cause the k3s service to fail on a number of
the other masters. A retry has been added to the task to attempt to
bring them all up until they stop failing.
2020-03-07 16:15:41 +00:00

82 lines
2.5 KiB
YAML

---
- name: Ensure ansible_host is mapped to inventory_hostname
lineinfile:
path: /tmp/inventory.txt
line: >-
{{ item }}
@@@
{{ hostvars[item].ansible_host | default(hostvars[item].ansible_fqdn) }}
@@@
C_{{ hostvars[item].k3s_control_node }}
@@@
P_{{ hostvars[item].k3s_primary_control_node | default(False) }}
create: true
loop: "{{ play_hosts }}"
- name: Lookup control node from file
command: "grep '{{ 'P_True' if (k3s_controller_count | length > 1) else 'C_True' }}' /tmp/inventory.txt"
changed_when: false
register: k3s_control_delegate_raw
- name: Ensure control node is delegated to for obtaining a token
set_fact:
k3s_control_delegate: "{{ k3s_control_delegate_raw.stdout.split(' @@@ ')[0] }}"
- name: Ensure the control node address is registered in Ansible
set_fact:
k3s_control_node_address: "{{ hostvars[k3s_control_delegate].ansible_host | default(hostvars[k3s_control_delegate].ansible_fqdn) }}"
when: k3s_control_node_address is not defined
- name: Ensure NODE_TOKEN is captured from control node
slurp:
path: "/var/lib/rancher/k3s/server/node-token"
register: k3s_slurped_control_token
delegate_to: "{{ k3s_control_delegate }}"
when: k3s_control_token is not defined
- name: Ensure NODE_TOKEN is formatted correctly for use in templates
set_fact:
k3s_control_token: "{{ k3s_slurped_control_token.content | b64decode }}"
when: k3s_control_token is not defined
- name: Ensure k3s service unit file is present
template:
src: k3s.service.j2
dest: /etc/systemd/system/k3s.service
notify:
- reload systemd
- restart k3s
- name: Ensure secondary masters are started
service:
name: k3s
state: started
enabled: true
register: ensure_secondary_masters_started
until: ensure_secondary_masters_started is succeeded
retries: "{{ play_hosts | length }}"
delay: 2
when: k3s_control_node and not k3s_primary_control_node
- name: Wait for control plane to be ready to accept connections
wait_for:
port: "{{ k3s_https_port }}"
delay: 5
sleep: 5
timeout: 300
when: k3s_control_node
- meta: flush_handlers
- name: Wait for all nodes to be ready
command: "{{ k3s_install_dir }}/kubectl get nodes"
changed_when: false
failed_when: false
register: kubectl_get_nodes_result
until: kubectl_get_nodes_result.rc == 0
and kubectl_get_nodes_result.stdout.find("NotReady") == -1
retries: 30
delay: 20
when: k3s_control_node and not k3s_no_flannel