okd/test/aws/scaleup.yml

---
- name: create new nodes
  hosts: localhost
  connection: local
  tasks:
  - import_tasks: ssh_bastion.yml

  - import_tasks: get_machinesets.yml

  - include_tasks: create_machineset.yml
    loop: "{{ machineset.resources }}"
    when:
    - item.status.replicas is defined
    - item.status.replicas != 0

- name: wait for nodes to become available
  hosts: new_workers
  gather_facts: false
  tasks:
  - wait_for_connection: {}
  - setup: {}
  - name: Copy ops-mirror.pem
    copy:
      src: ../../inventory/dynamic/injected/ops-mirror.pem
      dest: /var/lib/yum/ops-mirror.pem
      owner: root
      group: root
      mode: 0644
  - name: Initialize openshift repos
    import_tasks: additional_repos.yml

- import_playbook: ../../playbooks/scaleup.yml
  vars:
    openshift_kubeconfig_path: "{{ kubeconfig_path }}"

- name: wait for nodes to join
  hosts: new_workers
  tasks:
  - name: HACK disable selinux
    selinux:
      policy: targeted
      state: permissive
  - name: Create core user for storage tests to pass
    user:
      name: core
      group: wheel
  - name: Make sure core user has ssh config directory
    file:
      name: /home/core/.ssh
      state: directory
      owner: core
      group: wheel
      mode: 0700
  - name: Install nfs-utils for storage tests
    package:
      name: nfs-utils
      state: present
    register: result
    until: result is succeeded
  - name: Wait for new nodes to be ready
    k8s_facts:
      kubeconfig: "{{ kubeconfig_path }}"
      kind: Node
      name: "{{ node_name }}"
    delegate_to: localhost
    register: new_machine
    until:
    - new_machine.resources is defined
    - new_machine.resources | length > 0
    - new_machine.resources[0].status is defined
    - new_machine.resources[0].status.conditions is defined
    - new_machine.resources[0].status.conditions | selectattr('type', 'match', '^Ready$') | map(attribute='status') | join | bool == True
    # Give the node three minutes to come back online.
    retries: 48
    delay: 30
    ignore_errors: true
  - when: new_machine is failed
    block:
    - name: Collect a list of containers
      command: crictl ps -a -q
      ignore_errors: true
      register: crictl_ps_output
    - name: Collect container logs
      command: "crictl logs {{ item }}"
      register: crictl_logs_output
      with_items: "{{ crictl_ps_output.stdout_lines }}"
      ignore_errors: true
    - name: Get crio logs
      command: journalctl --no-pager -u crio
      register: crio_logs
      ignore_errors: true
    - name: Get kubelet logs
      command: journalctl --no-pager -u kubelet
      register: kubelet_logs
      ignore_errors: tru
    - debug:
        var: crictl_logs_output
    - debug:
        msg: "{{ kubelet_logs.stdout_lines }}"
    - debug:
        msg: "{{ crio_logs.stdout_lines }}"
    - fail:
        msg: Node failed to become Ready

- name: Remove CoreOS nodes
  hosts: localhost
  connection: local
  tasks:
  - name: Mark CoreOS nodes as unschedulable
    command: >
      oc adm cordon {{ item | lower }}
      --config={{ kubeconfig_path }}
    with_items: "{{ pre_scaleup_workers_name }}"

  - name: Drain CoreOS nodes
    command: >
      oc adm drain {{ item | lower }}
      --config={{ kubeconfig_path }}
      --force --delete-local-data --ignore-daemonsets
      --timeout=0s
    with_items: "{{ pre_scaleup_workers_name }}"

  - name: remove existing machinesets
    k8s:
      api_version: machine.openshift.io/v1beta1
      kubeconfig: "{{ kubeconfig_path }}"
      namespace: openshift-machine-api
      kind: MachineSet
      name: "{{ item }}"
      state: absent
    with_items: "{{ pre_scaleup_machineset_names }}"

  - name: Wait for worker configs to roll out
    command: oc wait machineconfigpool/worker --for=condition=Updated --timeout=5m