ceph-cm-ansible/tools/prep-fog-capture.yml at 9fa30e919a808274c475888ca6c6533cb0a4ec7f · ceph/ceph-cm-ansible · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
---
### This standalone playbook can be used to prep a COBBLER-IMAGED testnode
### so that it can be used to capture an OS image for FOG.
### This playbook is needed for a couple reasons
###   - NIC configs get hard coded into the captured FOG images so nodes reimaged by FOG don't come up with network

- hosts:
    - testnodes
  become: true
  gather_facts: false
  tasks:

  # (Missing in RHEL8)
  - name: Check for /usr/bin/python
    shell: echo marco
    register: polo
    ignore_errors: true

  - name: Set ansible_python_interpreter=/usr/bin/python3
    set_fact:
      ansible_python_interpreter: /usr/bin/python3
    when: polo is failed

  # Now that we know where python is, we can gather_facts
  - setup:

  # We need to leave /.cephlab_rc_local or else each FOG reimage would tell Cobbler to run ceph-cm-ansible
  - name: Remove lock files and udev rules
    file:
      path: "{{ item }}"
      state: absent
    with_items:
      - /etc/udev/rules.d/70-persistent-net.rules
      - /.cephlab_net_configured
      - /ceph-qa-ready

  - name: Get list of ifcfg scripts from host used to capture image
    shell: "ls -1 /etc/sysconfig/network-scripts/ifcfg-* | grep -v ifcfg-lo"
    register: ifcfg_scripts
    when: ansible_os_family == "RedHat"
    ignore_errors: true

  - name: Get list of ifcfg scripts from host used to capture image
    shell: "ls -1 /etc/sysconfig/network/ifcfg-* | grep -v ifcfg-lo"
    register: ifcfg_scripts
    when: ansible_os_family == "Suse"
    ignore_errors: true

  - name: Delete ifcfg scripts
    file:
      path: "{{ item }}"
      state: absent
    with_items: "{{ ifcfg_scripts.stdout_lines|default([]) }}"
    when: ifcfg_scripts is defined

  - name: Remove /var/lib/ceph mountpoint from fstab
    shell: sed -i '/\/var\/lib\/ceph/d' /etc/fstab

  - name: Unmount /var/lib/ceph
    ansible.posix.mount:
      path: /var/lib/ceph
      state: unmounted

  - name: Get list of SSH host keys
    shell: "ls -1 /etc/ssh/ssh_host_*"
    register: ssh_host_keys
    ignore_errors: true

  # Key regeneration is done automatically on CentOS firstboot.
  # For Ubuntu, we'll add `dpkg-reconfigure openssh-server` to rc.local
  - name: Delete SSH host keys so they're generated during firstboot on cloned machines
    file:
      path: "{{ item }}"
      state: absent
    with_items: "{{ ssh_host_keys.stdout_lines|default([]) }}"
    when: ssh_host_keys is defined

  - name: Unsubscribe RHEL
    command: subscription-manager unregister
    when: ansible_distribution == "RedHat"
    failed_when: false

  # A file gets leftover when a testnode is registered with Satellite that caused
  # each registered subsequent testnode to report the wrong hostname
  - name: Clean up katello facts
    file:
      path: /etc/rhsm/facts/katello.facts
      state: absent
    when: ansible_distribution == "RedHat"

  # https://bugzilla.redhat.com/show_bug.cgi?id=1814337
  - name: Disable dnf-makecache service
    service:
      name: dnf-makecache.timer
      state: stopped
      enabled: no
    when:
      - ansible_os_family == "RedHat"
      - ansible_distribution_major_version|int >= 8

  # Hopefully fixes https://github.com/ceph/ceph-cm-ansible/pull/544#issuecomment-599076564
  - name: Clean DNF cache
    shell: "dnf clean all && rm -rf /var/cache/dnf/*"
    when:
      - ansible_os_family == "RedHat"
      - ansible_distribution_major_version|int >= 8

  - set_fact:
      ntp_service: ntp
    when: ansible_os_family == "Debian"

  - set_fact:
      ntp_service: ntpd
    when: ansible_os_family == "RedHat" and ansible_distribution_major_version|int <= 7

  - set_fact:
      ntp_service: chronyd
    when: (ansible_os_family == "RedHat" and ansible_distribution_major_version|int >= 8) or
          ansible_os_family == "Suse"

  - name: "Stop {{ ntp_service }} service"
    service:
      name: "{{ ntp_service }}"
      state: stopped
    when: '"ntp" in ntp_service'

  # The theory here is although we do have the ntp service running on boot,
  # if the time is off, it slowly drifts back in sync.  Since our testnodes
  # are ephemeral, they don't ever have enough time to correctly drift
  # back to the correct time.  So we'll force it in the captured OS images.
  - name: Force time synchronization using stepping | ntp
    command: "ntpdate -b {{ ntp_servers|join(' ') }}"
    when: '"ntp" in ntp_service'

  - name: "Start {{ ntp_service }}"
    service:
      name: "{{ ntp_service }}"
      state: started

  # chronyd needs to be started in order to force time sync. This differs from ntpd.
  - name: Force time synchronization using stepping | chrony
    command: chronyc -a makestep
    when: '"chrony" in ntp_service'

  - name: Sync the hardware clock
    command: "hwclock --systohc --verbose"
    become: true
    become_method: sudo
    become_user: root
    environment:
      PATH: "/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin"
    ignore_errors: true