ceph-ansible/infrastructure-playbooks/cephadm-adopt.yml

1546 lines
59 KiB
YAML

---
#
# This playbook does a cephadm adopt for all the Ceph services
#
- name: confirm whether user really meant to adopt the cluster by cephadm
hosts: localhost
connection: local
become: false
gather_facts: false
vars_prompt:
- name: ireallymeanit
prompt: Are you sure you want to adopt the cluster by cephadm ?
default: 'no'
private: no
tasks:
- name: exit playbook, if user did not mean to adopt the cluster by cephadm
fail:
msg: >
Exiting cephadm-adopt playbook, cluster was NOT adopted.
To adopt the cluster, either say 'yes' on the prompt or
use `-e ireallymeanit=yes` on the command line when
invoking the playbook
when: ireallymeanit != 'yes'
- name: import_role ceph-defaults
import_role:
name: ceph-defaults
- name: check if a legacy grafana-server group exists
import_role:
name: ceph-facts
tasks_from: convert_grafana_server_group_name.yml
when: groups.get((grafana_server_group_name | default('grafana-server')), []) | length > 0
- name: gather facts and prepare system for cephadm
hosts:
- "{{ mon_group_name|default('mons') }}"
- "{{ osd_group_name|default('osds') }}"
- "{{ mds_group_name|default('mdss') }}"
- "{{ rgw_group_name|default('rgws') }}"
- "{{ mgr_group_name|default('mgrs') }}"
- "{{ rbdmirror_group_name|default('rbdmirrors') }}"
- "{{ nfs_group_name|default('nfss') }}"
- "{{ iscsi_gw_group_name|default('iscsigws') }}"
- "{{ monitoring_group_name|default('monitoring') }}"
become: true
any_errors_fatal: True
gather_facts: false
vars:
delegate_facts_host: true
tasks:
- import_role:
name: ceph-defaults
- name: gather facts
setup:
gather_subset:
- 'all'
- '!facter'
- '!ohai'
when: not delegate_facts_host | bool or inventory_hostname in groups.get(client_group_name, [])
- name: gather and delegate facts
setup:
gather_subset:
- 'all'
- '!facter'
- '!ohai'
delegate_to: "{{ item }}"
delegate_facts: true
with_items: "{{ groups['all'] | difference(groups.get('clients', [])) }}"
run_once: true
when: delegate_facts_host | bool
- name: fail if one osd node is using filestore
fail:
msg: >
filestore OSDs are not supported with cephadm.
Please convert them with the filestore-to-bluestore.yml playbook first.
when:
- osd_group_name in group_names
- osd_objectstore == 'filestore'
- import_role:
name: ceph-facts
tasks_from: container_binary.yml
- name: set_fact ceph_cmd
set_fact:
ceph_cmd: "{{ container_binary + ' run --rm --net=host -v /etc/ceph:/etc/ceph:z -v /var/lib/ceph:/var/lib/ceph:ro -v /var/run/ceph:/var/run/ceph:z --entrypoint=ceph ' + ceph_docker_registry + '/' + ceph_docker_image + ':' + ceph_docker_image_tag if containerized_deployment | bool else 'ceph' }} --cluster {{ cluster }}"
- name: check pools have an application enabled
command: "{{ ceph_cmd }} health detail --format json"
register: health_detail
run_once: true
changed_when: false
delegate_to: "{{ groups[mon_group_name][0] }}"
- name: check for POOL_APP_NOT_ENABLED warning
fail:
msg: "Make sure all your pool have an application enabled."
run_once: true
delegate_to: localhost
when:
- (health_detail.stdout | default('{}', True) | from_json)['status'] == "HEALTH_WARN"
- "'POOL_APP_NOT_ENABLED' in (health_detail.stdout | default('{}', True) | from_json)['checks']"
- import_role:
name: ceph-facts
tasks_from: convert_grafana_server_group_name.yml
when: groups.get((grafana_server_group_name|default('grafana-server')), []) | length > 0
- name: get the ceph version
command: "{{ container_binary + ' run --rm --entrypoint=ceph ' + ceph_docker_registry + '/' + ceph_docker_image + ':' + ceph_docker_image_tag if containerized_deployment | bool else 'ceph' }} --version"
changed_when: false
register: ceph_version_out
- name: set_fact ceph_version
set_fact:
ceph_version: "{{ ceph_version_out.stdout.split(' ')[2] }}"
- name: fail on pre octopus ceph releases
fail:
msg: >
Your Ceph version {{ ceph_version }} is not supported for this operation.
Please upgrade your cluster with the rolling_update.yml playbook first.
when: ceph_version is version('15.2', '<')
- name: check if it is atomic host
stat:
path: /run/ostree-booted
register: stat_ostree
- name: set_fact is_atomic
set_fact:
is_atomic: "{{ stat_ostree.stat.exists }}"
- import_role:
name: ceph-container-engine
when: not containerized_deployment | bool
- import_role:
name: ceph-container-common
tasks_from: registry.yml
when:
- not containerized_deployment | bool
- ceph_docker_registry_auth | bool
- name: "pulling {{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }} image"
command: "{{ timeout_command }} {{ container_binary }} pull {{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}"
changed_when: false
register: docker_image
until: docker_image.rc == 0
retries: "{{ docker_pull_retry }}"
delay: 10
when:
- not containerized_deployment | bool
- inventory_hostname in groups.get(mon_group_name, []) or
inventory_hostname in groups.get(osd_group_name, []) or
inventory_hostname in groups.get(mds_group_name, []) or
inventory_hostname in groups.get(rgw_group_name, []) or
inventory_hostname in groups.get(mgr_group_name, []) or
inventory_hostname in groups.get(rbdmirror_group_name, []) or
inventory_hostname in groups.get(iscsi_gw_group_name, []) or
inventory_hostname in groups.get(nfs_group_name, [])
- name: configure repository for installing cephadm
when: containerized_deployment | bool
tags: with_pkg
block:
- name: set_fact ceph_origin
set_fact:
ceph_origin: repository
when: ceph_origin == 'dummy'
- name: set_fact ceph_repository
set_fact:
ceph_repository: community
when: ceph_repository == 'dummy'
- name: validate repository variables
import_role:
name: ceph-validate
tasks_from: check_repository.yml
- name: configure repository
import_role:
name: ceph-common
tasks_from: "configure_repository.yml"
- name: install cephadm requirements
tags: with_pkg
package:
name: ['python3', 'lvm2']
register: result
until: result is succeeded
- name: install cephadm
tags: with_pkg
package:
name: cephadm
register: result
until: result is succeeded
- name: install cephadm mgr module
tags: with_pkg
package:
name: ceph-mgr-cephadm
register: result
until: result is succeeded
when:
- not containerized_deployment | bool
- mgr_group_name in group_names
- name: get current fsid
command: "{{ ceph_cmd }} fsid"
register: current_fsid
run_once: true
changed_when: false
delegate_to: "{{ groups[mon_group_name][0] }}"
- name: get a minimal ceph configuration
command: "{{ ceph_cmd }} config generate-minimal-conf"
register: minimal_config
run_once: true
changed_when: false
delegate_to: "{{ groups[mon_group_name][0] }}"
- name: set_fact fsid
set_fact:
fsid: "{{ current_fsid.stdout }}"
run_once: true
- name: enable cephadm mgr module
ceph_mgr_module:
name: cephadm
cluster: "{{ cluster }}"
state: enable
environment:
CEPH_CONTAINER_IMAGE: "{{ ceph_docker_registry + '/' + ceph_docker_image + ':' + ceph_docker_image_tag if containerized_deployment | bool else None }}"
CEPH_CONTAINER_BINARY: "{{ container_binary }}"
run_once: true
delegate_to: '{{ groups[mon_group_name][0] }}'
- name: set cephadm as orchestrator backend
command: "{{ ceph_cmd }} orch set backend cephadm"
changed_when: false
run_once: true
delegate_to: '{{ groups[mon_group_name][0] }}'
- name: check if there is an existing ssh keypair
stat:
path: "{{ item }}"
loop:
- "{{ cephadm_ssh_priv_key_path }}"
- "{{ cephadm_ssh_pub_key_path }}"
register: ssh_keys
changed_when: false
run_once: true
delegate_to: '{{ groups[mon_group_name][0] }}'
- name: set fact
set_fact:
stat_ssh_key_pair: "{{ ssh_keys.results | map(attribute='stat.exists') | list }}"
- name: fail if either ssh public or private key is missing
fail:
msg: "One part of the ssh keypair of user {{ cephadm_ssh_user }} is missing"
when:
- false in stat_ssh_key_pair
- true in stat_ssh_key_pair
- name: generate cephadm ssh key if there is none
command: "{{ ceph_cmd }} cephadm generate-key"
when: not true in stat_ssh_key_pair
changed_when: false
run_once: true
delegate_to: '{{ groups[mon_group_name][0] }}'
- name: use existing user keypair for remote connections
when: not false in stat_ssh_key_pair
delegate_to: "{{ groups[mon_group_name][0] }}"
run_once: true
command: >
{{ container_binary + ' run --rm --net=host --security-opt label=disable
-v /etc/ceph:/etc/ceph:z
-v /var/lib/ceph:/var/lib/ceph:ro
-v /var/run/ceph:/var/run/ceph:z
-v ' + item.1 + ':/etc/ceph/cephadm.' + item.0 + ':ro --entrypoint=ceph '+ ceph_docker_registry + '/' + ceph_docker_image + ':' + ceph_docker_image_tag if containerized_deployment | bool else 'ceph' }}
--cluster {{ cluster }} cephadm set-{{ item.0 }}-key -i /etc/ceph/cephadm.{{ item.0 }}
with_together:
- [ 'pub', 'priv' ]
- [ '{{ cephadm_ssh_pub_key_path }}', '{{ cephadm_ssh_priv_key_path }}' ]
- name: get the cephadm ssh pub key
command: "{{ ceph_cmd }} cephadm get-pub-key"
changed_when: false
run_once: true
register: cephadm_pubpkey
delegate_to: '{{ groups[mon_group_name][0] }}'
- name: allow cephadm key for {{ cephadm_ssh_user }} account
authorized_key:
user: "{{ cephadm_ssh_user }}"
key: '{{ cephadm_pubpkey.stdout }}'
- name: set cephadm ssh user to {{ cephadm_ssh_user }}
command: "{{ ceph_cmd }} cephadm set-user {{ cephadm_ssh_user }}"
changed_when: false
run_once: true
delegate_to: "{{ groups[mon_group_name][0] }}"
- name: run cephadm prepare-host
command: cephadm prepare-host
changed_when: false
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: set default container image in ceph configuration
command: "{{ ceph_cmd }} config set global container_image {{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}"
changed_when: false
run_once: true
delegate_to: '{{ groups[mon_group_name][0] }}'
- name: set container image base in ceph configuration
command: "{{ ceph_cmd }} config set mgr mgr/cephadm/container_image_base {{ ceph_docker_registry }}/{{ ceph_docker_image }}"
changed_when: false
run_once: true
delegate_to: '{{ groups[mon_group_name][0] }}'
- name: set dashboard container image in ceph mgr configuration
when: dashboard_enabled | bool
run_once: true
block:
- name: set alertmanager container image in ceph configuration
command: "{{ ceph_cmd }} config set mgr mgr/cephadm/container_image_alertmanager {{ alertmanager_container_image }}"
changed_when: false
delegate_to: '{{ groups[mon_group_name][0] }}'
- name: set grafana container image in ceph configuration
command: "{{ ceph_cmd }} config set mgr mgr/cephadm/container_image_grafana {{ grafana_container_image }}"
changed_when: false
delegate_to: '{{ groups[mon_group_name][0] }}'
- name: set node-exporter container image in ceph configuration
command: "{{ ceph_cmd }} config set mgr mgr/cephadm/container_image_node_exporter {{ node_exporter_container_image }}"
changed_when: false
delegate_to: '{{ groups[mon_group_name][0] }}'
- name: set prometheus container image in ceph configuration
command: "{{ ceph_cmd }} config set mgr mgr/cephadm/container_image_prometheus {{ prometheus_container_image }}"
changed_when: false
delegate_to: '{{ groups[mon_group_name][0] }}'
- name: enable the osd memory autotune for hci environment
command: "{{ ceph_cmd }} config set osd osd_memory_target_autotune true"
changed_when: false
run_once: true
delegate_to: '{{ groups[mon_group_name][0] }}'
when: is_hci | bool
- name: set autotune_memory_target_ratio
command: "{{ ceph_cmd }} config set mgr mgr/cephadm/autotune_memory_target_ratio {{ '0.2' if is_hci | bool else '0.7' }}"
changed_when: false
delegate_to: "{{ groups[mon_group_name][0] }}"
run_once: true
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: manage nodes with cephadm - ipv4
command: "{{ ceph_cmd }} orch host add {{ ansible_facts['nodename'] }} {{ ansible_facts['all_ipv4_addresses'] | ips_in_ranges(cephadm_mgmt_network.split(',')) | first }} {{ group_names | intersect(adopt_label_group_names) | join(' ') }}"
changed_when: false
delegate_to: '{{ groups[mon_group_name][0] }}'
when: cephadm_mgmt_network is ansible.utils.ipv4
- name: manage nodes with cephadm - ipv6
command: "{{ ceph_cmd }} orch host add {{ ansible_facts['nodename'] }} {{ ansible_facts['all_ipv6_addresses'] | ips_in_ranges(cephadm_mgmt_network.split(',')) | last | ansible.utils.ipwrap }} {{ group_names | intersect(adopt_label_group_names) | join(' ') }}"
changed_when: false
delegate_to: '{{ groups[mon_group_name][0] }}'
when: cephadm_mgmt_network is ansible.utils.ipv6
- name: add ceph label for core component
command: "{{ ceph_cmd }} orch host label add {{ ansible_facts['nodename'] }} ceph"
changed_when: false
delegate_to: '{{ groups[mon_group_name][0] }}'
when: inventory_hostname in groups.get(mon_group_name, []) or
inventory_hostname in groups.get(osd_group_name, []) or
inventory_hostname in groups.get(mds_group_name, []) or
inventory_hostname in groups.get(rgw_group_name, []) or
inventory_hostname in groups.get(mgr_group_name, []) or
inventory_hostname in groups.get(rbdmirror_group_name, [])
- name: get the client.admin keyring
ceph_key:
name: client.admin
cluster: "{{ cluster }}"
output_format: plain
state: info
environment:
CEPH_CONTAINER_IMAGE: "{{ ceph_docker_registry + '/' + ceph_docker_image + ':' + ceph_docker_image_tag if containerized_deployment | bool else None }}"
CEPH_CONTAINER_BINARY: "{{ container_binary }}"
run_once: true
delegate_to: '{{ groups[mon_group_name][0] }}'
register: client_admin_keyring
- name: copy the client.admin keyring
copy:
dest: "/etc/ceph/{{ cluster }}.client.admin.keyring"
content: "{{ client_admin_keyring.stdout + '\n' }}"
owner: "{{ ceph_uid | int if containerized_deployment | bool else 'ceph' }}"
group: "{{ ceph_uid | int if containerized_deployment | bool else 'ceph' }}"
mode: "{{ ceph_keyring_permissions }}"
run_once: true
delegate_to: "{{ item }}"
with_items:
- "{{ groups.get(osd_group_name, []) }}"
- "{{ groups.get(mds_group_name, []) }}"
- "{{ groups.get(rgw_group_name, []) }}"
- "{{ groups.get(mgr_group_name, []) }}"
- "{{ groups.get(rbdmirror_group_name, []) }}"
- name: assimilate ceph configuration
command: "{{ ceph_cmd }} config assimilate-conf -i /etc/ceph/{{ cluster }}.conf"
changed_when: false
when: inventory_hostname in groups.get(mon_group_name, []) or
inventory_hostname in groups.get(osd_group_name, []) or
inventory_hostname in groups.get(mds_group_name, []) or
inventory_hostname in groups.get(rgw_group_name, []) or
inventory_hostname in groups.get(mgr_group_name, []) or
inventory_hostname in groups.get(rbdmirror_group_name, [])
- name: set_fact cephadm_cmd
set_fact:
cephadm_cmd: "cephadm {{ '--docker' if container_binary == 'docker' else '' }}"
- name: set container registry info
command: "{{ ceph_cmd }} cephadm registry-login {{ ceph_docker_registry }} {{ ceph_docker_registry_username }} {{ ceph_docker_registry_password }}"
changed_when: false
no_log: true
run_once: true
delegate_to: '{{ groups[mon_group_name][0] }}'
when: ceph_docker_registry_auth | bool
- name: remove logrotate configuration
file:
path: /etc/logrotate.d/ceph
state: absent
when: inventory_hostname in groups.get(mon_group_name, []) or
inventory_hostname in groups.get(osd_group_name, []) or
inventory_hostname in groups.get(mds_group_name, []) or
inventory_hostname in groups.get(rgw_group_name, []) or
inventory_hostname in groups.get(mgr_group_name, []) or
inventory_hostname in groups.get(rbdmirror_group_name, []) or
inventory_hostname in groups.get(iscsi_gw_group_name, [])
- name: store existing rbd mirror peers in monitor config store
hosts: "{{ rbdmirror_group_name|default('rbdmirrors') }}"
become: true
any_errors_fatal: true
gather_facts: true
tasks:
- name: store existing rbd mirror peers in monitor config store
when:
- ceph_rbd_mirror_configure | default(True) | bool
- ceph_rbd_mirror_remote_user is undefined
block:
- name: import ceph-defaults
import_role:
name: ceph-defaults
- name: import ceph-validate
import_role:
name: ceph-validate
tasks_from: check_rbdmirror.yml
- name: import container_binary
import_role:
name: ceph-facts
tasks_from: container_binary.yml
- name: set_fact rbd_cmd
set_fact:
rbd_cmd: "{{ container_binary + ' run --rm --net=host -v /etc/ceph:/etc/ceph:z -v /var/lib/ceph:/var/lib/ceph:z -v /var/run/ceph:/var/run/ceph:z --entrypoint=rbd ' + ceph_docker_registry + '/' + ceph_docker_image + ':' + ceph_docker_image_tag if containerized_deployment | bool else 'ceph' }} --cluster {{ cluster }} -n client.rbd-mirror.{{ ansible_facts['hostname'] }} -k /etc/ceph/{{ cluster }}.client.rbd-mirror.{{ ansible_facts['hostname'] }}.keyring"
- name: set_fact admin_rbd_cmd
set_fact:
admin_rbd_cmd: "{{ container_binary + ' run --rm --net=host -v /etc/ceph:/etc/ceph:z -v /var/lib/ceph:/var/lib/ceph:z -v /var/run/ceph:/var/run/ceph:z --entrypoint=rbd ' + ceph_docker_registry + '/' + ceph_docker_image + ':' + ceph_docker_image_tag if containerized_deployment | bool else 'ceph' }} --cluster {{ cluster }}"
- name: get mirror pool info
command: "{{ rbd_cmd }} mirror pool info {{ ceph_rbd_mirror_pool }} --format json"
register: mirror_pool_info
changed_when: false
- name: set_fact mirror_peer_found
set_fact:
mirror_peer_uuid: "{{ ((mirror_pool_info.stdout | default('{}') | from_json)['peers'] | selectattr('site_name', 'match', '^'+ceph_rbd_mirror_remote_cluster+'$') | map(attribute='uuid') | list) }}"
- name: remove current rbd mirror peer, add new peer into mon config store
when: mirror_peer_uuid | length > 0
block:
- name: get remote user keyring
slurp:
src: "/etc/ceph/{{ ceph_rbd_mirror_remote_cluster }}.{{ ceph_rbd_mirror_remote_user }}.keyring"
register: remote_user_keyring
- name: get quorum_status
command: "{{ cephadm_cmd }} shell -k /etc/ceph/{{ cluster }}.client.admin.keyring --fsid {{ fsid }} -- ceph quorum_status --format json"
changed_when: false
delegate_to: "{{ groups[mon_group_name][0] }}"
register: quorum_status
run_once: true
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: set_fact mon_ip_list
set_fact:
mon_ip_list: "{{ mon_ip_list | default([]) | union([item['addr'].split(':')[0]]) }}"
loop: "{{ (quorum_status.stdout | default('{}') | from_json)['monmap']['mons'] }}"
run_once: true
- name: remove current mirror peer
command: "{{ admin_rbd_cmd }} mirror pool peer remove {{ ceph_rbd_mirror_pool }} {{ ((mirror_pool_info.stdout | default('{}') | from_json)['peers'] | selectattr('site_name', 'match', '^'+ceph_rbd_mirror_remote_cluster+'$') | map(attribute='uuid') | list)[0] }}"
delegate_to: "{{ groups.get(mon_group_name | default('mons'))[0] }}"
changed_when: false
- name: get remote user keyring secret
set_fact:
remote_user_keyring_secret: "{{ item.split('=', 1)[1] | trim }}"
with_items: "{{ (remote_user_keyring.content | b64decode).split('\n') }}"
when: "'key = ' in item"
- name: create a temporary file
tempfile:
path: /etc/ceph
state: file
suffix: _ceph-ansible
register: tmp_file
delegate_to: "{{ groups.get(mon_group_name | default('mons'))[0] }}"
- name: write secret to temporary file
copy:
dest: "{{ tmp_file.path }}"
content: "{{ remote_user_keyring_secret }}"
delegate_to: "{{ groups.get(mon_group_name | default('mons'))[0] }}"
- name: re-add mirror peer
command: "{{ admin_rbd_cmd }} mirror pool peer add {{ ceph_rbd_mirror_pool }} {{ ceph_rbd_mirror_remote_user }}@{{ ceph_rbd_mirror_remote_cluster }} --remote-mon-host {{ ','.join(mon_ip_list) }} --remote-key-file {{ tmp_file.path }}"
delegate_to: "{{ groups.get(mon_group_name | default('mons'))[0] }}"
changed_when: false
- name: rm temporary file
file:
path: "{{ tmp_file.path }}"
state: absent
delegate_to: "{{ groups.get(mon_group_name | default('mons'))[0] }}"
- name: adopt ceph mon daemons
hosts: "{{ mon_group_name|default('mons') }}"
serial: 1
become: true
gather_facts: false
any_errors_fatal: True
tasks:
- import_role:
name: ceph-defaults
- name: adopt mon daemon
cephadm_adopt:
name: "mon.{{ ansible_facts['hostname'] }}"
cluster: "{{ cluster }}"
image: "{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}"
docker: "{{ true if container_binary == 'docker' else false }}"
pull: false
firewalld: "{{ true if configure_firewall | bool else false }}"
- name: reset failed ceph-mon systemd unit
command: "systemctl reset-failed ceph-mon@{{ ansible_facts['hostname'] }}" # noqa 303
changed_when: false
failed_when: false
when: containerized_deployment | bool
- name: remove ceph-mon systemd files
file:
path: "{{ item }}"
state: absent
loop:
- /etc/systemd/system/ceph-mon@.service
- /etc/systemd/system/ceph-mon@.service.d
- /etc/systemd/system/ceph-mon.target
- name: waiting for the monitor to join the quorum...
command: "{{ cephadm_cmd }} shell -k /etc/ceph/{{ cluster }}.client.admin.keyring --fsid {{ fsid }} -- ceph quorum_status --format json"
changed_when: false
register: ceph_health_raw
until: >
ansible_facts['hostname'] in (ceph_health_raw.stdout | from_json)["quorum_names"]
retries: "{{ health_mon_check_retries }}"
delay: "{{ health_mon_check_delay }}"
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: adopt ceph mgr daemons
hosts: "{{ groups[mgr_group_name] | default(groups[mon_group_name]) }}"
serial: 1
become: true
gather_facts: false
any_errors_fatal: True
tasks:
- import_role:
name: ceph-defaults
- name: adopt mgr daemon
cephadm_adopt:
name: "mgr.{{ ansible_facts['hostname'] }}"
cluster: "{{ cluster }}"
image: "{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}"
docker: "{{ true if container_binary == 'docker' else false }}"
pull: false
firewalld: "{{ true if configure_firewall | bool else false }}"
- name: reset failed ceph-mgr systemd unit
command: "systemctl reset-failed ceph-mgr@{{ ansible_facts['hostname'] }}" # noqa 303
changed_when: false
failed_when: false
when: containerized_deployment | bool
- name: remove ceph-mgr systemd files
file:
path: "{{ item }}"
state: absent
loop:
- /etc/systemd/system/ceph-mgr@.service
- /etc/systemd/system/ceph-mgr@.service.d
- /etc/systemd/system/ceph-mgr.target
- name: stop and remove legacy iscsigw daemons
hosts: "{{ iscsi_gw_group_name|default('iscsigws') }}"
serial: 1
become: true
gather_facts: false
any_errors_fatal: True
tasks:
- import_role:
name: ceph-defaults
- name: stop and disable iscsigw systemd services
service:
name: '{{ item }}'
state: stopped
enabled: false
failed_when: false
with_items:
- rbd-target-api
- rbd-target-gw
- tcmu-runner
- name: reset failed iscsigw systemd units
command: 'systemctl reset-failed {{ item }}' # noqa 303
changed_when: false
failed_when: false
with_items:
- rbd-target-api
- rbd-target-gw
- tcmu-runner
when: containerized_deployment | bool
- name: remove iscsigw systemd unit files
file:
path: '/etc/systemd/system/{{ item }}.service'
state: absent
with_items:
- rbd-target-api
- rbd-target-gw
- tcmu-runner
when: containerized_deployment | bool
- name: redeploy iscsigw daemons
hosts: "{{ iscsi_gw_group_name|default('iscsigws') }}"
become: true
gather_facts: false
any_errors_fatal: True
tasks:
- import_role:
name: ceph-defaults
- name: update the placement of iscsigw hosts
command: "{{ cephadm_cmd }} shell -k /etc/ceph/{{ cluster }}.client.admin.keyring --fsid {{ fsid }} -- ceph orch apply iscsi {{ iscsi_pool_name | default('rbd') }} {{ api_user | default('admin') }} {{ api_password | default('admin') }} {{ trusted_ip_list | default('192.168.122.1') }} --placement='{{ groups.get(iscsi_gw_group_name, []) | length }} label:{{ iscsi_gw_group_name }}'"
run_once: true
changed_when: false
delegate_to: '{{ groups[mon_group_name][0] }}'
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: set osd flags
hosts: "{{ osd_group_name|default('osds') }}"
become: true
gather_facts: false
any_errors_fatal: True
tasks:
- import_role:
name: ceph-defaults
- name: get pool list
command: "{{ ceph_cmd }} --cluster {{ cluster }} osd pool ls detail -f json"
register: pool_list
run_once: true
delegate_to: "{{ groups[mon_group_name][0] }}"
changed_when: false
check_mode: false
- name: get balancer module status
command: "{{ ceph_cmd }} --cluster {{ cluster }} balancer status -f json"
register: balancer_status_adopt
run_once: true
delegate_to: "{{ groups[mon_group_name][0] }}"
changed_when: false
check_mode: false
- name: set_fact pools_pgautoscaler_mode
set_fact:
pools_pgautoscaler_mode: "{{ pools_pgautoscaler_mode | default([]) | union([{'name': item.pool_name, 'mode': item.pg_autoscale_mode}]) }}"
run_once: true
with_items: "{{ pool_list.stdout | default('{}') | from_json }}"
- name: disable balancer
command: "{{ ceph_cmd }} --cluster {{ cluster }} balancer off"
run_once: true
delegate_to: "{{ groups[mon_group_name][0] }}"
changed_when: false
when: (balancer_status_adopt.stdout | from_json)['active'] | bool
- name: disable pg autoscale on pools
ceph_pool:
name: "{{ item.name }}"
cluster: "{{ cluster }}"
pg_autoscale_mode: false
with_items: "{{ pools_pgautoscaler_mode }}"
delegate_to: "{{ groups[mon_group_name][0] }}"
run_once: true
when:
- pools_pgautoscaler_mode is defined
- item.mode == 'on'
environment:
CEPH_CONTAINER_IMAGE: "{{ ceph_docker_registry + '/' + ceph_docker_image + ':' + ceph_docker_image_tag if containerized_deployment | bool else None }}"
CEPH_CONTAINER_BINARY: "{{ container_binary }}"
- name: set osd flags
ceph_osd_flag:
cluster: "{{ cluster }}"
name: "{{ item }}"
state: present
with_items:
- noout
- nodeep-scrub
delegate_to: "{{ groups[mon_group_name][0] }}"
run_once: true
environment:
CEPH_CONTAINER_IMAGE: "{{ ceph_docker_registry + '/' + ceph_docker_image + ':' + ceph_docker_image_tag if containerized_deployment | bool else None }}"
CEPH_CONTAINER_BINARY: "{{ container_binary }}"
- name: adopt ceph osd daemons
hosts: "{{ osd_group_name|default('osd') }}"
serial: 1
become: true
gather_facts: false
any_errors_fatal: True
tasks:
- import_role:
name: ceph-defaults
- import_role:
name: ceph-facts
tasks_from: container_binary.yml
when: containerized_deployment | bool
- name: get osd list
ceph_volume:
cluster: "{{ cluster }}"
action: list
environment:
CEPH_CONTAINER_IMAGE: "{{ ceph_docker_registry + '/' + ceph_docker_image + ':' + ceph_docker_image_tag if containerized_deployment | bool else None }}"
CEPH_CONTAINER_BINARY: "{{ container_binary }}"
register: osd_list
- name: set osd fsid for containerized deployment
lineinfile:
path: '/var/lib/ceph/osd/{{ cluster }}-{{ item.key }}/fsid'
line: "{{ (item.value | selectattr('type', 'equalto', 'block') | map(attribute='tags') | first)['ceph.osd_fsid'] }}"
owner: '{{ ceph_uid }}'
group: '{{ ceph_uid }}'
create: true
with_dict: '{{ osd_list.stdout | from_json }}'
when: containerized_deployment | bool
- name: set osd type for containerized deployment
lineinfile:
path: '/var/lib/ceph/osd/{{ cluster }}-{{ item }}/type'
line: 'bluestore'
owner: '{{ ceph_uid }}'
group: '{{ ceph_uid }}'
create: true
loop: '{{ (osd_list.stdout | from_json).keys() | list }}'
when: containerized_deployment | bool
- name: adopt osd daemon
cephadm_adopt:
name: "osd.{{ item }}"
cluster: "{{ cluster }}"
image: "{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}"
docker: "{{ true if container_binary == 'docker' else false }}"
pull: false
firewalld: "{{ true if configure_firewall | bool else false }}"
loop: '{{ (osd_list.stdout | from_json).keys() | list }}'
- name: remove ceph-osd systemd and ceph-osd-run.sh files
file:
path: "{{ item }}"
state: absent
loop:
- /etc/systemd/system/ceph-osd@.service
- /etc/systemd/system/ceph-osd@.service.d
- /etc/systemd/system/ceph-osd.target
- "{{ ceph_osd_docker_run_script_path | default('/usr/share') }}/ceph-osd-run.sh"
- name: remove osd directory
file:
path: "/var/lib/ceph/osd/{{ cluster }}-{{ item }}"
state: absent
loop: '{{ (osd_list.stdout | from_json).keys() | list }}'
- name: remove any legacy directories in /var/lib/ceph/mon (workaround)
file:
path: "/var/lib/ceph/mon/{{ cluster }}-{{ ansible_facts['hostname'] }}"
state: absent
- name: waiting for clean pgs...
command: "{{ cephadm_cmd }} shell -k /etc/ceph/{{ cluster }}.client.admin.keyring --fsid {{ fsid }} -- ceph pg stat --format json"
changed_when: false
register: ceph_health_post
until: >
(((ceph_health_post.stdout | from_json).pg_summary.num_pg_by_state | length) > 0)
and
(((ceph_health_post.stdout | from_json).pg_summary.num_pg_by_state | selectattr('name', 'search', '^active\\+clean') | map(attribute='num') | list | sum) == (ceph_health_post.stdout | from_json).pg_summary.num_pgs)
delegate_to: "{{ groups[mon_group_name][0] }}"
retries: "{{ health_osd_check_retries }}"
delay: "{{ health_osd_check_delay }}"
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: unset osd flags
hosts: "{{ osd_group_name|default('osds') }}"
become: true
gather_facts: false
any_errors_fatal: True
tasks:
- import_role:
name: ceph-defaults
- name: re-enable pg autoscale on pools
ceph_pool:
name: "{{ item.name }}"
cluster: "{{ cluster }}"
pg_autoscale_mode: true
with_items: "{{ pools_pgautoscaler_mode }}"
delegate_to: "{{ groups[mon_group_name][0] }}"
run_once: true
when:
- pools_pgautoscaler_mode is defined
- item.mode == 'on'
environment:
CEPH_CONTAINER_IMAGE: "{{ ceph_docker_registry + '/' + ceph_docker_image + ':' + ceph_docker_image_tag if containerized_deployment | bool else None }}"
CEPH_CONTAINER_BINARY: "{{ container_binary }}"
- name: unset osd flags
ceph_osd_flag:
cluster: "{{ cluster }}"
name: "{{ item }}"
state: absent
with_items:
- noout
- nodeep-scrub
delegate_to: "{{ groups[mon_group_name][0] }}"
run_once: true
environment:
CEPH_CONTAINER_IMAGE: "{{ ceph_docker_registry + '/' + ceph_docker_image + ':' + ceph_docker_image_tag if containerized_deployment | bool else None }}"
CEPH_CONTAINER_BINARY: "{{ container_binary }}"
- name: re-enable balancer
command: "{{ ceph_cmd }} --cluster {{ cluster }} balancer on"
run_once: true
delegate_to: "{{ groups[mon_group_name][0] }}"
changed_when: false
when: (balancer_status_adopt.stdout | from_json)['active'] | bool
- name: redeploy mds daemons
hosts: "{{ mds_group_name|default('mdss') }}"
become: true
gather_facts: false
any_errors_fatal: True
tasks:
- import_role:
name: ceph-defaults
- name: update the placement of metadata hosts
command: "{{ cephadm_cmd }} shell -k /etc/ceph/{{ cluster }}.client.admin.keyring --fsid {{ fsid }} -- ceph orch apply mds {{ cephfs }} --placement='{{ groups.get(mds_group_name, []) | length }} label:{{ mds_group_name }}'"
run_once: true
changed_when: false
delegate_to: "{{ groups[mon_group_name][0] }}"
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: stop and remove legacy ceph mds daemons
hosts: "{{ mds_group_name|default('mdss') }}"
serial: 1
become: true
gather_facts: false
any_errors_fatal: True
tasks:
- import_role:
name: ceph-defaults
- name: stop and disable ceph-mds systemd service
service:
name: "ceph-mds@{{ ansible_facts['hostname'] }}"
state: stopped
enabled: false
failed_when: false
- name: stop and disable ceph-mds systemd target
service:
name: ceph-mds.target
state: stopped
enabled: false
failed_when: false
- name: reset failed ceph-mds systemd unit
command: "systemctl reset-failed ceph-mds@{{ ansible_facts['hostname'] }}" # noqa 303
changed_when: false
failed_when: false
when: containerized_deployment | bool
- name: remove ceph-mds systemd files
file:
path: "{{ item }}"
state: absent
loop:
- /etc/systemd/system/ceph-mds@.service
- /etc/systemd/system/ceph-mds@.service.d
- /etc/systemd/system/ceph-mds.target
- name: remove legacy ceph mds data
file:
path: "/var/lib/ceph/mds/{{ cluster }}-{{ ansible_facts['hostname'] }}"
state: absent
- name: redeploy rgw daemons
hosts: "{{ rgw_group_name | default('rgws') }}"
become: true
gather_facts: false
any_errors_fatal: True
tasks:
- import_role:
name: ceph-defaults
- import_role:
name: ceph-facts
tasks_from: set_radosgw_address.yml
- name: import rgw ssl certificate into kv store
when: radosgw_frontend_ssl_certificate | length > 0
block:
- name: slurp rgw ssl certificate
slurp:
src: "{{ radosgw_frontend_ssl_certificate }}"
register: rgw_ssl_cert
- name: store ssl certificate in kv store (not multisite)
command: >
{{ container_binary }} run --rm -i -v /etc/ceph:/etc/ceph:z --entrypoint=ceph {{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }} --cluster {{ cluster }}
config-key set rgw/cert/rgw.{{ ansible_facts['hostname'] }} -i -
args:
stdin: "{{ rgw_ssl_cert.content | b64decode }}"
stdin_add_newline: no
changed_when: false
when: not rgw_multisite | bool
delegate_to: "{{ groups[mon_group_name][0] }}"
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: store ssl certificate in kv store (multisite)
command: >
{{ container_binary }} run --rm -i -v /etc/ceph:/etc/ceph:z --entrypoint=ceph {{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }} --cluster {{ cluster }}
config-key set rgw/cert/rgw.{{ ansible_facts['hostname'] }}.{{ item.rgw_realm }}.{{ item.rgw_zone }}.{{ item.radosgw_frontend_port }} -i -
args:
stdin: "{{ rgw_ssl_cert.content | b64decode }}"
stdin_add_newline: no
changed_when: false
loop: "{{ rgw_instances }}"
when: rgw_multisite | bool
delegate_to: "{{ groups[mon_group_name][0] }}"
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: update the placement of radosgw hosts
command: >
{{ cephadm_cmd }} shell -k /etc/ceph/{{ cluster }}.client.admin.keyring --fsid {{ fsid }} --
ceph orch apply rgw {{ ansible_facts['hostname'] }}
--placement='count-per-host:{{ radosgw_num_instances }} {{ ansible_facts['nodename'] }}'
--port={{ radosgw_frontend_port }}
{{ '--ssl' if radosgw_frontend_ssl_certificate else '' }}
changed_when: false
delegate_to: "{{ groups[mon_group_name][0] }}"
when: not rgw_multisite | bool
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: update the placement of radosgw multisite hosts
command: >
{{ cephadm_cmd }} shell -k /etc/ceph/{{ cluster }}.client.admin.keyring --fsid {{ fsid }} --
ceph orch apply rgw {{ ansible_facts['hostname'] }}.{{ item.rgw_realm }}.{{ item.rgw_zone }}.{{ item.radosgw_frontend_port }}
--placement={{ ansible_facts['nodename'] }}
--realm={{ item.rgw_realm }} --zone={{ item.rgw_zone }}
--port={{ item.radosgw_frontend_port }}
{{ '--ssl' if radosgw_frontend_ssl_certificate else '' }}
changed_when: false
loop: "{{ rgw_instances }}"
when: rgw_multisite | bool
delegate_to: "{{ groups[mon_group_name][0] }}"
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: stop and remove legacy ceph rgw daemons
hosts: "{{ rgw_group_name|default('rgws') }}"
serial: 1
become: true
gather_facts: false
any_errors_fatal: True
tasks:
- import_role:
name: ceph-defaults
- import_role:
name: ceph-facts
tasks_from: set_radosgw_address.yml
- name: stop and disable ceph-radosgw systemd service
service:
name: "ceph-radosgw@rgw.{{ ansible_facts['hostname'] }}.{{ item.instance_name }}"
state: stopped
enabled: false
failed_when: false
loop: '{{ rgw_instances }}'
- name: stop and disable ceph-radosgw systemd target
service:
name: ceph-radosgw.target
state: stopped
enabled: false
failed_when: false
- name: reset failed ceph-radosgw systemd unit
command: "systemctl reset-failed ceph-radosgw@rgw.{{ ansible_facts['hostname'] }}.{{ item.instance_name }}" # noqa 303
changed_when: false
failed_when: false
loop: '{{ rgw_instances }}'
when: containerized_deployment | bool
- name: remove ceph-radosgw systemd files
file:
path: "{{ item }}"
state: absent
loop:
- /etc/systemd/system/ceph-radosgw@.service
- /etc/systemd/system/ceph-radosgw@.service.d
- /etc/systemd/system/ceph-radosgw.target
- name: remove legacy ceph radosgw data
file:
path: "/var/lib/ceph/radosgw/{{ cluster }}-rgw.{{ ansible_facts['hostname'] }}.{{ item.instance_name }}"
state: absent
loop: '{{ rgw_instances }}'
- name: remove legacy ceph radosgw directory
file:
path: "/var/lib/ceph/radosgw/{{ cluster }}-rgw.{{ ansible_facts['hostname'] }}"
state: absent
- name: stop and remove legacy ceph nfs daemons
hosts: "{{ nfs_group_name|default('nfss') }}"
tags: 'ceph_nfs_adopt'
serial: 1
become: true
gather_facts: false
any_errors_fatal: True
tasks:
- import_role:
name: ceph-defaults
- import_role:
name: ceph-nfs
tasks_from: create_rgw_nfs_user.yml
- name: enable ceph mgr nfs module
ceph_mgr_module:
name: "nfs"
cluster: "{{ cluster }}"
state: enable
environment:
CEPH_CONTAINER_IMAGE: "{{ ceph_docker_registry + '/' + ceph_docker_image + ':' + ceph_docker_image_tag if containerized_deployment | bool else None }}"
CEPH_CONTAINER_BINARY: "{{ container_binary }}"
delegate_to: "{{ groups[mon_group_name][0] }}"
- name: stop and disable ceph-nfs systemd service
service:
name: "ceph-nfs@{{ ansible_facts['hostname'] }}"
state: stopped
enabled: false
failed_when: false
- name: reset failed ceph-nfs systemd unit
command: "systemctl reset-failed ceph-nfs@{{ ansible_facts['hostname'] }}" # noqa 303
changed_when: false
failed_when: false
when: containerized_deployment | bool
- name: remove ceph-nfs systemd unit files
file:
path: "{{ item }}"
state: absent
loop:
- /etc/systemd/system/ceph-nfs@.service
- /etc/systemd/system/ceph-nfs@.service.d
- name: remove legacy ceph radosgw directory
file:
path: "/var/lib/ceph/radosgw/{{ cluster }}-rgw.{{ ansible_facts['hostname'] }}"
state: absent
- name: create nfs ganesha cluster
command: "{{ cephadm_cmd }} shell -k /etc/ceph/{{ cluster }}.client.admin.keyring --fsid {{ fsid }} -- ceph nfs cluster create {{ ansible_facts['hostname'] }} {{ ansible_facts['hostname'] }}"
changed_when: false
delegate_to: "{{ groups[mon_group_name][0] }}"
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: create cephfs export
command: "{{ cephadm_cmd }} shell -k /etc/ceph/{{ cluster }}.client.admin.keyring --fsid {{ fsid }} -- ceph nfs export create cephfs {{ cephfs }} {{ ansible_facts['hostname'] }} {{ ceph_nfs_ceph_pseudo_path }} --squash {{ ceph_nfs_ceph_squash }}"
changed_when: false
delegate_to: "{{ groups[mon_group_name][0] }}"
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
when: nfs_file_gw | bool
- name: create rgw export
command: "{{ cephadm_cmd }} shell -k /etc/ceph/{{ cluster }}.client.admin.keyring --fsid {{ fsid }} -- ceph nfs export create rgw --cluster-id {{ ansible_facts['hostname'] }} --pseudo-path {{ ceph_nfs_rgw_pseudo_path }} --user-id {{ ceph_nfs_rgw_user }} --squash {{ ceph_nfs_rgw_squash }}"
changed_when: false
delegate_to: "{{ groups[mon_group_name][0] }}"
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
when: nfs_obj_gw | bool
- name: redeploy rbd-mirror daemons
hosts: "{{ rbdmirror_group_name|default('rbdmirrors') }}"
become: true
gather_facts: false
any_errors_fatal: True
tasks:
- import_role:
name: ceph-defaults
- name: update the placement of rbd-mirror hosts
command: "{{ cephadm_cmd }} shell -k /etc/ceph/{{ cluster }}.client.admin.keyring --fsid {{ fsid }} -- ceph orch apply rbd-mirror --placement='{{ groups.get(rbdmirror_group_name, []) | length }} label:{{ rbdmirror_group_name }}'"
run_once: true
changed_when: false
delegate_to: "{{ groups[mon_group_name][0] }}"
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: stop and remove legacy rbd-mirror daemons
hosts: "{{ rbdmirror_group_name|default('rbdmirrors') }}"
serial: 1
become: true
gather_facts: false
any_errors_fatal: True
tasks:
- import_role:
name: ceph-defaults
- name: stop and disable rbd-mirror systemd service
service:
name: "ceph-rbd-mirror@rbd-mirror.{{ ansible_facts['hostname'] }}"
state: stopped
enabled: false
failed_when: false
- name: stop and disable rbd-mirror systemd target
service:
name: ceph-rbd-mirror.target
state: stopped
enabled: false
failed_when: false
- name: reset failed rbd-mirror systemd unit
command: "systemctl reset-failed ceph-rbd-mirror@rbd-mirror.{{ ansible_facts['hostname'] }}" # noqa 303
changed_when: false
failed_when: false
when: containerized_deployment | bool
- name: remove rbd-mirror systemd files
file:
path: "{{ item }}"
state: absent
loop:
- /etc/systemd/system/ceph-rbd-mirror@.service
- /etc/systemd/system/ceph-rbd-mirror@.service.d
- /etc/systemd/system/ceph-rbd-mirror.target
- name: redeploy ceph-crash daemons
hosts:
- "{{ mon_group_name|default('mons') }}"
- "{{ osd_group_name|default('osds') }}"
- "{{ mds_group_name|default('mdss') }}"
- "{{ rgw_group_name|default('rgws') }}"
- "{{ mgr_group_name|default('mgrs') }}"
- "{{ rbdmirror_group_name|default('rbdmirrors') }}"
become: true
gather_facts: false
any_errors_fatal: True
tasks:
- import_role:
name: ceph-defaults
- name: stop and disable ceph-crash systemd service
service:
name: "{{ 'ceph-crash@' + ansible_facts['hostname'] if containerized_deployment | bool else 'ceph-crash.service' }}"
state: stopped
enabled: false
failed_when: false
- name: remove ceph-crash systemd unit file
file:
path: /etc/systemd/system/ceph-crash@.service
state: absent
- name: update the placement of ceph-crash hosts
command: "{{ cephadm_cmd }} shell -k /etc/ceph/{{ cluster }}.client.admin.keyring --fsid {{ fsid }} -- ceph orch apply crash --placement='label:ceph'"
run_once: true
changed_when: false
delegate_to: '{{ groups[mon_group_name][0] }}'
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: redeploy alertmanager/grafana/prometheus daemons
hosts: "{{ monitoring_group_name|default('monitoring') }}"
serial: 1
become: true
gather_facts: false
any_errors_fatal: True
tasks:
- import_role:
name: ceph-defaults
- name: check whether a ceph config file is present
stat:
path: "/etc/ceph/{{ cluster }}.conf"
register: ceph_config
- name: ensure /etc/ceph is present
file:
path: /etc/ceph
state: directory
owner: "{{ ceph_uid | int if containerized_deployment | bool else 'ceph' }}"
group: "{{ ceph_uid | int if containerized_deployment | bool else 'ceph' }}"
mode: "{{ ceph_directories_mode }}"
- name: write a ceph.conf with minimal config
copy:
dest: "/etc/ceph/{{ cluster }}.conf"
content: "{{ minimal_config.stdout }}"
owner: "{{ ceph_uid | int if containerized_deployment | bool else 'ceph' }}"
group: "{{ ceph_uid | int if containerized_deployment | bool else 'ceph' }}"
mode: "{{ ceph_keyring_permissions }}"
when: not ceph_config.stat.exists | bool
- name: with dashboard enabled
when: dashboard_enabled | bool
block:
- name: ensure alertmanager/prometheus data directories are present
file:
path: "{{ item }}"
state: directory
owner: "{{ prometheus_user_id }}"
group: "{{ prometheus_user_id }}"
with_items:
- "{{ alertmanager_data_dir }}"
- "{{ prometheus_data_dir }}"
# (workaround) cephadm adopt alertmanager only stops prometheus-alertmanager systemd service
- name: stop and disable alertmanager systemd unit
service:
name: alertmanager
state: stopped
enabled: false
failed_when: false
# (workaround) cephadm adopt alertmanager only uses /etc/prometheus/alertmanager.yml
- name: create alertmanager config symlink
file:
path: /etc/prometheus/alertmanager.yml
src: '{{ alertmanager_conf_dir }}/alertmanager.yml'
state: link
# (workaround) cephadm adopt alertmanager only uses /var/lib/prometheus/alertmanager/
- name: create alertmanager data symlink
file:
path: '{{ prometheus_data_dir }}/alertmanager'
src: '{{ alertmanager_data_dir }}'
state: link
- name: adopt alertmanager daemon
cephadm_adopt:
name: "alertmanager.{{ ansible_facts['hostname'] }}"
cluster: "{{ cluster }}"
image: "{{ alertmanager_container_image }}"
docker: "{{ true if container_binary == 'docker' else false }}"
pull: false
firewalld: "{{ true if configure_firewall | bool else false }}"
- name: remove alertmanager systemd unit file
file:
path: /etc/systemd/system/alertmanager.service
state: absent
- name: remove the legacy alertmanager data
file:
path: '{{ alertmanager_data_dir }}'
state: absent
- name: stop and disable prometheus systemd unit
service:
name: prometheus
state: stopped
enabled: false
failed_when: false
- name: remove alertmanager data symlink
file:
path: '{{ prometheus_data_dir }}/alertmanager'
state: absent
# (workaround) cephadm adopt prometheus only uses /var/lib/prometheus/metrics/
- name: tmp copy the prometheus data
copy:
src: '{{ prometheus_data_dir }}/'
dest: /var/lib/prom_metrics
owner: 65534
group: 65534
remote_src: true
# (workaround) cephadm adopt prometheus only uses /var/lib/prometheus/metrics/
- name: restore the prometheus data
copy:
src: /var/lib/prom_metrics/
dest: /var/lib/prometheus/metrics
owner: 65534
group: 65534
remote_src: true
- name: remove the tmp prometheus data copy
file:
path: /var/lib/prom_metrics
state: absent
- name: adopt prometheus daemon
cephadm_adopt:
name: "prometheus.{{ ansible_facts['hostname'] }}"
cluster: "{{ cluster }}"
image: "{{ prometheus_container_image }}"
docker: "{{ true if container_binary == 'docker' else false }}"
pull: false
firewalld: "{{ true if configure_firewall | bool else false }}"
- name: remove prometheus systemd unit file
file:
path: /etc/systemd/system/prometheus.service
state: absent
- name: remove the legacy prometheus data
file:
path: '{{ prometheus_data_dir }}'
state: absent
# (workaround) cephadm adopt grafana only stops grafana systemd service
- name: stop and disable grafana systemd unit
service:
name: grafana-server
state: stopped
enabled: false
failed_when: false
- name: adopt grafana daemon
cephadm_adopt:
name: "grafana.{{ ansible_facts['hostname'] }}"
cluster: "{{ cluster }}"
image: "{{ grafana_container_image }}"
docker: "{{ true if container_binary == 'docker' else false }}"
pull: false
firewalld: "{{ true if configure_firewall | bool else false }}"
- name: remove grafana systemd unit file
file:
path: /etc/systemd/system/grafana-server.service
state: absent
- name: remove the legacy grafana data
file:
path: /var/lib/grafana
state: absent
- name: redeploy node-exporter daemons
hosts:
- "{{ mon_group_name|default('mons') }}"
- "{{ osd_group_name|default('osds') }}"
- "{{ mds_group_name|default('mdss') }}"
- "{{ rgw_group_name|default('rgws') }}"
- "{{ mgr_group_name|default('mgrs') }}"
- "{{ rbdmirror_group_name|default('rbdmirrors') }}"
- "{{ nfs_group_name|default('nfss') }}"
- "{{ iscsi_gw_group_name|default('iscsigws') }}"
- "{{ monitoring_group_name|default('monitoring') }}"
become: true
gather_facts: false
any_errors_fatal: True
tasks:
- import_role:
name: ceph-defaults
- name: with dashboard enabled
when: dashboard_enabled | bool
block:
- name: stop and disable node-exporter systemd service
service:
name: node_exporter
state: stopped
enabled: false
failed_when: false
- name: remove node_exporter systemd unit file
file:
path: /etc/systemd/system/node_exporter.service
state: absent
- name: update the placement of node-exporter hosts
command: "{{ cephadm_cmd }} shell -k /etc/ceph/{{ cluster }}.client.admin.keyring --fsid {{ fsid }} -- ceph orch apply node-exporter --placement='*'"
run_once: true
changed_when: false
delegate_to: '{{ groups[mon_group_name][0] }}'
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: adjust placement daemons
hosts: "{{ mon_group_name|default('mons') }}[0]"
become: true
gather_facts: false
any_errors_fatal: True
tasks:
- import_role:
name: ceph-defaults
- name: update the placement of monitor hosts
command: "{{ cephadm_cmd }} shell -k /etc/ceph/{{ cluster }}.client.admin.keyring --fsid {{ fsid }} -- ceph orch apply mon --placement='{{ groups.get(mon_group_name, []) | length }} label:{{ mon_group_name }}'"
changed_when: false
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: set_fact mgr_placement
set_fact:
mgr_placement_count: "{{ groups.get(mgr_group_name, []) | length if groups.get(mgr_group_name, []) | length > 0 else groups.get(mon_group_name, []) | length }}"
- name: set_fact mgr_placement_label
set_fact:
mgr_placement_label: "{{ mgr_group_name if groups.get(mgr_group_name, []) | length > 0 else mon_group_name }}"
- name: update the placement of manager hosts
command: "{{ cephadm_cmd }} shell -k /etc/ceph/{{ cluster }}.client.admin.keyring --fsid {{ fsid }} -- ceph orch apply mgr --placement='{{ mgr_placement_count }} label:{{ mgr_placement_label }}'"
changed_when: false
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: with dashboard enabled
when: dashboard_enabled | bool
block:
- name: update the placement of alertmanager hosts
command: "{{ cephadm_cmd }} shell -k /etc/ceph/{{ cluster }}.client.admin.keyring --fsid {{ fsid }} -- ceph orch apply alertmanager --placement='{{ groups.get(monitoring_group_name, []) | length }} label:{{ monitoring_group_name }}'"
changed_when: false
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: update the placement of grafana hosts
command: "{{ cephadm_cmd }} shell -k /etc/ceph/{{ cluster }}.client.admin.keyring --fsid {{ fsid }} -- ceph orch apply grafana --placement='{{ groups.get(monitoring_group_name, []) | length }} label:{{ monitoring_group_name }}'"
changed_when: false
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: update the placement of prometheus hosts
command: "{{ cephadm_cmd }} shell -k /etc/ceph/{{ cluster }}.client.admin.keyring --fsid {{ fsid }} -- ceph orch apply prometheus --placement='{{ groups.get(monitoring_group_name, []) | length }} label:{{ monitoring_group_name }}'"
changed_when: false
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: show ceph orchestrator status
hosts: "{{ mon_group_name|default('mons') }}[0]"
become: true
gather_facts: false
any_errors_fatal: True
tasks:
- import_role:
name: ceph-defaults
- name: show ceph orchestrator services
command: "{{ cephadm_cmd }} shell -k /etc/ceph/{{ cluster }}.client.admin.keyring --fsid {{ fsid }} -- ceph orch ls --refresh"
changed_when: false
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: show ceph orchestrator daemons
command: "{{ cephadm_cmd }} shell -k /etc/ceph/{{ cluster }}.client.admin.keyring --fsid {{ fsid }} -- ceph orch ps --refresh"
changed_when: false
environment:
CEPHADM_IMAGE: '{{ ceph_docker_registry }}/{{ ceph_docker_image }}:{{ ceph_docker_image_tag }}'
- name: inform users about cephadm
debug:
msg: |
This Ceph cluster is now managed by cephadm. Any new changes to the
cluster need to be achieved by using the cephadm CLI and you don't
need to use ceph-ansible playbooks anymore.