ceph-ansible/infrastructure-playbooks/switch-from-non-containeriz...

467 lines
13 KiB
YAML
Raw Normal View History

---
# This playbook switches from non-containerized to containerized Ceph daemons
- name: confirm whether user really meant to switch from non-containerized to containerized ceph daemons
hosts:
- localhost
gather_facts: false
vars_prompt:
- name: ireallymeanit
prompt: Are you sure you want to switch from non-containerized to containerized ceph daemons?
default: 'no'
private: no
tasks:
- name: exit playbook, if user did not mean to switch from non-containerized to containerized daemons?
fail:
msg: >
"Exiting switch-from-non-containerized-to-containerized-ceph-daemons.yml playbook,
cluster did not switch from non-containerized to containerized ceph daemons.
To switch from non-containerized to containerized ceph daemons, either say 'yes' on the prompt or
or use `-e ireallymeanit=yes` on the command line when
invoking the playbook"
when: ireallymeanit != 'yes'
- name: make sure docker is present and started
hosts:
- "{{ mon_group_name|default('mons') }}"
- "{{ mgr_group_name|default('mgrs') }}"
- "{{ osd_group_name|default('osds') }}"
- "{{ mds_group_name|default('mdss') }}"
- "{{ rgw_group_name|default('rgws') }}"
- "{{ rbdmirror_group_name|default('rbdmirrors') }}"
- "{{ nfs_group_name|default('nfss') }}"
become: true
tasks:
- name: check if selinux is enabled
command: getenforce
register: sestatus
changed_when: false
when: ansible_os_family == 'RedHat'
- name: set selinux permissions
command: chcon -Rt svirt_sandbox_file_t "{{ item }}"
with_items:
- /etc/ceph
- /var/lib/ceph
changed_when: false
when:
- sestatus.stdout != 'Disabled'
- ansible_os_family == 'RedHat'
- name: gather and delegate facts
setup:
delegate_to: "{{ item }}"
delegate_facts: True
with_items: "{{ groups['all'] }}"
- name: switching from non-containerized to containerized ceph mon
vars:
health_mon_check_retries: 5
health_mon_check_delay: 15
containerized_deployment: true
mon_group_name: mons
restapi_group_name: restapis
hosts:
- "{{ mon_group_name|default('mons') }}"
serial: 1
become: true
pre_tasks:
- name: select a running monitor
set_fact: mon_host={{ item }}
with_items: "{{ groups[mon_group_name] }}"
when: item != inventory_hostname
- name: stop non-containerized ceph mon
service:
name: "ceph-mon@{{ ansible_hostname }}"
state: stopped
enabled: no
- set_fact:
ceph_uid: 64045
when: ceph_docker_image_tag | match("latest") or ceph_docker_image_tag | search("ubuntu")
- set_fact:
ceph_uid: 167
when: ceph_docker_image_tag | search("centos") or ceph_docker_image | search("rhceph") or ceph_docker_image_tag | search("fedora")
- name: set proper ownership on ceph directories
file:
path: "{{ item }}"
owner: "{{ ceph_uid }}"
recurse: yes
with_items:
- /var/lib/ceph
- /etc/ceph
- name: check for existing old leveldb file extension (ldb)
shell: stat /var/lib/ceph/mon/*/store.db/*.ldb
changed_when: false
failed_when: false
register: ldb_files
- name: rename leveldb extension from ldb to sst
shell: rename -v .ldb .sst /var/lib/ceph/mon/*/store.db/*.ldb
changed_when: false
failed_when: false
when: ldb_files.rc == 0
roles:
- ceph-defaults
- ceph-docker-common
- ceph-mon
post_tasks:
# We don't do a container test by running 'docker exec ...' since not all the monitors have switched to containers yet.
# Thus, we continue to use the 'ceph' binary from the host, there is no issue with that.
- name: non container | waiting for the monitor to join the quorum...
command: ceph --cluster "{{ cluster }}" -s --format json
register: ceph_health_raw
until: >
hostvars[mon_host]['ansible_hostname'] in (ceph_health_raw.stdout | from_json)["quorum_names"]
retries: "{{ health_mon_check_retries }}"
delay: "{{ health_mon_check_delay }}"
delegate_to: "{{ mon_host }}"
- name: switching from non-containerized to containerized ceph mgr
hosts:
- "{{ mgr_group_name|default('mgrs') }}"
vars:
containerized_deployment: true
mgr_group_name: mgrs
serial: 1
become: true
pre_tasks:
- name: stop non-containerized ceph mgr(s)
service:
name: "ceph-mgr@{{ ansible_hostname }}"
state: stopped
enabled: no
- set_fact:
ceph_uid: 64045
when: ceph_docker_image_tag | match("latest") or ceph_docker_image_tag | search("ubuntu")
- set_fact:
ceph_uid: 167
when: ceph_docker_image_tag | search("centos") or ceph_docker_image | search("rhceph") or ceph_docker_image_tag | search("fedora")
- name: set proper ownership on ceph directories
file:
path: "{{ item }}"
owner: "{{ ceph_uid }}"
recurse: yes
with_items:
- /var/lib/ceph
- /etc/ceph
roles:
- ceph-defaults
- ceph-docker-common
- ceph-mgr
- name: switching from non-containerized to containerized ceph osd
vars:
health_osd_check_retries: 5
health_osd_check_delay: 15
containerized_deployment: true
osd_group_name: osds
hosts:
- "{{ osd_group_name|default('osds') }}"
serial: 1
become: true
pre_tasks:
- name: collect running osds
shell: |
systemctl list-units | grep "loaded active" | grep -Eo 'ceph-osd@[0-9]{1,2}.service'
register: running_osds
changed_when: false
failed_when: false
- name: collect osd devices
shell: |
blkid | awk '/ceph data/ { sub ("1:", "", $1); print $1 }'
register: collect_devices
changed_when: false
when:
- devices is not defined or (devices is defined and devices == [])
- name: set devices
set_fact:
devices: "{{ collect_devices.stdout_lines | list }}"
when:
- collect_devices is defined
- not collect_devices.get("skipped")
- collect_devices != []
- name: stop non-containerized ceph osd(s)
systemd:
name: "{{ item }}"
state: stopped
enabled: no
masked: yes
with_items: "{{ running_osds.stdout_lines | default([])}}"
when: running_osds != []
- set_fact:
ceph_uid: 64045
when: ceph_docker_image_tag | match("latest") or ceph_docker_image_tag | search("ubuntu")
- set_fact:
ceph_uid: 167
when: ceph_docker_image_tag | search("centos") or ceph_docker_image | search("rhceph") or ceph_docker_image_tag | search("fedora")
- name: set proper ownership on ceph directories
file:
path: "{{ item }}"
owner: "{{ ceph_uid }}"
recurse: yes
with_items:
- /var/lib/ceph
- /etc/ceph
- name: check for existing old leveldb file extension (ldb)
shell: stat /var/lib/ceph/osd/*/current/omap/*.ldb
changed_when: false
failed_when: false
register: ldb_files
- name: rename leveldb extension from ldb to sst
shell: rename -v .ldb .sst /var/lib/ceph/osd/*/current/omap/*.ldb
changed_when: false
failed_when: false
when:
- ldb_files.rc == 0
- name: check if containerized osds are already running
shell: |
docker ps | grep -sq {{ item | regex_replace('/', '') }}
changed_when: false
failed_when: false
with_items: "{{ devices }}"
register: osd_running
- name: unmount all the osd directories
command: umount "{{ item.0 }}"1
changed_when: false
failed_when: false
with_together:
- "{{ devices }}"
- "{{ osd_running.results }}"
when:
- item.1.get("rc", 0) != 0
roles:
- ceph-defaults
- ceph-docker-common
- ceph-osd
post_tasks:
- name: get num_pgs
command: docker exec ceph-mon-{{ hostvars[groups[mon_group_name][0]]['ansible_hostname'] }} ceph --cluster "{{ cluster }}" -s --format json
register: ceph_pgs
delegate_to: "{{ groups[mon_group_name][0] }}"
- name: container - waiting for clean pgs...
command: "docker exec ceph-mon-{{ hostvars[groups[mon_group_name][0]]['ansible_hostname'] }} ceph --cluster {{ cluster }} -s --format json"
register: ceph_health_post
until: >
((ceph_health_post.stdout | from_json).pgmap.pgs_by_state | length) == 1
and
(ceph_health_post.stdout | from_json).pgmap.pgs_by_state.0.state_name == "active+clean"
delegate_to: "{{ groups[mon_group_name][0] }}"
retries: "{{ health_osd_check_retries }}"
delay: "{{ health_osd_check_delay }}"
when:
- (ceph_pgs.stdout | from_json).pgmap.num_pgs != 0
- name: switching from non-containerized to containerized ceph mds
hosts:
- "{{ mds_group_name|default('mdss') }}"
vars:
containerized_deployment: true
mds_group_name: mdss
serial: 1
become: true
pre_tasks:
- name: stop non-containerized ceph mds(s)
service:
name: "ceph-mds@{{ ansible_hostname }}"
state: stopped
enabled: no
- set_fact:
ceph_uid: 64045
when: ceph_docker_image_tag | match("latest") or ceph_docker_image_tag | search("ubuntu")
- set_fact:
ceph_uid: 167
when: ceph_docker_image_tag | search("centos") or ceph_docker_image | search("rhceph") or ceph_docker_image_tag | search("fedora")
- name: set proper ownership on ceph directories
file:
path: "{{ item }}"
owner: "{{ ceph_uid }}"
recurse: yes
with_items:
- /var/lib/ceph
- /etc/ceph
roles:
- ceph-defaults
- ceph-docker-common
- ceph-mds
- name: switching from non-containerized to containerized ceph rgw
hosts:
- "{{ rgw_group_name|default('rgws') }}"
vars:
containerized_deployment: true
rgw_group_name: rgws
serial: 1
become: true
pre_tasks:
- name: stop non-containerized ceph rgw(s)
service:
name: "ceph-radosgw@rgw.{{ ansible_hostname }}"
state: stopped
enabled: no
- set_fact:
ceph_uid: 64045
when: ceph_docker_image_tag | match("latest") or ceph_docker_image_tag | search("ubuntu")
- set_fact:
ceph_uid: 167
when: ceph_docker_image_tag | search("centos") or ceph_docker_image | search("rhceph") or ceph_docker_image_tag | search("fedora")
- name: set proper ownership on ceph directories
file:
path: "{{ item }}"
owner: "{{ ceph_uid }}"
recurse: yes
with_items:
- /var/lib/ceph
- /etc/ceph
roles:
- ceph-defaults
- ceph-docker-common
- ceph-rgw
- name: switching from non-containerized to containerized ceph rbd-mirror
hosts:
- "{{ rbdmirror_group_name|default('rbdmirrors') }}"
vars:
containerized_deployment: true
rbdmirror_group_name: rbdmirrors
serial: 1
become: true
pre_tasks:
- name: stop non-containerized ceph rbd mirror(s)
service:
name: "ceph-rbd-mirror@rbd-mirror.{{ ansible_hostname }}"
state: stopped
enabled: no
- set_fact:
ceph_uid: 64045
when: ceph_docker_image_tag | match("latest") or ceph_docker_image_tag | search("ubuntu")
- set_fact:
ceph_uid: 167
when: ceph_docker_image_tag | search("centos") or ceph_docker_image | search("rhceph") or ceph_docker_image_tag | search("fedora")
- name: set proper ownership on ceph directories
file:
path: "{{ item }}"
owner: "{{ ceph_uid }}"
recurse: yes
with_items:
- /var/lib/ceph
- /etc/ceph
roles:
- ceph-defaults
- ceph-docker-common
- ceph-rbd-mirror
- name: switching from non-containerized to containerized ceph nfs
hosts:
- "{{ nfs_group_name|default('nfss') }}"
vars:
containerized_deployment: true
nfs_group_name: nfss
serial: 1
become: true
pre_tasks:
- name: stop non-containerized ceph nfs(s)
service:
name: nfs-ganesha
state: stopped
enabled: no
- set_fact:
ceph_uid: 64045
when: ceph_docker_image_tag | match("latest") or ceph_docker_image_tag | search("ubuntu")
- set_fact:
ceph_uid: 167
when: ceph_docker_image_tag | search("centos") or ceph_docker_image | search("rhceph") or ceph_docker_image_tag | search("fedora")
- name: set proper ownership on ceph directories
file:
path: "{{ item }}"
owner: "{{ ceph_uid }}"
recurse: yes
with_items:
- /var/lib/ceph
- /etc/ceph
roles:
- ceph-defaults
- ceph-docker-common
- ceph-nfs