How to specify different hosts for different playbooks in one ansible script

ansible

As I understand, each playbook takes one hosts entry. I want to know that if I create a container playbook that includes other playbooks, can I parameterize the hosts for each playbook include. So something like

---
- include playbook_1.yml
    hosts: tag_postgres
- include: playbook_2.yml
    hosts: tag_rabbitmq

I am able to put all different playbooks in one script and call, but then this way I am not able to reuse some set_fact from one playbook into another and hence there is a lot of task duplication.

Another corollary of the question is, can I launch ec2s on amazon which would have

hosts: localhost

and configuration of the launched ec2s, which would basically configure inventory from -i ec2.py, and have hosts specified as

hosts: tag_<some_tag>

happening through the same playbook or a set of included playbooks (different roles)?

Best Answer

Actually, you can have more than one hosts: section per playbook. It appears that a hosts: starts a new play. See http://www.tecmint.com/use-ansible-playbooks-to-automate-complex-tasks-on-multiple-linux-servers/, for example.

Something like this works for me (ansible 2.2):

---
- hosts: localhost 
  connection: local
  roles:
    - { role: ec2,
        tag: 'master',
        instance_type: t2.2xlarge,
        count: 1
      }
  tasks:
  - shell: hostname # reports localhost

- hosts: tag_master
  tasks:
  - shell: hostname # reports instance(s) with tag 'master'

So, put hosts: at the top of each included .yml, not after the include:.

Related Solutions

Optimizing Ansible playbooks to run against many hosts

First of all, you need to consider caching the facts.

Take a look here for how to:

http://docs.ansible.com/ansible/playbooks_variables.html#fact-caching

You will see an amazing performance on gather-facts, even with caching to a file.

Then you may consider of improving the level of parallelism with -f

man ansible-playbook

   -f NUM, --forks=NUM
       Level of parallelism.  NUM is specified as an integer, the default is 5.

to something bigger than 5

How to run all Ansible plays at first host (all of them), then at second host (all of them) and so on – hosts one by one

Q "All plays should be run at a first host and finish. All plays should be run at a second host and finish. And so on at arbitrary quantity of hosts."

A: It's not possible. The scope of serial and strategy is a play. All hosts shall complete a play before the next one could be started. An import of a playbook is equivalent to starting a new play.

Q: "If a host is not from a group for a playbook, play should not be applied at this host."

A: It's possible to create a new group in the first play and use it by the imported playbooks. For example the inventory

shell> cat hosts
[test]
test_01
test_02
test_03

[satellites]
test_03

and the playbook

shell> cat play-do-11.yml
- hosts: my_dynamic_group
  tasks:
    - debug:
        var: inventory_hostname

imported from the playbook

shell> cat site.yml
- hosts: test
  tasks:
    - add_host:
        name: "{{ item }}"
        groups: my_dynamic_group
      loop: "{{ groups.test }}"
      when: item in groups.satellites
      run_once: true
    - debug:
        var: inventory_hostname

- import_playbook: play-do-11.yml

gives

shell> ansible-playbook site.yml

PLAY [test] ***

TASK [add_host] ***
skipping: [test_01] => (item=test_01) 
skipping: [test_01] => (item=test_02) 
changed: [test_01] => (item=test_03)

TASK [debug] ***
ok: [test_01] => {
    "inventory_hostname": "test_01"
}
ok: [test_03] => {
    "inventory_hostname": "test_03"
}
ok: [test_02] => {
    "inventory_hostname": "test_02"
}

PLAY [my_dynamic_group] ***

TASK [debug] ***
ok: [test_03] => {
    "inventory_hostname": "test_03"
}

PLAY RECAP ***

Best Answer

Related Solutions

Optimizing Ansible playbooks to run against many hosts

How to run all Ansible plays at first host (all of them), then at second host (all of them) and so on – hosts one by one

Related Topic