Reducing duplication of pillar data

[hashi]

Hey all,

Just want to know what everyone is doing around reducing duplication of pillar data for pillars that need to share values? is there any consensus on "best practice"?. So far the only thing I've seen that helps is by creating jinja files and using load_yaml then referencing it in pillar.

Im really not sure how to address such things for example:

2 pillars using same rabbitmq host

/srv/pillar/openstack/cinder.sls

cinder:

  rabbitmq:

    host: 192.168.1.1

    user: admin

    password: blah

    

/srv/pillar/openstack/nova.sls

nova:

  rabbitmq:

    host: 192.168.1.1

    user: admin

    password: blah

[Sylvain Viart]

Hi,

Here what I've found:

Le 19/08/2016 à 11:52, hashi a écrit :

Just want to know what everyone is doing around reducing duplication of pillar data for pillars that need to share values? is there any consensus on "best practice"?. So far the only thing I've seen that helps is by creating jinja files and using load_yaml then referencing it in pillar.

Same approach for me:

{% import_yaml "/srv/salt/pillar/auto/managed_password.yaml" as pass with context %}
mysql:
  user:
    user1:
      password: "{{ pass['user1']['mysql'] }}"

So I decided to generate pillar…
https://github.com/opensource-expert/customers-formula/

2 pillars using same rabbitmq host

/srv/pillar/openstack/cinder.sls

cinder:

  rabbitmq:

    host: 192.168.1.1

    user: admin

    password: blah

    

/srv/pillar/openstack/nova.sls

nova:

  rabbitmq:

    host: 192.168.1.1

    user: admin

    password: blah

or you could loop or use macro in jinja, if you did in some formula, there is many loop in state.

{% for h in salt['pillar.get']('mykey:rabbitmq', {}).itemps() %}

/srv/pillar/openstack/{{ h }}.sls

{{ h }}:

  rabbitmq:

    host: 192.168.1.1

    user: admin

    password: blah

{% endfor %}

Regards,
Sylvain.

-- 
Sylvain Viart - Consultant informatique système linux