[ansible-project] Variable Structure/Precedence

[Johnathan Komara]

I am having trouble wrapping my head around how i should setup my variable
structure. I have two locations with multiple environments at each. I can
have different variables depending on the location and production env.

For example, I have a dev web server at location A and a prod web server at
location B. I want them to each have a separate virtual host. What is the
best way to accomplish this?

I tried putting variables in the group_vars directory with a separate
directory for each location, then subdirectories for each environment, but
then I each location is going to pull in every variable from all
environments.

I tried just using group_vars (all, webserver, etc) but I couldn't get this
to work either unless I have a group_var file for every group at every
location for each environment.

The only other thing that I can think of to do to accomplish this is to put
the defaults for all groups in group_vars. In my inventory file for each
location or environment I can assign a group_var (i.e. env_type=dev). Then
in my playbook I can do:

var_files:
- /var/{{ location }}/{{ env_type }}/group_name.yml

I want to avoid duplicating variables that are the same for each env and
make the roles as dynamic as possible. I've read through the docs and can't
seem to make sense of it. I am sure that Ansible or someone else has
figured out. I just cant find it. If someone can point me in the right
direction I would really appreciate it.

Thanks.

[Johnathan Komara]

I was trying to avoid a group_var file for everything but it looks like
that is the route that I am going to have to go. I have set the
parent-child relationship of the groups in my inventory files. So in my
group_vars directory I have location_a, location_b, dev, prod, qa,
location_a_dev, location_b_dev, location_a_web, location_b_web, etc. This
accomplishes what I want but I feel like it is going to get out of control
very quickly. If there is a better way to accomplish this please let me
know.

Post by Johnathan Komara
I am having trouble wrapping my head around how i should setup my variable
structure. I have two locations with multiple environments at each. I can
have different variables depending on the location and production env.
For example, I have a dev web server at location A and a prod web server
at location B. I want them to each have a separate virtual host. What is
the best way to accomplish this?
I tried putting variables in the group_vars directory with a separate
directory for each location, then subdirectories for each environment, but
then I each location is going to pull in every variable from all
environments.
I tried just using group_vars (all, webserver, etc) but I couldn't get
this to work either unless I have a group_var file for every group at every
location for each environment.
The only other thing that I can think of to do to accomplish this is to
put the defaults for all groups in group_vars. In my inventory file for
each location or environment I can assign a group_var (i.e. env_type=dev).
- /var/{{ location }}/{{ env_type }}/group_name.yml
I want to avoid duplicating variables that are the same for each env and
make the roles as dynamic as possible. I've read through the docs and can't
seem to make sense of it. I am sure that Ansible or someone else has
figured out. I just cant find it. If someone can point me in the right
direction I would really appreciate it.
Thanks.

[koka424]

So I figured out a setup that works for what I am trying to do. I was
trying to set a hiera structure so that I could set global variables at the
top and have more specific groups overwrite these variables. I ran into a
lot of issues trying to set this up for multiple environments that span
multiple datacenters.

I have an inv file for each environment (dev, test, prod). In my ansible
home directory I have a dir called 'env'. Inside there I have a dir for
each environment. In each environment I have a playbooks/group_vars
directory. So for each environment I have:

/etc/ansible/dev <- Inventory
/etc/ansible/group_vars <-global group_vars
/etc/ansible/env/dev/playbooks <- Environment specific playbooks
/etc/ansible/env/dev/playbooks/group_vars <- Environment specific variables
since group vars in the playbooks directory overwrite the inv group_vars

When I run ansible I run:

ansible-playbook -i dev env/dev/playbooks/web-server.yml

I still have to specify each location. It works out though. I create a main
group for example [web-server] then i set its children to web-server-loca
and web-server-locb.

#Location

[loca:children]
web-server-loca

[locb:children]
web-server-locb

[web-server:children]
web-server-loca
web-server-locb

This way site specific vars overwrite role vars, and role-location vars
overwrite location specific vars.

I have hash merging enabled as well. This setup allows me to set something
in /etc/ansible/group_vars and have the more specific environment variables
(/etc/ansible/env/dev/playbooks/group_vars) overwrite the global vars, or
if needed merged via hash.

I hope this helps someone else.

Post by Johnathan Komara
I was trying to avoid a group_var file for everything but it looks like
that is the route that I am going to have to go. I have set the
parent-child relationship of the groups in my inventory files. So in my
group_vars directory I have location_a, location_b, dev, prod, qa,
location_a_dev, location_b_dev, location_a_web, location_b_web, etc. This
accomplishes what I want but I feel like it is going to get out of control
very quickly. If there is a better way to accomplish this please let me
know.

Post by Johnathan Komara
I am having trouble wrapping my head around how i should setup my
variable structure. I have two locations with multiple environments at
each. I can have different variables depending on the location and
production env.
For example, I have a dev web server at location A and a prod web server
at location B. I want them to each have a separate virtual host. What is
the best way to accomplish this?
I tried putting variables in the group_vars directory with a separate
directory for each location, then subdirectories for each environment, but
then I each location is going to pull in every variable from all
environments.
I tried just using group_vars (all, webserver, etc) but I couldn't get
this to work either unless I have a group_var file for every group at every
location for each environment.
The only other thing that I can think of to do to accomplish this is to
put the defaults for all groups in group_vars. In my inventory file for
each location or environment I can assign a group_var (i.e. env_type=dev).
- /var/{{ location }}/{{ env_type }}/group_name.yml
I want to avoid duplicating variables that are the same for each env and
make the roles as dynamic as possible. I've read through the docs and can't
seem to make sense of it. I am sure that Ansible or someone else has
figured out. I just cant find it. If someone can point me in the right
direction I would really appreciate it.
Thanks.