Ansible can handle variables in different ways. The easiest way is to pass variables on command line. This will always overwrite pre-defines.
Two examples:
create and start virtual machine with file injection (inject up to 5 max 1k base64 encoded files):
ansible-playbook ecs_create.yml -e "ecs_fileinject_1=/etc/hosts ecs_fileinject_data_1=$(base64 -w 0 hosts.txt) ecs_fileinject_2=/root/README.md2 ecs_fileinject_data_2=$(base64 -w 0 hallo.txt)" --vault-password-file vaultpass.txt
create and start virtual machine with injection user_data (inject max 32k base64 encoded user-data files):
ansible-playbook ecs_create.yml -e "ecs_user_data=$(base64 -w 0 user-data.txt)" --vault-password-file vaultpass.txt
You can also put the variables in an extra file and pass them on command line:
ansible-playbook ecs_create.yml -e "@my_ecs_data.yml"
More options are described on Ansible Lookups
A human readable form of configuration files in key=value format. The INI file is divided into section. A section is a definition of an ECS instance, or EVS/ELB, or DNS zone. A [default] section on topic is valid for each other section where values are overwritten. Example is in playbooks/vars/tenant.ini
YAML are also human readable. The format is key:value, but the value can also include another key, so you can group your items like:
default-default_of_whatever
vms-vm1
-vm2
-vm3
evs-evs1
-evs2
-evs3
elb-elb1
-elb2
-elb3
Example is on playbooks/vars/tenant.yml
JSON Schema can be a very powerful tool for configuration handling. Basically it’s also a key:value store like YAML or INI, but in JSON you can define your own schema. This includes the definition of field types, like strings or values. It would be useful if you know the schema definition of OTC API. Then you could copy the schema one to one. Or you start with your own definition which should be well considered. A basic example of using JSON is on playbooks/vars/tenant.json