Specify Host (inventory) in Python Script

Archives Ansible Project
1

Hi,

I’m trying to create a python script where I can specify the hosts within the script.

Something like:

webuptime = ansible.runner.Runner(
pattern=‘all’, forks=10,
module_name=‘command’, module_args=‘uptime’,
inventory = “webserver1,webserver2”
).run()

…or with a variable for the webserver bit. I can’t seem to figure it out though. Is there a simple way to do this without writing the names out to a file and reading them back in again? I can use an IP address instead if necessary.

Many thanks, Seumas

2

Why not use an executable inventory script? We include several examples of these in the ansible git repository (for example, the ec2.py script).

3

If you are doing inventory via the API you should use the Inventory class.

–Michael

4

Hi Michael,

Is there any documentation or examples of how to use that to do what I want?

Thanks, Seumas

5

Hi James,

I guess I could. I thought there must be an easy way of doing it directly though?

6

/usr/bin/ansible and /usr/bin/ansible-playbook both construct Inventory() objects.

7

Ok, it seems it’s not too complicated. You can do something like:

hosts = [“webserver1”,“webserver2”]

webInventory = ansible.inventory.Inventory(hosts)

webuptime = ansible.runner.Runner(
pattern=‘webserver*’, forks=10,
module_name=‘command’, module_args=‘uptime’,
inventory = webInventory
).run()

Hope this helps anyone looking for the same.

Cheers, Seumas

8

This is useful, but where do the host specific ansible_ssh_* values get passed? I’m not seeing how to do it in the dynamic inventory json style either.

9

They get passed as standard inventory variables

[webservers]
foo ansible_ssh_port=4000

or just

[webservers]
foo:4000

etc

In external inventory those variables are just like any other variable.