# Job Templates' variables & awx arc implementation

**URL:** <https://forum.ansible.com/t/job-templates-variables-awx-arc-implementation/25808>\
**Category:** AWX Project\
**Tags:** awx\
**Created:** [October 24, 2017, 10:58am UTC](https://forum.ansible.com/t/job-templates-variables-awx-arc-implementation/25808 "2017-10-24T10:58:41Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![nkb11193](https://avatars.discourse-cdn.com/v4/letter/n/e36b37/32.png) [@nkb11193](https://forum.ansible.com/u/nkb11193)\
**Post date:** [October 24, 2017, 10:58am UTC](https://forum.ansible.com/t/job-templates-variables-awx-arc-implementation/25808/1 "2017-10-24T10:58:41Z")

</div>

Hi there,

I’ve been testing awx for couple of days and I’m trying to integrate it inside my cloud environment (openstack), and I’m trying to use awx and ansible to provision my cloud’s resources, like creating an instance,  
and to run a deployment process, like installing java on the created instance. moreover, i would like it to be part of a single awx’s workflow. if assumed that creating an instance and installing java are separated ansible projects inside awx,  
is there a way to pass variables between job templates that part of the same workflow? can you suggest a better practices for this scenario?  
I would like my awx’s workflow to look like:

Creating instance → syncing cloud inventory → (passing the created instance inventory id) → installing java

Or, each template must have all it’s variable set prior the running process (like the instance’s hostname that would be created)?  
In addition, as awx already stores openstack credentials for syncing the dynamic inventory, is it possible to use them part of an ansible module like os\_server (openstack’s create\_instance ansible module)?

Last one, i still thinking about which implementation is better:  
creating a single site awx cluster and run jobs through ssh proxy for each tenant, or maybe it better to use “security isolated instance” feature for each tenant, and as a result run ansible right from the private network.  
I would like to hear some recommendation from your personal experience.

Regards,  
NB.

---

<div class="post-metadata">

**Author:** ![AlanCoding](https://sea1.discourse-cdn.com/ansible/user_avatar/forum.ansible.com/alancoding/32/301_2.png) [@AlanCoding](https://forum.ansible.com/u/AlanCoding)\
**Post date:** [October 24, 2017, 11:36am UTC](https://forum.ansible.com/t/job-templates-variables-awx-arc-implementation/25808/2 "2017-10-24T11:36:53Z")

</div>

You can pass variables from a workflow node to a downstream node by using set\_stats. When you use this, the variables you define will be shown in the “artifacts” field of the job that ran in the API. Those artifacts from the upstream job will then be passed to subsequent jobs as extra\_vars.

[https://docs.ansible.com/ansible-tower/latest/html/userguide/workflows.html#extra-variables](https://docs.ansible.com/ansible-tower/latest/html/userguide/workflows.html#extra-variables)

---

<div class="post-metadata">

**Author:** ![nkb11193](https://avatars.discourse-cdn.com/v4/letter/n/e36b37/32.png) [@nkb11193](https://forum.ansible.com/u/nkb11193)\
**Post date:** [October 24, 2017, 7:11pm UTC](https://forum.ansible.com/t/job-templates-variables-awx-arc-implementation/25808/3 "2017-10-24T19:11:21Z")

</div>

Thanks!  
NB

בתאריך יום שלישי, 24 באוקטובר 2017 בשעה 13:58:41 UTC+3, מאת [nkb1...@gmail.com](mailto:nkb1...@gmail.com):