Hi everyone! I'm still pretty new to the whole SD-WAN and network automation world, so please bear with me if this seems basic. I've been trying to wrap my head around how to actually *manage* a Versa deployment at scale without clicking through the Director GUI for every single branch.
I stumbled on something in the documentation that was a real lightbulb moment for me, and I wanted to share it and also ask for some advice. Apparently, Versa exposes YANG data models for their CPE and Director. This means you can use tools like Ansible to interact with them programmatically.
I just finished a small test where I used the `ansible.netcommon.netconf_config` module to push a simple change—just adding a new local user—to a branch device. It felt like magic compared to doing it manually. The idea that I could, in theory, write a playbook to stand up a whole new branch by scripting the configuration against these models is incredibly exciting.
For those of you who have been doing this longer, is this the standard way you automate Versa? Are there any big "gotchas" or best practices I should know about before I try to script something bigger, like a full site deployment? Also, are the YANG models for all Versa features pretty well documented and stable?
Any guidance would be hugely appreciated. I'm just starting to see how powerful automation can be, and it's a game-changer for someone like me who's managing a growing number of sites. 😅
Oh, that's an awesome starting point. Pushing that local user change and feeling that "magic" is exactly how the addiction to automation starts, haha.
Your instinct is spot on - using Ansible with the NETCONF modules against those YANG models is absolutely a standard method for Versa automation, especially for branch rollouts. The big thing you'll want to do next is get familiar with the exact YANG models for the Director itself, not just the CPE. The Director API is how you orchestrate the whole template-and-device attachment flow.
One gotcha I ran into early on: the state of the config on the Director vs. the CPE. Make sure your playbooks account for the commit-and-push cycle from the Director down to the branch. If you configure a device object directly, it won't hit the live box until that's done.
Have you looked at using `ansible.utils.netconf_rpc` for the operational ("get") commands? It's super useful for pre-change validation.
Automate all the things.