AVD Data Model for EVPN/VXLAN Deployment Guide - Domain A #61
Conversation
There was a problem hiding this comment.
Make file must be in the lab directory root as it must be available from the container workspace. Also Makefile for every lab will have a standard and templated layout in future shared by other labs. I'd suggest to keep existing Makefile for now and add AVD shortcuts there.
There was a problem hiding this comment.
Everything in avd/config_backup must be gitignored. Example here: https://github.com/arista-netdevops-community/one-click-se-demos/blob/6bb2b2aabc02a98837af27ebcc3e4b4f1509111a/demos/cvaas-cvaas-and-avd-demo--evpn-mlag/.gitignore#L113
There was a problem hiding this comment.
I'd suggest to avoid generating AVD docs and configs in advance. That can be done when lab starts and keeps inventory compact. However it's just an opinion and adding files in advance has some minor advantages. For ex. it's easier to inspect inventory with running the lab.
|
|
||
| ################################################### | ||
| # Customized Structured Config # | ||
| ################################################### |
There was a problem hiding this comment.
Custom structured section is quite long. Are we planning to advertise that as best practice for AVD data model? If yes, we have to sync with AVD maintainers team and find a better way to implement missing features. Otherwise we can consider dropping some non-essential configuration from custom structured section.
| ansible_httpapi_use_ssl: true | ||
| ansible_httpapi_validate_certs: false | ||
| ansible_python_interpreter: $(which python3) | ||
| avd_data_validation_mode: warning |
There was a problem hiding this comment.
I'd suggest changing avd_data_validation_mode to error.
There was a problem hiding this comment.
Do not add any ANTA generated files to the lab inventory by default. Especially result jsons.
There was a problem hiding this comment.
Same as above - we should have one single Makefile in the container workspace.
There was a problem hiding this comment.
Why do we need duplicate files in tech-library/datacenter if we already have the in /labs ?
There was a problem hiding this comment.
Is this the final structure of the inventory.yml ? Where are the management addresses?
Change Summary
techlib-vxlan-domain-acontainer to uselab-base:python3.11-avd-v5.0.0-clab0.57.5-rev1.1avddirectory, containing all validated AVD data models fordomain-ato thelabs/techlib-vxlan-domain-adirectorymake build,make deploy, andmake validatein the Codespace lab. All worked as intended.Repository Checklist
mainbranch