-
|
I have a tenant with a vrf. That vrf has SVI's over 6 nodes. But now I see that those are applied to all nodes, and not only to the those with the BGP peering.. Is there a cleaner way to do this ? I don't like that orphaned config, as it looks like something was configured incorrect |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments 4 replies
-
|
For now the only way would be to use some inline-jinja to check if you are on one of the 6 nodes. Written from memory, to may need some adjustment. Refer to Ansible docs for guidance. structured_config: |-
{% if inventory_hostname in ['node1', 'node2'] %}
route_maps: ...
{% elif %}
null
{% endif %}We have implemented We would like to implement similar "catalogs" of route-maps and other areas which will be applied when referred to under the BGP peers, peer-groups etc. So the route-maps and prefix-lists them selves would also get proper models in eos_designs. |
Beta Was this translation helpful? Give feedback.
-
|
I wrote a 'conditional catalog' to kind of get around it.. any design variable can be appended to _catalog_
then I added a custom_structured_configuration_prefix and an ansible set_fact now any variable starting with catalog will create a generated_ variable, and only if any of the apply_group is part of the ansible groups, or if the inventory_hostname matches the hostname Works pretty good :-) of course, I still have to manually decide what prefix list, route map, access-list goes on what devices, but at least they are not lingering on devices they are not needed on |
Beta Was this translation helpful? Give feedback.
I wrote a 'conditional catalog' to kind of get around it..
There is a catalog created like this
any design variable can be appended to _catalog_
then I added a custom_structured_configuration_prefix
generated_