Support for maintenance mode profile configuration

Issue Type

• Template enhancement
• Role enhancement
• Documentation enhancement (avd.sh)

Is your feature request related to a problem? Please describe.

We have to adjust the maintenance mode profile configuration on all L3/MLAG-Leaf switches to make maintenance mode work in our environment.

Describe the solution you’d like

We have to configure a maintenance profile interface with certain values like the following one and make it the default.

maintenance
   profile interface int_maint default
   !
   profile interface int_maint
      rate-monitoring load-interval 100
      rate-monitoring threshold 1000
      shutdown max-delay 100

There is currently no template available within the eos_cli_config_gen role. The data model could perhaps look like this:

maintenance:
  profile:
    interface:
      < profile_interface_1 >:
        rate_monitoring:
          load_interval: < seconds >
          threshold: < kbps >
        shutdown:
          max_delay: < seconds >     
        default: < true | false >

Describe alternatives you’ve considered

The current alternative is, to create a config_let within CVP and bind it to the containers of our leafs manually within CVP or “automated” by adding the config_let via AVD. Our goal is that Ansible code is the single source of configuration.

Additional context

Add any other context or screenshots about the feature request here.

[daanvdsanden]

I would like to suggest a different data model:

1. move the default out of the profiles (no need to loop all the profiles to get the default, also what behavior do you want when two profiles have set the default to true).
2. flatten the profiles to have it match the config hierarchy, this also makes the plural form a bit more logical

This makes the model match the cli config hierarchy better and less logic is needed (which can be embedded in the design role).

maintenance:
  default_interface_profile: <interface_profile_1>
  default_bgp_profile: < bgp_profile_1 >
  default_unit_profile: < unit_profile_1 >
  interface_profiles:
    < interface_profile_1 >:
      rate_monitoring:
        load_interval: < seconds >
        threshold: < kbps >
      shutdown:
        max_delay: < seconds >     
  bgp_profiles:
    < bgp_profile_1 >:
      initiator:
        route_map_inout: < route_map >
  unit_profiles:
    < unit_profile_1 >:
      on_boot:
        duration: < 300-3600 >

EDIT: changed unit and bgp profiles to be more consistent with the rest of the template.

[daanvdsanden]

I like the fact that the generated config is an exact match of the running config. So when I started I also used the “lazy” method and I thought I could flatten the array the same way as is done with vlans on a trunk port. That was impossible for me since to do this properly I need to know for what kind of device I’m creating the config for. That was the point I was trying to make with the Ethernet2/2 example. It can be two different things which can impact the result when the array is flattened.

E.G. the following happens on a 7050CX switch with QSFPs:

EOS(config-group-if-TEST)#interface ethernet 1/1-4
EOS(config-group-if-TEST)#interface ethernet 2/1-4
EOS(config-group-if-TEST)#show active
group interface TEST
   interface Et1/1-2/4
   exit

However when the same commands are given on a arista with linecard where SFP(+) modules inserted in slot 1 and 2, I (assume) it will not get flatten the same, I expect this to happen (assumption, I don’t have a 7500 to test this):

EOS(config-group-if-TEST)#interface ethernet 1/1-4
EOS(config-group-if-TEST)#interface ethernet 2/1-4
EOS(config-group-if-TEST)#show active
group interface TEST
   interface Et1/1-4,2/1-4
   exit