Feat(eos_cli_config_gen,eos_designs): Add support for 'l2 mru'

[m-rhode]

Change Summary

add support for setting the "l2 mru" on interfaces and port-channels

Related Issue(s)

none

Component(s) name

arista.avd.eos_cli_config_gen

Proposed changes

I duplicated all "l2_mtu" assets and replaced "mtu" with mru"

How to test

did a config generation and deployment with the branch:

r2-lab2-de#show int eth14 | grep MRU
  IP MTU 1500 bytes (default), Ethernet MRU 10240 bytes, BW 10000000 kbit

ansible run:

ok: [r2-lab2-de] => 
  msg:
  - |-
    --- system:/running-config
    +++ session:/1481195311-session-config
    interface Ethernet14
    +   l2 mru 9000

r2-lab2-de#show int eth14 | grep MRU
  IP MTU 1500 bytes (default), Ethernet MRU 9000 bytes, BW 10000000 kbit

Checklist

User Checklist

ansible_collections/arista/avd/roles/eos_cli_config_gen](support-l2-mru) $ grep -r "l2_mtu\|l2_mru"

then eyeballed if every l2_mtu has a l2_mru equivalent (actually good that I did :) )

Repository Checklist

• My code has been rebased from devel before I start
• I have read the CONTRIBUTING document.
• My change requires a change to the documentation and documentation have been updated accordingly.
• I have updated molecule CI testing accordingly. (check the box if not applicable)

[carlbuchmann]

@m-rhode thank you for this contribution, I converted it back to draft as you also need to include molecule unit tests for eos_designs and eos_cli_config_gen.
I see you are a first-time contributor to the project, and can assist on a zoom if required. Please contact me @ carl.buchmann@arista.com.

[carlbuchmann]

Also, you need to generate documentation. This is done by running pre-commit locally. Install pre-commit and then run:
pre-commit run shemas --all

[m-rhode]

@m-rhode thank you for this contribution, I converted it back to draft as you also need to include molecule unit tests for eos_designs and eos_cli_config_gen. I see you are a first-time contributor to the project, and can assist on a zoom if required. Please contact me @ carl.buchmann@arista.com.

Hello @carlbuchmann, thanks for the zoom offer, much appreciated!
Let's first try without. I assume it's basically reproducing the l2_mtu config under "ansible_collections/arista/avd/molecule" for the molecule stuff. And as this is not tested on real-life hardware I could simply use the same ports as they are used for the l2_mtu - I suppose?

Is there anything to do in "ansible_collections/arista/avd/roles/eos_designs"... as l2_mru` is not part of a design I suppose not?

Also, you need to generate documentation. This is done by running pre-commit locally. Install pre-commit and then run:
pre-commit run shemas --all

thanks for pointing this out...

[gmuloc]

Hello @m-rhode

Answering some of the questions here as @carlbuchmann is in another timezone

So in general we try to avoid introducing some new knobs in eos_cli_config_gen and eos_designs in one go to make things clearer. Usually, we just recommend one PR to add the low-level CLI in eos_cli_config_gen and then a new one on top to do the eos_designs bit. Thus I will structure a list of tasks based on this assumed separation and tick the steps you have followed already.

Make sure to install the requirements-dev.txt to be able to complete the process (I believe you have already but mentioning it just in case)

eos_cli_config_gen

This part of the change is aimed at adding some new knobs. The following tasks are required:

• Adding the knob to the appropriate schema fragments file in the roles/eos_cli_config_gen/schema/schema_fragments directory (or create a new file if adding a top-level feature)
• Editing the Jinja2 templates in roles/eos_cli_config_gen/templates/eos and roles/eos_cli_config_gen/templates/documentation (or adding new template if adding a top-level feature, in that case, you should also modify the roles/eos_cli_config_gen/templates/eos-device-documentation.j2 and ``roles/eos_cli_config_gen/templates/eos-intended-config.j2 to add these new templates, trying to respect the order in the EOS CLI)
  • NOTE: This one is partially done in your PR as you have modified only the ethernet-interface.j2 template for config and not the port-channel.j2 one where you have added the l2mru knob in the schema. Also for now you have not added anything to the documentation templates (it may or may not be necessary)
• Run pre-commit run shemas --all to re-generate the eos_cli_config_gen schema with your modifications (you need to do this every time you change the schema, even if this is just a description changed). This is also used to add the new options to our documentation.
• Add some molecule test in n the molecule/eos_cli_config_gen scenario exercising your new knob configuration
  • NOTE: In your case you would need to add to both molecule/eos_cli_config_gen/inventory/host_vars/ethernet-interfaces.yml and molecule/eos_cli_config_gen/inventory/host_vars/port-channel-interfaces.yml (re-using existing interfaces is fine as this is only for testing purposes)
• Run the molecule test locally to generate the new expected configuration and documentation for these two hosts (we recommend running the full scenario using from ansible_collections/arista/avd/molecule folder the command make converge MOLECULE=eos_cli_config_gen
• Run all pre-commit checks pre-commit run --all to make sure that all files you have added are correct and won't be rejected by CI (this command will take ~1min)
• Review that everything added looks correct and push to Github!

eos_designs

Is there anything to do in "ansible_collections/arista/avd/roles/eos_designs"... as l2_mru` is not part of a design I suppose not?

I just would like to point out that if you don't need the l2_mru in eos_designs you should remove the changes you have made to the python code there. Otherwise, there is a bit more work to do indeed. This can be done only after the eos_cli_config_gen part has been completed (hence why we usually recommend 2 PRs)

The process is very similar (for simple changes like yours) except for the jinja2 template part where you need to edit some Python code

• Adding the knob to the appropriate schema fragments file in the roles/eos_designs/schema/schema_fragments directory
  • NOTE: For you it would be eos_designs/schemas/schema_fragments/defs_adapter_config.schema.yml - something like https://github.com/aristanetworks/ansible-avd/blob/devel/ansible_collections/arista/avd/roles/eos_designs/schemas/schema_fragments/defs_adapter_config.schema.yml#L44 .
• Editing the relevant python_modules
  • NOTE: This has been done for you but you now need to add the knob to the eos_designs schema for adapters or it will never be possible to use it
• Run pre-commit run shemas --all to re-generate the eos_designs schema with your modifications (you need to do this every time you change the schema, even if this is just a description changed). This is also used to add the new options to our documentation.
• Add some molecule test in n the molecule/eos_designs_unit_tests scenario exercising your new knob configuration
  • NOTE: In your case you would need to add the l2mru knob to some adapters, one for port-channel and one for ethertnet-interfaces. It may be simpler in your case to add to the molecule/eos_designs_unit_tests/inventory/host_vars/connected_endpoints.yml file
• Run the molecule test locally to generate the new expected configuration and documentation for this two hosts (we recommend running the full scenario using from ansible_collections/arista/avd/molecule folder the command make converge MOLECULE=eos_designs_unit_tests
• Run all pre-commit checks pre-commit run --all to make sure that all files you have added are correct and won't be rejected by CI (this command will take ~1min)
• Review that everything added looks correct and push to Github!

Let us know if you need further help

approved by mistake

[m-rhode]

Dear @carlbuchmann and @gmuloc ,
thanks a lot for your explanations, really appreciated. I was on PTO and today is my first day back. I'd have taken care of working further on this but as it seems I was to slow. Sorry for that and thanks for implementing this!

[github-actions]

This pull request has conflicts, please resolve those before we can evaluate the pull request.

[github-actions]

Conflicts have been resolved. A maintainer will review the pull request shortly.

[ClausHolbechArista]

@carlbuchmann @m-rhode I have adjusted the schema a bit for min/max. For l2_mtu the min is 68 on all the device types I could find. This was set to 64, so please let me know if you have a platform saying 64.

[m-rhode]

@ClausHolbechArista thanks for working on this. Min 68 is fine for us.

[gmuloc]

LGTM