EC2: Change logic of deleting resources

[klaphi]

This PR changes the logic of deleting resources of the EC2 driver. Without that change security groups and keypairs will stay because the tasks will be skipped.

The tasks Write Molecule instance configs changes the when condition of the block immediately. This will skip the following tasks regardless of their when conditions. Moving the tasks Write Molecule instance configs to the end of the block will execute the cleanup of Security groups and keypairs according to their own when condition.

[klaphi]

This will fix: #280

[apatard]

While the change looks fine (the cleanup should always be done in the end imho), some questions for you:

• I've given a brief look and I'm having hard time to understand how it's solving your issue. Can you please give more details ?
• Which version of molecule are you using ? iirc, the cookiecutter stuff is not used anymore theses days.

[bparry02]

I was running into this issue as well (ephemeral security groups not being deleted during destroy), and this change resolved the issue. Thank you @klaphi !

I've given a brief look and I'm having hard time to understand how it's solving your issue. Can you please give more details ?

I assume there is some lazy loading of the platform variable going on, and when the backing file is updated it affects usage of that variable in later code.

Which version of molecule are you using ? iirc, the cookiecutter stuff is not used anymore theses days.

The scenario init with ec2 driver did not provide a working configuration for me, so I had to manually copy down create.yml and destroy.yml from GitHub. Ansible and molecule versions are below:

ansible                   8.7.0
ansible-compat            24.10.0
ansible-core              2.15.13
ansible-lint              24.12.2
molecule                  24.12.0
molecule-plugins          23.6.0

$ molecule --version
molecule 24.12.0 using python 3.12 
    ansible:2.15.13
    gce:23.6.0 from molecule_plugins requiring collections: google.cloud>=1.0.2 community.crypto>=1.8.0
    podman:23.6.0 from molecule_plugins requiring collections: containers.podman>=1.7.0 ansible.posix>=1.3.0
    ec2:23.6.0 from molecule_plugins
    default:24.12.0 from molecule
    vagrant:23.6.0 from molecule_plugins
    docker:23.6.0 from molecule_plugins requiring collections: community.docker>=3.4.11 ansible.posix>=1.4.0
    containers:23.6.0 from molecule_plugins requiring collections: ansible.posix>=1.3.0 community.docker>=1.9.1 containers.podman>=1.8.1
    openstack:23.6.0 from molecule_plugins requiring collections: openstack.cloud>=2.1.0
    azure:23.6.0 from molecule_plugins

[apatard]

I was running into this issue as well (ephemeral security groups not being deleted during destroy), and this change resolved the issue. Thank you @klaphi !

Thanks for confirming it's working for someone else. It's helping me making a decision :)

I've given a brief look and I'm having hard time to understand how it's solving your issue. Can you please give more details ?

I assume there is some lazy loading of the platform variable going on, and when the backing file is updated it affects usage of that variable in later code.

hm. I see...

Which version of molecule are you using ? iirc, the cookiecutter stuff is not used anymore theses days.

The scenario init with ec2 driver did not provide a working configuration for me, so I had to manually copy down create.yml and destroy.yml from GitHub. Ansible and molecule versions are below:

Ok, so the files were manually copied. That's confirming what I've suspected.

[apatard]

Not sure how it's working but someone else confirmed the issue and that the fix is solving it, so I'm approving