Pulp

Molecule: Enable Ansible pipelining via group_vars

Variables are currently sufficient for enabling pipelining in molecule because the role we tried (gzm55.smart_ssh_pipelining) did not need to modify system configuration for all 3 containers; it only set the variable.

Fixes: #5817 ansible-pulp CI is slower than it should be due to a lack of ansible pipelining [noissue]

Molecule: Enable Ansible pipelining via group_vars

Variables are currently sufficient for enabling pipelining in molecule because the role we tried (gzm55.smart_ssh_pipelining) did not need to modify system configuration for all 3 containers; it only set the variable.

Fixes: #5817 ansible-pulp CI is slower than it should be due to a lack of ansible pipelining [noissue]

Problem: provisioning is slower than it could be

Solution: Enable ansible pipelining. It's safe to do on all our current boxes. If it ever isn't safe to do for 1, we can use this role to conditionally enable it: https://galaxy.ansible.com/gzm55/smart_ssh_pipelining

On my quick test (file module loop to touch /tmp/[1-20], it reduced the time per loop item from .41s .21s. Tasks without a loop receive a similar reduction in overhead.

re: #5817 ansible-pulp CI is slower than it should be due to a lack of ansible pipelining https://pulp.plan.io/issues/5817