Story #5762: [Epic] CI Improvements
ansible-pulp molecule CI seems slower than it should be
Ticket moved to GitHub: "pulp/pulpcore/1857":https://github.com/pulp/pulpcore/issues/1857
ansible-pulp CI is currently taking like 30 minutes.
Some theories on why this is happening:
1. ansible pipelining is not working. So there is a large overhead per every ansible task.
2. The Travis VM is low on RAM as it tests 3 containers at once (C7, Fedora & Debian).
3. Although purely an improvement, we could implement libeatmydata/nosync, or at other layers (like docker) make all disk writes unsafely cached.