This is hack around the infinite dangling ansible-playbook processes which happen against the obsolescent fedorainfracloud.org now (copr vm manager thinks that the stucked processes do something, and respects the max_spawn_processes -- so in other words, nothing is spawned).

We should either implement some timeout in code, and drop the hack - or make the hack reproducible after re-deploy, so either:
- implement some configurable timeout in code
- provide /usr/bin/copr-ansible-playbook-timeout script, and use that by our code, or
- provide config option for the ansible-playbook script, and configure the rest by ansible