Greetings -
I am sure this question has been asked and answered many times but I could not find anything that specifically addressed my questions via google search.
I have modified a spec file for a python based application (weewx specifically although that is not too relevant) to target Fedora specifically. The upstream developers have a spec file that covers el8 and fedora both. The upstream version uses init.d based init scripts to manage the application daemon.
I modified the spec file to target fedora specifically using systemd to manage the daemon. This is working well for me on test VMs.
I would also like to be able to install the rpm on a fedora container image. I can do that with either the upstream or my current spec file, of course, but I am thinking that systemd, in this particular scenario, is overkill. The application daemon would be managed at the container level not within the container per se.
What would be a good way to structure the spec file to cover both scenarios? With systemd for install on a fedora VM or fedora bare metal machine and without systemd for a fedora container install?
Would using subpackages be appropriate? One rpm for the base install and the second to add systemd integration? Or is there another alternative? A pointer to relevant discussions or possible solution would be most helpful and appreciated.
Many thanks
regards
Brad