Description of problem: If users wish to install our Celery packages on a system that does not have httpd installed, the apache user will not exist. However, the Celery packages all assume that the apache user does exist. There is a reasonable workaround[0], but we could also take steps to fix this such as creating a "pulp" user and running all processes as it. We could also attempt to create the apache user ourselves, though I believe the pulp user may be more appropriate. In any case, it will be important for us to ensure that the user always has the same UID/GID. Using the existing "apache" user might help with this, as it seems to always have UID:GID of 48:48 for Red Hat based systems (I checked F20 and RHEL 7.) Version-Release number of selected component (if applicable): 2.4.3-1 How reproducible: Every time. Steps to Reproduce: 1. Configure a Pulp worker box that does not have httpd installed on it. 2. Attempt to start the workers. Actual results: The workers will fail to start, because the system will attempt to run them as the apache user, which does not exist. Expected results: The workers should start out of the box. Additional info: [0] A workaround is to manually create the apache user, using the same UID/GID as it has on the other systems that are sharing the filesystem.
Moved to https://pulp.plan.io/issues/678