I understand the distinction, but I'm a developer on the project. I don't think it's intuitive to the end user to tell them to edit a Pulp configuration file (/etc/pulp/client.conf) and a completely non-Pulp branded configuration file (/etc/gofer/agent.conf) in order to run a client on a different machine than the server. A "pulp user" probably won't know what gofer is or think to look there (the same argument can probably be applied to debugging issues as well). I think that's such a common use case that we're not making intuitive enough for users to adopt. It should all be configured through one Pulp configuration file.
Agreed. Updated plugin descriptor for the pulp (gofer) plugin to reference the (host) in /etc/pulp/client.conf. This way, users only need to edit the pulp branded file. Works in conjunction with the fix for bug 669498. This should be intuitive for pulp users. To point the client at a remote pulp server, the user does: 1) vi /etc/pulp/client.conf 2) service pulp-agent restart Done! I'll update the UG at the next community release. Commit: 557a3aed6f2dbe58493a56ec2b2f2a3fa4167bf7
Fixed in 0.136.
verified [root@10 ~]# service pulp-agent restart Stopping goferd [FAILED] Starting goferd [ OK ] [root@10 ~]# rpm -q pulp package pulp is not installed [root@10 ~]# rpm -q pulp-client pulp-client-0.0.137-1.fc14.noarch [root@10 ~]#
Closing with community release pulp-0.0.139-1.fc14.noarch