Setting up a repeater is as simple as installing an RSCD agent and NSH and making su re that the exports files allow connectivity. If you are not using "* rw" or similar, then agents that will be using a repeater should have the repeater listed as a host in the exports.
If you are using the optional function of secure certificates between the App Server and clients, there are additional steps documented in the "Certificates" section of the Admin Guide.
From what you are saying I understand that the repeater role is only acting as a kind of centralised channeling point for software/package distribution to all clients.
Can the repeater act as a repository for commonly distributed packages/files, so for example, when I create a job to deploy a package the instruction is sent to the repeater to execute and deploy the software thus avoiding the transfer of the package from the application server to the repeater?
There is no concept of permanently keeping a package on a repeater, it is used as a cache during deployments.
A repeater is most useful when you want to deploy a package to multiple targets at once. The app server will stage the package to the repeater once. The repeater will then fan-out the package to all targets that use it as a repeater.
you are looking for the holy grail, the distributed depot. It's in the road-map, but we're not quite there yet!
Any update on this for 7.6?