I do not understand the functionality of "repository" of the agent handler. Cached packages as requested.
Agents have different policies for assigning a repository or another, using the "same" idea with the agent handlers. From where the nodes will be updated? Who has priority, distributed repositories or agent handlers? Or I'm completly wrong?
Not wrong, just at a tangent
Where an agent gets updates from is controlled by the sitelist.xml
The sitelist is effectively recieved by the agent though a policy update.
An agent handler 'cache' behaves like a mini-master repository in a way. At least that is how it looks to the agents.
Assume: Agent wants to speak to ePO server, but policy says use agent handler 'x' instead,
Agent speaks to Agent handler as if it were the ePO server.
Agent handler requests files as needed from the ePO server and caches them.
Agent handler then serves the data to the agent as if it were the ePO server.
But I don't understand, if I have a policy that says to my agents to get updates from a repository A, for example an UNC repo, and I install an agent handler on B and says to agents to use this agent handler. From where will the agents get the updates, from de repo or from the agent handler?
If you have a policy which says 'get updates from repository A' then that is where it will get updates from
If you have a policy which says 'speak to Agent Handler X' then it will get policies and tasks, and send event through Agent Handler X - but it will *still* get updates from repository A.
If you then change it so it says 'get updates from ePO server', it will request them through Agent Handler 'X' not directly to the ePO server.