An ePO Agent Handler must have a high availability and high bandwidth connection to the ePO database:
Inevitably, the question of whether it is more appropriate to add another Agent Handler or deploy another ePO server arises. As a general response, Agent Handlers should be used when:
-
The existing ePO infrastructure needs to be expanded to handle more agents, more products, or a higher load due to more frequent ASCI communication.
-
You want to ensure that agents continue to dial in and receive policy, task, and product updates even if the application server is unavailable.
-
You want to expand ePO management into a disconnected network segment, where there is still a high-bandwidth link between the Agent Handler and the ePO database.
An Agent Handler should not be installed to:
-
Replace distributed repositories. Distributed repositories exist to distribute large files throughout an organization, and do not contain any logic.
-
Connect a disconnected network segment where there is limited or irregular connectivity to the ePO database.
Usually it is preferable to add an Agent Handler rather than a new ePO server. A new ePO server is the best choice only for separate IT infrastructures, separate administrative groups, or test environments.
A high availability and high-bandwidth connection is required because all data that the managed clients send to the Agent Handler must be provided to the database over this connection. The requirements of the high availability and high bandwidth connection required is less than 10-ms connection time. If the connection between the handler and the database is unavailable or has insufficient bandwidth to provide the data in a timely manner, there are various potential negative results, including but not limited to:
- Max connection errors in the Apache log (server.log) because the handler cannot process agent-to-server communication requests in a timely fashion
- Agent-to-server communication failures
- A backlog of events in the <Agent Handler install directory>\DB\Events folder