We are trying to have a flume cluster built which can fallback to a secondary incase of any system failure.
Could you please guide on how to setup the same with either of the two options:
1) Master/Slave configuration via the cloudera manager with mechanism to auto poll alternate master if the master goes down.
2) To build a flume cluster of agents to listen to an input stream and be reliable (any one agent goes down others would continue to get the stream). How to load balance the cluster cross the agents to avoid duplication of the stream.
Please help point me to some articles for the same or steps to accomplish the same.