Flume ships with ready to use Channel implementations. Lets look at them
- Memory Channel
- File Channel
- JDBC Channel
Memory Channel keeps all the Events stored in memory. In case, the Flume process goes down, the Events are lost. Max capacity of the Channel is configurable
File Channel stored all the Events on File system. A more detailed discussion on the design can be found here
JDBC Channel stores Events in a database. Currently Derby database is supported.