1. In general, each subscriber gets only 1 copy of the data. Occasionally, a subscribe can get a small # of duplicated messages and the application has to deal with it. We are trying to reduce # of dups in KAFKA-228.
2. Kafka has low requirements on CPU and memory. The I/O requirement depends on your application. 3. Kafka is still a young project. At this moment, you will need some monitoring of Kafka. Jun On Wed, Dec 21, 2011 at 2:26 PM, S Ahmed <sahmed1...@gmail.com> wrote: > Just reading through a kafka doc: > > 1. "Kafka > is a multi-subscriber system and a single message may be > consumed multiple times by different consumer applications." So this means > the clients that are consuming the messages are in charge of duplication > detection so as to not process the same message twice? > > 2. is there a minimum suggest hardware to setup kafka? (# of servers, > suggest ram, virtual versus dedicated bare metal hardware) > > 3. is kafka a setup and forget type system or does it require constant > monitoring as it can easily go down etc? (seeing as it hasn't hit 1.0 yet) >
