I usually am determined enough to push through the learning when it comes to new frameworks, etc. However, making the move from Storm to Trident has melted my brain. Even though I have already developed Kakfa/Avro/Storm the Trident documentation leaves me chasing answers/examples in circles.
*"the word counts are kept in memory, but this can be trivially swapped to use Memcached, Cassandra, or any other persistent store."* Yet it seems anything but trivial when attempted in the case of Cassandra. *"Suppose you have a home-grown database..." *Do people write their own databases? Has this example ever been used anywhere? Is this something I should be doing? I know how bad it looks to be complaining about documentation on an open source project that is so totally awesome. However, I believe that most users are going to use Trident primarily, if not exclusively, for the AtMostOnceProcessing which depends on the persistence of state to something like Cassandra (as mentioned in the docs). So far, all of the information I have consumed (online and in book form) has missing pieces that make, what I believe to be, a VERY common configuration/use-case very difficult to dissect. Kafka->Trident->Cassandra I was hoping that someone could share a working example of a Trident Topology that 1. Persists state to Cassandra so that AtMostOnce can be guaranteed. 2. Has a working implementation of an IOpaqueParitionedTridentSpout that reads from a Kafka Topic. I am sorry to come off like a jerk. I know there is something small that my tiny brain cannot conceptualize about Trident. The Storm docs are great! I am just getting desperate in trying to find the elusive Kafka->Trident->Cassandra example. I will share back what I am able to synthesize from your kind donation of knowledge.
