It is absolutely ok what you are doing. No need to worry about anything.

If your tuples really time out, you can increase the timeout via
TOPOLOGY_MESSAGE_TIMEOUT_SECS (the default value is 30 seconds).

-Matthias


On 01/10/2016 06:50 PM, John Yost wrote:
> Hi Everyone,
> 
> I have a topology that, for every tuple coming in from a KafkaSpout the
> following occurs:
> 
> 1. processing in Bolt A generates approximately 20 tuples that are
> anchored and sent to Bolt B
> 2. each anchored, processed tuple is cached in Bolt B and eventually
> written to a NoSQL database
> 
> Question--since there are roughly 20 tuples to be acked before the
> anchor tuple coming from Kafka is acked, and there will be some form of
> delay for each, is this a scenario that guaranteed message processing in
> Storm is designed to handle? Are there any best practices that may make
> this process more efficient or more easily fault tolerant, or is Storm
> good-to-go as-is?
> 
> Any guidance or thoughts would be great--thanks! :)
> 
> --John

Attachment: signature.asc
Description: OpenPGP digital signature

Reply via email to