Thank you. Is there any other optimizing method such as modifying storm config?
I set the TOPOLOGY_DISRUPTOR_BATCH_SIZE and
TOPOLOGY_DISRUPTOR_BATCH_TIMEOUT_MILLIS to 1, it seems better only when I set
the worker number to 1.
发件人: Kevin Conaway mailto:kevin.a.cona...@gmail.com>>
答复: "user@storm
Hello.
I do test with a simple topology to test the intercommunication latency of
spout/bolt. It’s just emit the current nano timestamp from a spout and print
the time difference when a bolt receive it.
I deploy my storm cluster in my own machine with docker container (one nimbus,
one supervisor