storm-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sai SaiGraph <saisaigr...@gmail.com>
Subject Storm Nub questions.
Date Tue, 22 Jul 2014 07:41:19 GMT
Hi
I m trying to understand parellism in Storm a few questions about it:

1. Lets say i have only a single file i am trying to test
with a single line L1
Now if i use 2 spouts, will there be 2 spouts created
or is it only 1 spout which will b created as there is only a line.

2. Is parallelism hint mean that it is not for sure
that multiple objects are created and it is upto
storm to create them based on the size of the data.

3. In a single node this will create 1 spout that is 1 object in 1 thread:

builder.setBolt(SPLIT_BOLT_ID, splitBolt, 1)
    .shuffleGrouping(SENTENCE_SPOUT_ID);

and In a single node this will create 2 spouts  that is 2 objects in 1
thread:

builder.setBolt(SPLIT_BOLT_ID, splitBolt, 2)
    .shuffleGrouping(SENTENCE_SPOUT_ID);

My questions is, is this really parallelism.
Is it true to say this is parallelism at a thread level
in a single node will this really help achieve the parallelism.

The reason i ask is because in hadoop we achieve parallelisn
at a process level which is easy to digest
that we are achieving parallelism, but in storm i am
confused. Can some one pls pour some light.
Thanks
Sai

Mime
View raw message