If maintaining the order of the messages is a requirement, fields grouping
seems to be the only strategy that ensures that all tuples of the same
partition will be sent to the same task ID.
https://storm.apache.org/releases/current/Concepts.html
Stream groupings
> Part of defining a topology i
Hi,
Is there a standard way of avoiding bottleneck which arises due to fields
grouping from one bolt to another. I have a use case where half a million
tuples have the same field and go to the same bolt task because of field
grouping. I cannot use shuffle grouping here because it is important that
