I am currently scheduling to run multiple instances of a bolt to run on a
single worker slot.
Say for example, i have a bolt with name BOLT_TESTA with parallelism 4.
I am running this bolt on 4 executors, but on the same worker slot.

Is this same as running single executor with number of tasks as 4 ?

I was expecting that it runs 4 different instances of the bolt. But when i
check the logs i see that the data is always going to the single instance
even though my grouping is based on fields and the parameter that it is
grouping on - i gave it as a sequence with atleast 10 input records.

But when i run single instance on each worker slots (I have 4 worker slots
in this case), i see that the data is going to all the 4 slots.

Can someone please enlighten me on this.


Reply via email to