>From reading the docs what I concluded is that all tuples know their parent
spout tuple so all tuples from the same tree go to the same acker task.
 However, for one instance I saw all tuples going to only one acker task no
matter how many acker threads I gave the topology.

I was working with a topology where I was reading data from a file where
each line is a new record. For each record, I would emit a tuple from the
spout.  All tuples generated by a file were emitted at the same time (what
I mean is when the storm framework called execute I would emit a bunch of
tuples by doing a for loop and reading the file line by line).  I know this
is a bad design and we have moved away from it.  However, when we tried
this I noticed all tuples were being tracked by only one acker task even
though I gave the topology 10 acker threads.  Why would this be the case?
 My understanding from the docs is that each spout tuple will go to a
different acker task and all of its children tuples will go to that acker
task.  However, what I saw is that all tuples in this topology were going
to just one acker.  Am I not understanding something about the way acker
tasks work?

Naresh

Reply via email to