Thank you Erik weathers. next time for sure I will place the source.
*Best regards,* *K.Sai Dilip Reddy.* On Wed, Apr 27, 2016 at 12:16 PM, Erik Weathers <eweath...@groupon.com> wrote: > So, you got those quotes from the following page it seems (please cite > your source next time): > > https://storm.apache.org/releases/current/FAQ.html > > > *Number of workers a multiple of number of machines * > > I'm not sure why that statement is there. It entirely depends on how many > topologies you are running, what they are doing, etc. > > > *Number of kafka partitions a multiple of number of spout parallelism* > > This is a bit confusing, but the point is that a given kafka partition can > only be handled by a single spout executor. You cannot have 2 spouts > reading from the same partition. So if you have say 10 kafka partitions > for a given topic, then the number of spout executors (i.e., the > parallelism) should be 1, 2, 5, or 10. Numbers that are a factor of 10. > Otherwise you'll have unused executors. > > - Erik > > On Tue, Apr 26, 2016 at 9:07 PM, Sai Dilip Reddy Kiralam < > dkira...@aadhya-analytics.com> wrote: > >> can some one explain me this ? >> >> >> >> *Best regards,* >> >> *K.Sai Dilip Reddy.* >> >> On Mon, Apr 25, 2016 at 1:48 PM, Sai Dilip Reddy Kiralam < >> dkira...@aadhya-analytics.com> wrote: >> >>> Hi, >>> >>> I'm confused in setting up workers to storm cluster. >>> please some one explain the following with example. >>> >>> >>> * "1) Number of workers a multiple of number of machines ** 2) * >>> >>> *Number of kafka partitions a multiple of number of spout parallelism;" >>> * >>> >>> >>> >>> >>> *Best regards,* >>> >>> *K.Sai Dilip Reddy.* >>> >> >> >
