Hello, I had the same issue when I tried to collocate all my executors to one worker. I believe it has to do with resource contention among the different threads (at least the spouts that are not starting). I would advise you to make sure that you do not over-provision the system's resources and end up with "thrashing" threads.
Regards, Nikos 2015-07-28 9:11 GMT-04:00 Aditya Rajan <[email protected]>: > Hello, > > We recently migrated our topology from a 12 x 4 core machine cluster to a > 1x 32 core machine. We have suddenly started facing issues like Spouts not > starting on occasion . Also sometimes the topology dies and restarts > without reason. > > We assume that we are making errors with the storm.yaml or other > configuration files. We;ve changed all the child opts to higher values. > > Is there something we are missing? > > Thanks and Regards > Aditya Rajan > -- Nikolaos Romanos Katsipoulakis, University of Pittsburgh, PhD candidate
