Hi,
We are running 3 ec2 zookeeper instances solely for Storm. Once a month or so the Zookeeper cluster completely collapses, and Storm supervisors goes out of service. We keep tweaking the settings but it would be much faster if we can compare notes. Itai
