Hello everybody, I have a question concerning storm data consistency. We are using AWS and have deployment process when we deploy storm supervisor using auto scaling groups. So new autoscaling group created and deployed, then old one starts shrinking until reaching 0 instances.
All instances inside old autoscaling group shutting down gracefully(waiting until all init scripts stopped). We also have custom supervisor script which is running on "stop" until supervisor java process is gone. Is there any possibility that data could be lost during this deployment? BR, Andrii
