How can we manage cluster-wide atomic operations? Such as maintaining an auto-increment counter.
Does Hadoop provide native support for these kinds of operations? An in case ultimate answer involves zookeeper, I'd love to work out doing this in AWS/EMR.
