We are pleased to announce the availability of Akka 2.2.0 "Coltrane." This is the second evolutionary step of the Akka 2 series outlined in our roadmap. The biggest feature is cluster support, which has already been part of the 2.1 release as an experimental preview. We have improved, tested and hardened it and, thanks to the valuable feedback from our users, can confidently remove the "experimental" flag: Akka cluster is now officially supported.
The focus of our development efforts for this milestone was to further unify the semantics of actor communications between local and remote deployment. The remoting layer has been enhanced to support reliable delivery of system-internal messages and connection failure detection. Building on this we can now offer the same reliability of supervision and lifecycle monitoring in an actor hierarchy independent of whether it runs on a single node or a cluster. In addition to these changes within Akka's internals, we also prepared the future removal of some API methods which break location transparency, for example calling "isTerminated" on an actor should be replaced by using lifecycle monitoring. These changes benefit your code to make it more robust, also in the face of future changes to your application's deployment. In addition we have been working on several other features which include:
This is just a very high-level overview of the biggest pieces, please refer to the announcement on the Akka team blog for more details, a series of blog posts have been published over the past weeks to cast some spotlights. To get started with Akka try out Typesafe Activator which offers several Akka tutorials based on this new release.
We would like to send special thanks to all our external contributors, listed in decreasing order of commits: Mathias Doenitz, Johannes Rudolph, Dario Rexin, Rick Latrine, Christophe Pache, Raman Gupta, Kaspar Fischer, Ricky Elrod, Kevin Wright, Raymond Roestenburg, Michael Pilquist, Helena Edelson, Matthew Neeley, Thomas Lockney and Derek Mahar.