we—the Akka committers—are pleased to be able to announce the availability of Akka 2.4.0-RC3 (THIRD RELEASE CANDIDATE). This is what we intend to ship as 2.4.0 final unless issues are found, so please test it thoroughly and report back. Failures are important to hear about, but praise also does not hurt :-)
Since the last release candidate two weeks ago we have mostly improved the documentation, but there are few other changes also:
- make Persistence Query API explorable (more IDE friendly), thanks to Brian Scully for the design suggestion
- make event adapter lookup more efficient, to fix this we had to break binary compatibility with the Persistence journal plugins released for 2.4.0-RC2 so we kindly ask the plugin maintainers to publish new versions
- Persistence schema evolution section of the documentation has been ported to Java
- fixed LevelDB journal issue when deleting all events
- Utility program that can remove the internal data stored by the Cluster Sharding coordinator
- fixed regression in Cluster Singleton configuration
The complete list of closed tickets can be found in the 2.4.0-RC3 github issues milestone.
The key features of the 2.4.0 release are:
- we dropped support for Java 6 & 7 as announced in the last roadmap update, and now require Java 8 or later
- we dropped support for Scala 2.10, kept 2.11 and added 2.12 (which is at milestone 2 currently); this is in keeping with our policy to support the Scala version that is “current” when the first milestone comes out plus any later version that is published during this series’ lifetime
- Akka Persistence was promoted to a fully supported module (not experimental)
- added experimental stream based API for the Query Side of Akka Persistence (implementations must be provided by query plugins corresponding to the journal plugins for the various data stores)
- improved support for Akka Persistence event migrations by Serializer with String Manifest and Event Adapters and prepared comprehensive documentation of strategies for schema evolution of persistent events
- made Akka Persistence failure handling more robust and support rejections of events
- Akka Persistence can now use multiple different Journals within one ActorSystem (thanks to Andrei Pozolotin, who also split out ClusterMetrics into their own module)
- introduced Persistent FSM, thanks to an awesome contribution from leonidb
- simplified setting up Akka Clusters behind NAT (including inside Docker containers)
- we added the experimental Akka Typed module previously codenamed Project Gålbma, a new way of formulating Actor interactions with full type-system support
- we promoted the ClusterSingleton, ClusterClient, DistributedPubSub and ClusterSharding patterns to fully a supported Cluster Tools module, including various API and configuration improvements
- ClusterSharding also learnt a few new tricks (graceful shutdown, asynchronous shard allocation, reviving entries after migration thanks to Dominic Black, and more flexible use of roles [thanks to Richard Marscher)
- use custom failure detection in ClusterClient to avoid quarantining in case of transient network failures
- added experimental Akka Distributed Data which was previously known as Akka Data Replication (see akka-data-replication migration guide), the Distributed Data module is useful for sharing eventually consistent data (CRDTs) between nodes in an Akka Cluster
- new experimental feature that makes it possible to promote new members to WeaklyUp during network partitions, thanks to Hector Veiga for contributing
- the SLF4J logging adapter can now apply the log filtering rules prior to sending to the EventStream, see the migration guide
- make it possible to use any version of Protobuf for serialization of application messages, the internal dependency to Protobuf has been replaced by embedding a shaded (renamed package) version, if you use Protobuf in your application you need to add the dependency to your build, see migration guide
- but the biggest feature is probably that Akka 2.4 will be binary backwards compatible with Akka 2.3, see the detailed description below.
Things that are Known Missing
We will not be able to complete the Java API and the documentation for Akka Typed for 2.4.0, but the Scala API will be released in order to gather early feedback on this rather exciting new module.
We will incorporate Streams & HTTP once ready (to become an experimental module, as usual) into 2.4.x at a later time. Note that you can use Akka Streams 1.0 together with Akka 2.4.0-RC3.
Akka 2.4.x is backwards binary compatible with previous 2.3.x versions (exceptions listed below). This means that the new JARs are a drop-in replacement for the old one (but not the other way around) as long as your build does not enable the inliner (Scala-only restriction). It should be noted that Scala 2.11.x is is not binary compatible with Scala 2.10.x, which means that Akka’s binary compatibility property only holds between versions that were built for a given Scala version—akka-actor_2.11-2.4.0-RC3.jar is compatible with akka-actor_2.11-2.3.14.jar but not with akka-actor_2.10-2.3.14.jar.
Binary compatibility is not maintained for the following:
- akka-testkit and akka-multi-node-testkit
- experimental modules, such as akka-persistence and akka-contrib
- features, classes, methods that were deprecated in 2.3.x and removed in 2.4.x
Being binary compatible means that applications and libraries built on top of Akka 2.3.x continue to work with Akka 2.4.x without recompilation (subject to the conditions below), which implies that Akka Streams & HTTP as well as Play Framework 2.4 can be combined with Akka 2.4.
The dependency to Netty has been updated from version 3.8.0.Final to 3.10.3.Final. The changes in those versions might not be fully binary compatible, but we believe that it will not be a problem in practice. No changes were needed to the Akka source code for this update. Users of libraries that depend on 3.8.0.Final that break with 3.10.3.Final should be able to manually downgrade the dependency to 3.8.0.Final and Akka will still work with that version.
The dependency to Typesafe Config has been updated from 1.2.1 to 1.3.0 which should be binary compatible for the vast majority users, except for obscure edge cases as its changelog points out. This change was made in order in order to use new JDK8 specific features in the library as well as to align Akka with Play which is now also depending on 1.3.0.
When migrating a code base to 2.4 please refer to the migration guide in order to profit from some of the improvements.
Old Akka Persistence plugins released for 2.3 are not compatible with the changes that were done in 2.4-M2. This is a good time for Akka Persistence plugin maintainers to migrate and release a preview for 2.4.0-RC3. Even though there are rather many API changes since 2.3 in the Persistence plugin API it should not be difficult to migrate. See the Persistence Plugin APIs section in the migration guide.
We have carefully made changes in Akka Persistence with the goal that events stored with Akka Persistence 2.3.14 should still be possible to replay with 2.4.x, i.e. no lost data. Specific journal implementations may of course violate that goal.
Please do what you usually do so well: try out this release candidate and report back when things break, not work as advertised, feel strange, or even when you are happy :-) Especially concerning binary compatibility we will need help from the community (you!) since we cannot run all possible programs ourselves; we base our BC efforts on the MiMa plugin but that is no perfect guarantee that everything will work out of the box.
The artifacts comprising this release have been published to https://repo1.maven.org/maven2/ and also to Maven Central. In addition, we adopted the sbt standard of encoding the Scala binary version in the artifact name, i.e. the core actor package’s artifactId is “akka-actor_2.11” or “akka-actor_2.12.0-M2”, respectively.
33 issues were closed since 2.4.0-RC2. The complete list of closed tickets can be found in the 2.4.0-RC3 github issues milestone.
commits added removed 26 4460 1360 Patrik Nordwall 9 500 106 Johan Andrén 2 5 3 Konrad Malawski 2 3 2 Jamie 1 21 6 Marek Kadek 1 1 0 luben karavelov 1 39 9 Daryl Odnert