Cassandra Journal for Akka Persistence

A replicated Akka Persistence journal backed by Apache Cassandra.

Installation

Plugin dependency

To include this journal plugin into your sbt project, add the following lines to build.sbt:

resolvers += "krasserm at bintray" at "http://dl.bintray.com/krasserm/maven"

libraryDependencies += "com.github.krasserm" %% "akka-persistence-cassandra" % "0.2.1"

This version of the plugin depends on Akka 2.3.2 and is cross-built against Scala 2.10.4 and 2.11.0.

Cassandra database

The plugin requires an exiting installation of Cassandra 2.0.3 or higher. You may want to follow this Getting Started guide for basic installation instructions.

Features

• All operations required by the Akka Persistence journal plugin API are fully supported.
• The plugin uses Cassandra in a pure log-oriented way i.e. data are only ever inserted but never updated (deletions are made on user request only or by persistent channels, see also Caveats).
• Writes of messages and confirmations are batched to optimize throughput. See batch writes for details how to configure batch sizes. The plugin was tested to work properly under high load.
• Messages written by a single processor are partitioned across the cluster to achieve scalability with data volume by adding nodes.

Configuration

To activate the plugin, add the following line to your Akka application.conf:

akka.persistence.journal.plugin = "cassandra-journal"

This will run the journal with its default settings. The default settings can be changed with the following configuration keys:

• cassandra-journal.contact-points. A comma-separated list of contact points in a Cassandra cluster. Default value is [127.0.0.1].
• cassandra-journal.keyspace. Name of the keyspace to be used by the plugin. If the keyspace doesn't exist it is automatically created. Default value is akka.
• cassandra-journal.table. Name of the table to be used by the plugin. If the table doesn't exist it is automatically created. Default value is messages.
• cassandra-journal.replication-factor. Replication factor to use when a keyspace is created by the plugin. Default value is 1.
• cassandra-journal.max-partition-size. Maximum number of entries (messages, confirmations and deletion markers) per partition. Default value is 5000000. Do not change this setting after table creation (not checked yet).
• cassandra-journal.max-result-size. Maximum number of entries returned per query. Queries are executed recursively, if needed, to achieve recovery goals. Default value is 50001.
• cassandra-journal.write-consistency. Write consistency level. Default value is QUORUM.
• cassandra-journal.read-consistency. Read consistency level. Default value is QUORUM.

The default read and write consistency levels ensure that processors can read their own writes. During normal operation, processors only write to the journal, reads occur only during recovery.

Caveats

• Detailed tests under failure conditions are still missing.
• Persistent channel recovery times are not optimized yet (see also issue 4).
  • Recovery time increases with the number of tombstones and drops to a minimum after compaction.
  • Persistent channel throughput is independent of the number of tombstones after successful recovery.
• Range deletion performance (i.e. deleteMessages up to a specified sequence number) depends on the extend of previous deletions
  • linearly increases with the number of tombstones generated by previous permanent deletions and drops to a minimum after compaction
  • linearly increases with the number of plugin-level deletion markers generated by previous logical deletions (recommended: always use permanent range deletions)

These issues are likely to be resolved in future versions of the plugin.