Quasar Redis Example

Getting Started

Redis and make must be installed to run the tests. make test will run a get/set example and a pub-sub one.

Quasar and Redis integration

Redis is an in-memory data structure store, used as a database, cache and message broker.

The comsat-redis module is compatible with JDK 8 and higher; it uses Lettuce's async API, which is based on JDK8's CompletableStage, to offer a synchronous fiber-blocking Jedis API. At present it supports most basic commands and Redis pub-sub.

The co.paralleluniverse.fibers.redis.BinaryJedis, Jedis, BinaryJedisPubSub and JedisPubSub classes extend the corresponding Jedis' ones so in most cases using the integration is as simple as including the module as a compile-time dependency, enabling Quasar instrumentation and replacing references to the redis.clients.jedis package with co.paralleluniverse.fibers.redis.

There are some difference between comsat-redis and Jedis:

• When invoked from fibers comsat-redis will block fibers while from threads it will block threads; this means it's generally better to use it inside fibers in order to get maximum concurrency and efficiency but it can also be used as a traditional Jedis driver from Java threads.
• When using pub-sub, the callback handlers have to extend Comsat's BinaryJedisPubSub or JedisPubSub rather than Jedis' (but usually it's just a matter of switching package). Also, each callback method will run in a fiber and will have to be annotated as @Suspendable if it calls fiber-blocking methods.
• Lettuce's driver is thread-safe for basic commands so there's usually no need to use pooling: a single Jedis instance will suffice and will yield better performance.

Here's a complete list of features that are currently supported:

• Pub-sub
• All basic commands and operations except geo commands, getClient, zrevrangeByLex, geohash, rpushx, lpushx, spop with count, zincrby with count, scan commands with parameters, bit operations with parameters, zstore operations with parameters and Jedis-deprecated calls.

Clustering, sharding, replication, transaction/multi, pipelining, scripting, monitoring and all other features not explicitly listed above are not yet supported.