- Native Scala types Set and List responses.
- Transparent serialization
- Connection pooling
- Consistent Hashing on the client.
- Support for Clustering of Redis nodes.
Redis is a key-value database. It is similar to memcached but the dataset is not volatile, and values can be strings, exactly like in memcached, but also lists and sets with atomic operations to push/pop elements.
- Fast in-memory store with asynchronous save to disk.
- Key value get, set, delete, etc.
- Atomic operations on sets and lists, union, intersection, trim, etc.
- sbt (get it at http://www.scala-sbt.org/)
Add to build.sbt
libraryDependencies ++= Seq(
"net.debasishg" %% "redisclient" % "3.30"
)Start your redis instance (usually redis-server will do it)
$ cd scala-redis
$ sbt
> update
> console
And you are ready to start issuing commands to the server(s)
Redis 2 implements a new protocol for binary safe commands and replies
Let us connect and get a key:
scala> import com.redis._
import com.redis._
scala> val r = new RedisClient("localhost", 6379)
r: com.redis.RedisClient = localhost:6379
scala> r.set("key", "some value")
res3: Boolean = true
scala> r.get("key")
res4: Option[String] = Some(some value)
Let us try out some List operations:
scala> r.lpush("list-1", "foo")
res0: Option[Long] = Some(1)
scala> r.rpush("list-1", "bar")
res1: Option[Long] = Some(2)
scala> r.llen("list-1")
res2: Option[Long] = Some(2)
Let us look at some serialization stuff:
scala> import serialization._
import serialization._
scala> import Parse.Implicits._
import Parse.Implicits._
scala> r.hmset("hash", Map("field1" -> "1", "field2" -> 2))
res0: Boolean = true
scala> r.hmget[String,String]("hash", "field1", "field2")
res1: Option[Map[String,String]] = Some(Map(field1 -> 1, field2 -> 2))
scala> r.hmget[String,Int]("hash", "field1", "field2")
res2: Option[Map[String,Int]] = Some(Map(field1 -> 1, field2 -> 2))
scala> val x = "debasish".getBytes("UTF-8")
x: Array[Byte] = Array(100, 101, 98, 97, 115, 105, 115, 104)
scala> r.set("key", x)
res3: Boolean = true
scala> import Parse.Implicits.parseByteArray
import Parse.Implicits.parseByteArray
scala> val s = r.get[Array[Byte]]("key")
s: Option[Array[Byte]] = Some([B@6e8d02)
scala> new String(s.get)
res4: java.lang.String = debasish
scala> r.get[Array[Byte]]("keey")
res5: Option[Array[Byte]] = None
scala-redis is a blocking client, which serves the purpose in most of the cases since Redis is also single threaded. But there may be situations when clients need to manage multiple RedisClients to ensure thread-safe programming.
scala-redis includes a Pool implementation which can be used to serve this purpose. Based on Apache Commons Pool implementation, RedisClientPool maintains a pool of instances of RedisClient, which can grow based on demand. Here's a sample usage ..
val clients = new RedisClientPool("localhost", 6379)
def lp(msgs: List[String]) = {
clients.withClient {
client => {
msgs.foreach(client.lpush("list-l", _))
client.llen("list-l")
}
}
}Using a combination of pooling and futures, scala-redis can be throttled for more parallelism. This is the typical recommended strategy if you are looking forward to scale up using this redis client. Here's a sample usage .. we are doing a parallel throttle of an lpush, rpush and set operations in redis, each repeated a number of times ..
If we have a pool initialized, then we can use the pool to repeat the following operations.
// lpush
def lp(msgs: List[String]) = {
clients.withClient {
client => {
msgs.foreach(client.lpush("list-l", _))
client.llen("list-l")
}
}
}
// rpush
def rp(msgs: List[String]) = {
clients.withClient {
client => {
msgs.foreach(client.rpush("list-r", _))
client.llen("list-r")
}
}
}
// set
def set(msgs: List[String]) = {
clients.withClient {
client => {
var i = 0
msgs.foreach { v =>
client.set("key-%d".format(i), v)
i += 1
}
Some(1000)
}
}
}And here's the snippet that throttles our redis server with the above operations in a non blocking mode using Scala futures:
val l = (0 until 5000).map(_.toString).toList
val fns = List[List[String] => Option[Int]](lp, rp, set)
val tasks = fns map (fn => scala.actors.Futures.future { fn(l) })
val results = tasks map (future => future.apply())scala-redis is a blocking client for Redis. But you can develop high performance asynchronous patterns of computation using scala-redis and Futures. RedisClientPool allows you to work with multiple RedisClient instances and Futures offer a non-blocking semantics on top of this. The combination can give you good numbers for implementing common usage patterns like scatter/gather. Here's an example that you will also find in the test suite. It uses the scatter/gather technique to do loads of push across many lists in parallel. The gather phase pops from all those lists in parallel and does some compuation over them.
Here's the main routine that implements the pattern:
// set up Executors
val system = ActorSystem("ScatterGatherSystem")
import system.dispatcher
val timeout = 5 minutes
private[this] def flow[A](noOfRecipients: Int, opsPerClient: Int, keyPrefix: String,
fn: (Int, String) => A) = {
(1 to noOfRecipients) map {i =>
Future {
fn(opsPerClient, keyPrefix + i)
}
}
}
// scatter across clients and gather them to do a sum
def scatterGatherWithList(opsPerClient: Int)(implicit clients: RedisClientPool) = {
// scatter
val futurePushes = flow(100, opsPerClient, "list_", listPush)
// concurrent combinator: Future.sequence
val allPushes = Future.sequence(futurePushes)
// sequential combinator: flatMap
val allSum = allPushes flatMap {result =>
// gather
val futurePops = flow(100, opsPerClient, "list_", listPop)
val allPops = Future.sequence(futurePops)
allPops map {members => members.sum}
}
Await.result(allSum, timeout).asInstanceOf[Long]
}
// scatter across clietns and gather the first future to complete
def scatterGatherFirstWithList(opsPerClient: Int)(implicit clients: RedisClientPool) = {
// scatter phase: push to 100 lists in parallel
val futurePushes = flow(100, opsPerClient, "seq_", listPush)
// wait for the first future to complete
val firstPush = Future.firstCompletedOf(futurePushes)
// do a sum on the list whose key we got from firstPush
val firstSum = firstPush map {key =>
listPop(opsPerClient, key)
}
Await.result(firstSum, timeout).asInstanceOf[Int]
}See an example implementation using Akka at https://github.com/debasishg/akka-redis-pubsub.
RedisCluster uses data sharding (partitioning) which splits all data across available Redis instances,
so that every instance contains only a subset of the keys. Such process allows mitigating data grown
by adding more and more instances and dividing the data to smaller parts (shards or partitions).
RedisCluster allows user to pass a special KeyTag, that helps to distribute keys according to special
requirements. Otherwise node is selected by hashing the whole key with CRC-32 function.
RedisCluster also allows for dynamic nodes modification with addServer, replaceServer and removeServer
methods. Note that data on the disconnected node will be lost immediately.
What is more, since modification of the cluster impacts key distribution, some of the data scattered
across the cluster could be lost as well.
For automatic node downtime handling, by disconnecting the offline node and reconnecting it as it comes back up,
there is a Reconnectable trait. To allow such behaviour mix it into RedisCluster instance:
new RedisCluster(nodes, Some(NoOpKeyTag)) with Reconnectable
you can observe it's behaviour in ReconnectableSpec test.
This software is licensed under the Apache 2 license, quoted below.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.