I am using redis with Akka so I need no blocking calls. Lettuce has async-future call built into it. But Jedis is the recommended client by Redis. Can someone tell me if I am using both of them the right way. If so which one is better.
JEDIS I am using a static Jedis connection pool to get con and using Akka future callback to process the result. My concern here is when I use another thread (callable) to get the result that thread is eventually going to block for the result. While Lettuce might have some more efficient way of doing this.
private final class OnSuccessExtension extends OnSuccess<String> {
private final ActorRef senderActorRef;
private final Object message;
@Override
public void onSuccess(String valueRedis) throws Throwable {
log.info(getContext().dispatcher().toString());
senderActorRef.tell((String) message, ActorRef.noSender());
}
public OnSuccessExtension(ActorRef senderActorRef,Object message) {
this.senderActorRef = senderActorRef;
this.message=message;
}
}
ActorRef senderActorRef = getSender(); //never close over a future
if (message instanceof String) {
Future<String> f =akka.dispatch.Futures.future(new Callable<String>() {
public String call() {
String result;
try(Jedis jedis=JedisWrapper.redisPool.getResource()) {
result = jedis.get("name");
}
return result;
}
}, ex);
f.onSuccess(new OnSuccessExtension(senderActorRef,message), ex);
}
LETTUCE
ExecutorService executorService = Executors.newFixedThreadPool(10);
public void onReceive(Object message) throws Exception {
ActorRef senderActorRef = getSender(); //never close over a future
if (message instanceof String) {
final RedisFuture<String> future = lettuce.connection.get("name");
future.addListener(new Runnable() {
final ActorRef sender = senderActorRef;
final String msg =(String) message;
@Override
public void run() {
try {
String value = future.get();
log.info(value);
sender.tell(message, ActorRef.noSender());
} catch (Exception e) {
}
}
}, executorService);
If lettuce is a better option for Async calls. Then what type of executor should I go with in production environment. If possible can I use a Akka dispatcher as an execution context for Letture future call.
