一、摘要
Lettuce 是 Redis 的一款高级 Java 客户端,与 Jedis 并列成为最热门的客户端之一,目前已成为 SpringBoot 2.0 版本默认的 redis 客户端。
相比老牌 Jedis,Lettuce 属于后起之秀,不仅功能丰富,而且提供了很多新的功能特性,比如异步操作、响应式编程等等,同时还解决了 Jedis 中线程不安全的问题。
废话不多说了,如何使用呢?请看下文!
二、Lettuce
2.1、基本使用
首先,创建一个 maven 项目,引入lettuce-core
包,就可以使用了。
<dependency>
<groupId>io.lettuce</groupId>
<artifactId>lettuce-core</artifactId>
<version>5.3.1.RELEASE</version>
</dependency>
使用 lettuce 连接 redis,测试是否能正常联通!
publicclassLettuceMain{
publicstaticvoidmain(String[]args){
RedisURIredisUri=RedisURI.builder()
.withHost("127.0.0.1")
.withPort(6379)
.withPassword("111111")
.withTimeout(Duration.of(10,ChronoUnit.SECONDS))
.build();
RedisClientredisClient=RedisClient.create(redisUri);
StatefulRedisConnection<String,String>connection=redisClient.connect();
RedisCommands<String,String>commands=connection.sync();
System.out.println(commands.ping());
connection.close();
redisClient.shutdown();
}
}
2.2、同步操作
基本上只要是 Jedis 支持的同步命令操作,Lettuce 都支持。
下面,我们以同步操作字符串为例,Lettuce 的 api 操作如下!
publicclassLettuceSyncMain{
publicstaticvoidmain(String[]args){
RedisURIredisUri=RedisURI.builder()
.withHost("127.0.0.1").withPort(6379).withPassword("111111")
.withTimeout(Duration.of(10,ChronoUnit.SECONDS))
.build();
RedisClientredisClient=RedisClient.create(redisUri);
StatefulRedisConnection<String,String>connection=redisClient.connect();
//获取同步操作命令工具
RedisCommands<String,String>commands=connection.sync();
System.out.println("清空数据:"+commands.flushdb());
System.out.println("判断某个键是否存在:"+commands.exists("username"));
System.out.println("新增<'username','xmr'>的键值对:"+commands.set("username","xmr"));
System.out.println("新增<'password','password'>的键值对:"+commands.set("password","123"));
System.out.println("获取<'password'>键的值:"+commands.get("password"));
System.out.println("系统中所有的键如下:"+commands.keys("*"));
System.out.println("删除键password:"+commands.del("password"));
System.out.println("判断键password是否存在:"+commands.exists("password"));
System.out.println("设置键username的过期时间为5s:"+commands.expire("username",5L));
System.out.println("查看键username的剩余生存时间:"+commands.ttl("username"));
System.out.println("移除键username的生存时间:"+commands.persist("username"));
System.out.println("查看键username的剩余生存时间:"+commands.ttl("username"));
System.out.println("查看键username所存储的值的类型:"+commands.type("username"));
connection.close();
redisClient.shutdown();
}
}
2.3、异步操作
除此之外,Lettuce 还支持异步操作,将上面的操作改成异步处理,结果如下!
publicclassLettuceASyncMain{
publicstaticvoidmain(String[]args)throwsException{
RedisURIredisUri=RedisURI.builder()
.withHost("127.0.0.1").withPort(6379).withPassword("111111")
.withTimeout(Duration.of(10,ChronoUnit.SECONDS))
.build();
RedisClientredisClient=RedisClient.create(redisUri);
StatefulRedisConnection<String,String>connection=redisClient.connect();
//获取异步操作命令工具
RedisAsyncCommands<String,String>commands=connection.async();
System.out.println("清空数据:"+commands.flushdb().get());
System.out.println("判断某个键是否存在:"+commands.exists("username").get());
System.out.println("新增<'username','xmr'>的键值对:"+commands.set("username","xmr").get());
System.out.println("新增<'password','password'>的键值对:"+commands.set("password","123").get());
System.out.println("获取<'password'>键的值:"+commands.get("password").get());
System.out.println("系统中所有的键如下:"+commands.keys("*").get());
System.out.println("删除键password:"+commands.del("password").get());
System.out.println("判断键password是否存在:"+commands.exists("password").get());
System.out.println("设置键username的过期时间为5s:"+commands.expire("username",5L).get());
System.out.println("查看键username的剩余生存时间:"+commands.ttl("username").get());
System.out.println("移除键username的生存时间:"+commands.persist("username").get());
System.out.println("查看键username的剩余生存时间:"+commands.ttl("username").get());
System.out.println("查看键username所存储的值的类型:"+commands.type("username").get());
connection.close();
redisClient.shutdown();
}
}
2.4、响应式编程
Lettuce 除了支持异步编程以外,还支持响应式编程,Lettuce 引入的响应式编程框架是Project Reactor
,如果没有响应式编程经验可以先自行了解一下,访问地址https://projectreactor.io/。
响应式编程使用案例如下:
publicclassLettuceMain{
publicstaticvoidmain(String[]args)throwsException{
RedisURIredisUri=RedisURI.builder()
.withHost("127.0.0.1").withPort(6379).withPassword("111111")
.withTimeout(Duration.of(10,ChronoUnit.SECONDS))
.build();
RedisClientredisClient=RedisClient.create(redisUri);
StatefulRedisConnection<String,String>connection=redisClient.connect();
//获取响应式API操作命令工具
RedisReactiveCommands<String,String>commands=connection.reactive();
Mono<String>setc=commands.set("name","mayun");
System.out.println(setc.block());
Mono<String>getc=commands.get("name");
getc.subscribe(System.out::println);
Flux<String>keys=commands.keys("*");
keys.subscribe(System.out::println);
//开启一个事务,先把count设置为1,再将count自增1
commands.multi().doOnSuccess(r->{
commands.set("count","1").doOnNext(value->System.out.println("count1:"+value)).subscribe();
commands.incr("count").doOnNext(value->System.out.println("count2:"+value)).subscribe();
}).flatMap(s->commands.exec())
.doOnNext(transactionResult->System.out.println("transactionResult:"+transactionResult.wasDiscarded())).subscribe();
Thread.sleep(1000*5);
connection.close();
redisClient.shutdown();
}
}
2.5、发布和订阅
Lettuce 还支持 redis 的消息发布和订阅,具体实现案例如下:
publicclassLettuceReactiveMain1{
publicstaticvoidmain(String[]args)throwsException{
RedisURIredisUri=RedisURI.builder()
.withHost("127.0.0.1").withPort(6379).withPassword("111111")
.withTimeout(Duration.of(10,ChronoUnit.SECONDS))
.build();
RedisClientredisClient=RedisClient.create(redisUri);
//获取发布订阅操作命令工具
StatefulRedisPubSubConnection<String,String>pubsubConn=redisClient.connectPubSub();
pubsubConn.addListener(newRedisPubSubListener<String,String>(){
@Override
publicvoidunsubscribed(Stringchannel,longcount){
System.out.println("[unsubscribed]"+channel);
}
@Override
publicvoidsubscribed(Stringchannel,longcount){
System.out.println("[subscribed]"+channel);
}
@Override
publicvoidpunsubscribed(Stringpattern,longcount){
System.out.println("[punsubscribed]"+pattern);
}
@Override
publicvoidpsubscribed(Stringpattern,longcount){
System.out.println("[psubscribed]"+pattern);
}
@Override
publicvoidmessage(Stringpattern,Stringchannel,Stringmessage){
System.out.println("[message]"+pattern+"->"+channel+"->"+message);
}
@Override
publicvoidmessage(Stringchannel,Stringmessage){
System.out.println("[message]"+channel+"->"+message);
}
});
RedisPubSubAsyncCommands<String,String>pubsubCmd=pubsubConn.async();
pubsubCmd.psubscribe("CH");
pubsubCmd.psubscribe("CH2");
pubsubCmd.unsubscribe("CH");
Thread.sleep(100*5);
pubsubConn.close();
redisClient.shutdown();
}
}
2.6、客户端资源与参数配置
Lettuce 客户端的通信框架集成了 Netty 的非阻塞 IO 操作,客户端资源的设置与 Lettuce 的性能、并发和事件处理紧密相关,如果不是特别熟悉客户端参数配置,不建议在没有经验的前提下凭直觉修改默认值,保持默认配置就行。
非集群环境下,具体的配置案例如下:
publicclassLettuceMain{
publicstaticvoidmain(String[]args)throwsException{
ClientResourcesresources=DefaultClientResources.builder()
.ioThreadPoolSize(4)//I/O线程数
.computationThreadPoolSize(4)//任务线程数
.build();
RedisURIredisUri=RedisURI.builder()
.withHost("127.0.0.1").withPort(6379).withPassword("111111")
.withTimeout(Duration.of(10,ChronoUnit.SECONDS))
.build();
ClientOptionsoptions=ClientOptions.builder()
.autoReconnect(true)//是否自动重连
.pingBeforeActivateConnection(true)//连接激活之前是否执行PING命令
.build();
RedisClientclient=RedisClient.create(resources,redisUri);
client.setOptions(options);
StatefulRedisConnection<String,String>connection=client.connect();
RedisCommands<String,String>commands=connection.sync();
commands.set("name","关羽");
System.out.println(commands.get("name"));
connection.close();
client.shutdown();
resources.shutdown();
}
}
集群环境下,具体的配置案例如下:
publicclassLettuceMain{
publicstaticvoidmain(String[]args)throwsException{
ClientResourcesresources=DefaultClientResources.builder()
.ioThreadPoolSize(4)//I/O线程数
.computationThreadPoolSize(4)//任务线程数
.build();
RedisURIredisUri=RedisURI.builder()
.withHost("127.0.0.1").withPort(6379).withPassword("111111")
.withTimeout(Duration.of(10,ChronoUnit.SECONDS))
.build();
ClusterClientOptionsoptions=ClusterClientOptions.builder()
.autoReconnect(true)//是否自动重连
.pingBeforeActivateConnection(true)//连接激活之前是否执行PING命令
.validateClusterNodeMembership(true)//是否校验集群节点的成员关系
.build();
RedisClusterClientclient=RedisClusterClient.create(resources,redisUri);
client.setOptions(options);
StatefulRedisClusterConnection<String,String>connection=client.connect();
RedisAdvancedClusterCommands<String,String>commands=connection.sync();
commands.set("name","张飞");
System.out.println(commands.get("name"));
connection.close();
client.shutdown();
resources.shutdown();
}
}
2.7、线程池配置
Lettuce 连接设计的时候,就是线程安全的,所以一个连接可以被多个线程共享,同时 lettuce 连接默认是自动重连的,使用单连接基本可以满足业务需求,大多数情况下不需要配置连接池,多连接并不会给操作带来性能上的提升。
但在某些特殊场景下,比如事物操作,使用连接池会是一个比较好的方案,那么如何配置线程池呢?
publicclassLettuceMain{
publicstaticvoidmain(String[]args)throwsException{
RedisURIredisUri=RedisURI.builder()
.withHost("127.0.0.1")
.withPort(6379)
.withPassword("111111")
.withTimeout(Duration.of(10,ChronoUnit.SECONDS))
.build();
RedisClientclient=RedisClient.create(redisUri);
//连接池配置
GenericObjectPoolConfigpoolConfig=newGenericObjectPoolConfig();
poolConfig.setMaxIdle(2);
GenericObjectPool<StatefulRedisConnection<String,String>>pool=ConnectionPoolSupport.createGenericObjectPool(client::connect,poolConfig);
StatefulRedisConnection<String,String>connection=pool.borrowObject();
RedisCommands<String,String>commands=connection.sync();
commands.set("name","张飞");
System.out.println(commands.get("name"));
connection.close();
pool.close();
client.shutdown();
}
}
2.8、主从模式配置
redis 一般采用主从复制模式,搭建高可用的架构,简单的说就一个主节点,多个从节点,自动从主节点同步最新数据。
Lettuce 支持自动发现主从模式下的节点信息,然后保存到本地,具体配置如下:
publicclassLettuceMain{
publicstaticvoidmain(String[]args)throwsException{
//这里只需要配置一个节点的连接信息,不一定需要是主节点的信息,从节点也可以;可以自动发现主从节点
RedisURIuri=RedisURI.builder().withHost("192.168.31.111").withPort(6379).withPassword("123456").build();
RedisClientclient=RedisClient.create(uri);
StatefulRedisMasterReplicaConnection<String,String>connection=MasterReplica.connect(client,StringCodec.UTF8,uri);
//从节点读取数据
connection.setReadFrom(ReadFrom.REPLICA);
RedisCommands<String,String>commands=connection.sync();
commands.set("name","张飞");
System.out.println(commands.get("name"));
connection.close();
client.shutdown();
}
}
当然我们也可以手动指定集群节点来加载,具体配置如下:
publicclassLettuceMain{
publicstaticvoidmain(String[]args)throwsException{
//集群节点
List<RedisURI>uris=newArrayList();
uris.add(RedisURI.builder().withHost("192.168.31.111").withPort(6379).withPassword("111111").build());
uris.add(RedisURI.builder().withHost("192.168.31.112").withPort(6379).withPassword("111111").build());
uris.add(RedisURI.builder().withHost("192.168.31.113").withPort(6379).withPassword("111111").build());
RedisClientclient=RedisClient.create();
StatefulRedisMasterReplicaConnection<String,String>connection=MasterReplica.connect(client,StringCodec.UTF8,uris);
//从节点读取数据
connection.setReadFrom(ReadFrom.REPLICA);
RedisCommands<String,String>commands=connection.sync();
commands.set("name","张飞");
System.out.println(commands.get("name"));
connection.close();
client.shutdown();
}
}
2.9、哨兵模式配置
哨兵模式,也是 redis 实现服务高可用的一大亮点,具体配置实现如下:
publicclassLettuceMain{
publicstaticvoidmain(String[]args)throwsException{
//集群节点
List<RedisURI>uris=newArrayList();
uris.add(RedisURI.builder().withSentinel("192.168.31.111",26379).withSentinelMasterId("mymaster").withPassword("123456").build());
uris.add(RedisURI.builder().withSentinel("192.168.31.112",26379).withSentinelMasterId("mymaster").withPassword("123456").build());
uris.add(RedisURI.builder().withSentinel("192.168.31.113",26379).withSentinelMasterId("mymaster").withPassword("123456").build());
RedisClientclient=RedisClient.create();
StatefulRedisMasterReplicaConnection<String,String>connection=MasterReplica.connect(client,StringCodec.UTF8,uris);
//从节点读取数据
connection.setReadFrom(ReadFrom.REPLICA);
RedisCommands<String,String>commands=connection.sync();
commands.set("name","赵云");
System.out.println(commands.get("name"));
connection.close();
client.shutdown();
}
}
2.10、Cluster 集群模式配置
Cluster 集群模式,是之后推出的一种高可用的架构模型,主要是采用分片方式来存储数据,具体配置如下:
publicclassLettuceReactiveMain4{
publicstaticvoidmain(String[]args)throwsException{
Set<RedisURI>uris=newHashSet<>();
uris.add(RedisURI.builder().withHost("192.168.31.111").withPort(7000).withPassword("123456").build());
uris.add(RedisURI.builder().withHost("192.168.31.112").withPort(7000).withPassword("123456").build());
uris.add(RedisURI.builder().withHost("192.168.31.113").withPort(7000).withPassword("123456").build());
uris.add(RedisURI.builder().withHost("192.168.31.114").withPort(7000).withPassword("123456").build());
uris.add(RedisURI.builder().withHost("192.168.31.115").withPort(7000).withPassword("123456").build());
uris.add(RedisURI.builder().withHost("192.168.31.116").withPort(7001).withPassword("123456").build());
RedisClusterClientclient=RedisClusterClient.create(uris);
StatefulRedisClusterConnection<String,String>connection=client.connect();
RedisAdvancedClusterCommands<String,String>commands=connection.sync();
commands.set("name","关羽");
System.out.println(commands.get("name"));
//选择从节点,只读
NodeSelection<String,String>replicas=commands.replicas();
NodeSelectionCommands<String,String>nodeSelectionCommands=replicas.commands();
Executions<List<String>>keys=nodeSelectionCommands.keys("*");
keys.forEach(key->System.out.println(key));
connection.close();
client.shutdown();
}
}
三、小结
Lettuce 相比老牌的 Jedis 客户端,功能更加强大,不仅解决了线程安全的问题,还支持异步和响应式编程,支持集群,Sentinel,管道和编码器等等功能。
以上介绍的可能只是冰山一角,如果想要了解更多的信息,可以访问它的官网地址:https://lettuce.io/
到此这篇关于Redis实战之Lettuce的使用技巧详解的文章就介绍到这了,更多相关Redis Lettuce内容请搜索我们以前的文章或继续浏览下面的相关文章希望大家以后多多支持我们!