How to use lettuce in the right way

Bug Report

When the server is under a lot of pressure, call lettucePool.acquire().get(1, TimeUnit.SECONDS) and the program has the following exception.

Exception

java.util.NoSuchElementException: Pool exhausted
...
Other exception stack information

Java code

// config  RedisClusterClient
ClientResources clientResources = DefaultClientResources.builder()
                                     .ioThreadPoolSize(3 * Runtime.getRuntime().availableProcessors())
                                     .computationThreadPoolSize(4)
                                     .reconnectDelay(Delay.constant(Duration.ofSeconds(1)))
                                     .build();

RedisClusterClient clusterClient = RedisClusterClient.create(clientResources, redisURIS);
ClusterTopologyRefreshOptions topologyRefreshOptions = ClusterTopologyRefreshOptions
                                                         .builder()
                                                         .enableAdaptiveRefreshTrigger(ClusterTopologyRefreshOptions.RefreshTrigger.MOVED_REDIRECT,ClusterTopologyRefreshOptions.RefreshTrigger.PERSISTENT_RECONNECTS)
                                                         .adaptiveRefreshTriggersTimeout(Duration.ofMinutes(3))
                                                         .build();
clusterClient.setOptions(ClusterClientOptions.builder()
                                                         .topologyRefreshOptions(topologyRefreshOptions)
                                                         .autoReconnect(true)
                                                         .pingBeforeActivateConnection(true)
                                                         .build());

// config lettucePool
// maxIdle = 8
// minIdle = 1
// maxActive = 8
// testOnCreate = true
final BoundedPoolConfig boundedPoolConfig = BoundedPoolConfig
                .builder()
                .minIdle(minIdle)
                .maxIdle(maxIdle)
                .maxTotal(maxActive)
                .testOnCreate(testOnCreate)
                .testOnAcquire(true).build();
final BoundedAsyncPool<StatefulRedisClusterConnection<String, String>> lettucePool =
                AsyncConnectionPoolSupport.createBoundedObjectPool(() -> clusterClient.connectAsync(StringCodec.UTF8), boundedPoolConfig);
===============================================================


// way of use
@Autowired
BoundedAsyncPool<StatefulRedisClusterConnection<String, String>> lettucePool;
// way of use
final StatefulRedisClusterConnection<String, String> connection = lettucePool.acquire().get(1, TimeUnit.SECONDS);

[mp911de]

Lettuce connection are designed to be long-lived and they are thread-safe. If you use Redis Cluster with simple GET and SET commands, then you don’t need pooling at all. The only thing to do is injecting StatefulRedisClusterConnection into your code and simply using it.

[10119255]

Lettuce connection are designed to be long-lived and they are thread-safe. If you use Redis Cluster with simple GET and SET commands, then you don’t need pooling at all. The only thing to do is injecting StatefulRedisClusterConnection into your code and simply using it.

Thank you very much for your reply. I will try your suggestion.