Stay organized with collections
Save and categorize content based on your preferences.
This page provides examples of Memorystore for Redis Cluster configurations for
different client libraries.
Lettuce
We recommend using Lettuce, versions 6.2.4 and later.
Lettuce
// Create RedisURI from the MRC discovery endpoint
RedisURI redisUri = RedisURI.Builder.redis(CLUSTER_DISC_EP_ADDR, CLUSTER_DISC_EP_PORT).build();
// Configure client' resources
// Configure reconnectDelay with exponential backoff and full jitter
ClientResources resources = DefaultClientResources.builder()
.reconnectDelay(Delay.fullJitter(
Duration.ofMillis(100), // minimum 100 millisecond delay
Duration.ofSeconds(5), // maximum 5 second delay
100, TimeUnit.MILLISECONDS) // 100 millisecond base
).build();
// Create a cluster client with the URI and resources
RedisClusterClient clusterClient = RedisClusterClient.create(resources, redisUri);
// Configure the topology refreshment options
// Enable periodic cluster topology updates so that the client updates the cluster topology in the intervals of
// 60 seconds
// Enable adaptive topology refresh that uses all triggers: MOVED_REDIRECT, ASK_REDIRECT,
// PERSISTENT_RECONNECTS, UNCOVERED_SLOT, UNKNOWN_NODE
// Disable dynamicRefreshSources so that only the initial seed nodes (Memorystore for Redis Cluster
// discovery endpoint) will be used as the source for topology discovery
// Enable closing stale connections when refreshing the cluster topology. This reduces the need to handle
// failed connections during command runtime.
ClusterTopologyRefreshOptions topologyRefreshOptions = ClusterTopologyRefreshOptions.builder()
.enablePeriodicRefresh(1, TimeUnit.MINUTES)
.enableAllAdaptiveRefreshTriggers()
.dynamicRefreshSources(false)
.closeStaleConnections(true)
.build();
// Configure the socket options
// Set connectTimeout based on your application requirements and workload
// Enable TCP keepAlive to reduce the need to handle failed connections during command runtime
SocketOptions socketOptions = SocketOptions.builder()
.connectTimeout(CONNECT_TIMEOUT)
.keepAlive(true)
.build();
// Configure the client options
// Enable AutoReconnect when connection is lost
// Set nodeFilter to filter out failed nodes from the topology
// Disable validateClusterNodeMembership to allow redirecting commands to newly added nodes
clusterClient.setOptions(ClusterClientOptions.builder()
.topologyRefreshOptions(topologyRefreshOptions)
.socketOptions(socketOptions)
.autoReconnect(true)
.nodeFilter(it ->
! (it.is(RedisClusterNode.NodeFlag.FAIL)
|| it.is(RedisClusterNode.NodeFlag.EVENTUAL_FAIL)
|| it.is(RedisClusterNode.NodeFlag.NOADDR)))
.validateClusterNodeMembership(false)
.build());
// Create a connection pool
GenericObjectPool<StatefulRedisClusterConnection<String, String> pool = ConnectionPoolSupport.createGenericObjectPool(() -> clusterClient.connect(), new GenericObjectPoolConfig());
pool.setMaxTotal(MAX_CONNECTIONS_IN_CONNECTION_POOL);
// Get a connection from the connection pool
StatefulRedisClusterConnection<String, String> connection = pool.borrowObject();
// Get a cluster sync command and call 'set'
RedisAdvancedClusterCommands<String, String> syncCommands = connection.sync();
syncCommands.set(key, value);
redis-py
We recommend using redis-py, versions 5.1
and later.
redis-py
// Create a cluster client. A connection pool is used internally.
cluster_client = redis.RedisCluster(
host=CLUSTER_DISC_EP_ADDR,
port=CLUSTER_DISC_EP_PORT,
retry=Retry(EqualJitterBackoff(), 3))
cluster_client.set('key', 'value')
PHPRedis
We recommend using PHPRedis, versions
5.0.0 and later.
PHPRedis
// In redis.ini
redis.clusters.seeds = "cluster[]=CLUSTER_DISC_EP_ADDR:CLUSTER_DISC_EP_PORT"
redis.clusters.cache_slots = "cluster=1"
redis.clusters.persistent = "cluster=1"
$cluster = new RedisCluster('cluster');
$cluster->set("key", "value");
Redisson
We recommend using Redisson, versions 3.20.1 and later.
[[["Easy to understand","easyToUnderstand","thumb-up"],["Solved my problem","solvedMyProblem","thumb-up"],["Other","otherUp","thumb-up"]],[["Hard to understand","hardToUnderstand","thumb-down"],["Incorrect information or sample code","incorrectInformationOrSampleCode","thumb-down"],["Missing the information/samples I need","missingTheInformationSamplesINeed","thumb-down"],["Other","otherDown","thumb-down"]],["Last updated 2025-08-29 UTC."],[],[],null,["# Client library code samples\n\nThis page provides examples of Memorystore for Redis Cluster configurations for\ndifferent client libraries.\n\nLettuce\n-------\n\nWe recommend using [Lettuce](https://lettuce.io/), versions 6.2.4 and later.\n**Caution** : If you use Lettuce to connect to a Memorystore for Redis Cluster instance, then change the `validateClusterNodeMembership` parameter to `false`. Otherwise, when the topology changes, you might get `unknownPartition` errors. \n\n### Lettuce\n\n```\n\n// Create RedisURI from the MRC discovery endpoint\nRedisURI redisUri = RedisURI.Builder.redis(CLUSTER_DISC_EP_ADDR, CLUSTER_DISC_EP_PORT).build();\n\n// Configure client' resources\n// Configure reconnectDelay with exponential backoff and full jitter\nClientResources resources = DefaultClientResources.builder()\n .reconnectDelay(Delay.fullJitter(\n Duration.ofMillis(100), // minimum 100 millisecond delay\n Duration.ofSeconds(5), // maximum 5 second delay\n 100, TimeUnit.MILLISECONDS) // 100 millisecond base\n).build();\n\n// Create a cluster client with the URI and resources\nRedisClusterClient clusterClient = RedisClusterClient.create(resources, redisUri);\n\n// Configure the topology refreshment options\n// Enable periodic cluster topology updates so that the client updates the cluster topology in the intervals of\n// 60 seconds\n// Enable adaptive topology refresh that uses all triggers: MOVED_REDIRECT, ASK_REDIRECT,\n// PERSISTENT_RECONNECTS, UNCOVERED_SLOT, UNKNOWN_NODE\n// Disable dynamicRefreshSources so that only the initial seed nodes (Memorystore for Redis Cluster\n// discovery endpoint) will be used as the source for topology discovery\n// Enable closing stale connections when refreshing the cluster topology. This reduces the need to handle\n// failed connections during command runtime.\nClusterTopologyRefreshOptions topologyRefreshOptions = ClusterTopologyRefreshOptions.builder()\n .enablePeriodicRefresh(1, TimeUnit.MINUTES)\n .enableAllAdaptiveRefreshTriggers()\n .dynamicRefreshSources(false)\n .closeStaleConnections(true)\n .build();\n\n// Configure the socket options\n// Set connectTimeout based on your application requirements and workload\n// Enable TCP keepAlive to reduce the need to handle failed connections during command runtime\nSocketOptions socketOptions = SocketOptions.builder()\n .connectTimeout(CONNECT_TIMEOUT)\n .keepAlive(true)\n .build();\n\n// Configure the client options\n// Enable AutoReconnect when connection is lost\n// Set nodeFilter to filter out failed nodes from the topology\n// Disable validateClusterNodeMembership to allow redirecting commands to newly added nodes\n\nclusterClient.setOptions(ClusterClientOptions.builder()\n .topologyRefreshOptions(topologyRefreshOptions)\n .socketOptions(socketOptions)\n .autoReconnect(true)\n .nodeFilter(it -\u003e\n ! (it.is(RedisClusterNode.NodeFlag.FAIL)\n || it.is(RedisClusterNode.NodeFlag.EVENTUAL_FAIL)\n || it.is(RedisClusterNode.NodeFlag.NOADDR)))\n .validateClusterNodeMembership(false)\n .build());\n\n// Create a connection pool\nGenericObjectPool\u003cStatefulRedisClusterConnection\u003cString, String\u003e pool = ConnectionPoolSupport.createGenericObjectPool(() -\u003e clusterClient.connect(), new GenericObjectPoolConfig());\npool.setMaxTotal(MAX_CONNECTIONS_IN_CONNECTION_POOL);\n\n// Get a connection from the connection pool\nStatefulRedisClusterConnection\u003cString, String\u003e connection = pool.borrowObject();\n\n// Get a cluster sync command and call 'set'\nRedisAdvancedClusterCommands\u003cString, String\u003e syncCommands = connection.sync();\nsyncCommands.set(key, value);\n\n```\n\nredis-py\n--------\n\nWe recommend using [redis-py](https://github.com/redis/redis-py), versions 5.1\nand later. \n\n### redis-py\n\n```\n\n// Create a cluster client. A connection pool is used internally.\ncluster_client = redis.RedisCluster(\n host=CLUSTER_DISC_EP_ADDR,\n port=CLUSTER_DISC_EP_PORT,\n retry=Retry(EqualJitterBackoff(), 3))\n\ncluster_client.set('key', 'value')\n\n```\n\nPHPRedis\n--------\n\nWe recommend using [PHPRedis](https://github.com/phpredis/phpredis), versions\n5.0.0 and later. \n\n### PHPRedis\n\n```\n\n// In redis.ini\nredis.clusters.seeds = \"cluster[]=CLUSTER_DISC_EP_ADDR:CLUSTER_DISC_EP_PORT\"\nredis.clusters.cache_slots = \"cluster=1\"\nredis.clusters.persistent = \"cluster=1\"\n\n$cluster = new RedisCluster('cluster');\n$cluster-\u003eset(\"key\", \"value\");\n\n```\n\nRedisson\n--------\n\nWe recommend using [Redisson](https://redisson.org/), versions 3.20.1 and later. \n\n### Redisson\n\n```\n\nConfig config = new Config();\nClusterServersConfig clusterConfig = config.useClusterServers()\n .addNodeAddress(\"redis://CLUSTER_DISC_EP_ADDR:CLUSTER_DISC_EP_PORT\")\n .setScanInterval(60000) // cluster topology periodic refresh\n .setMasterConnectionMinimumIdleSize(1)\n .setMasterConnectionPoolSize(10)\n .setSlaveConnectionMinimumIdleSize(1)\n .setSlaveConnectionPoolSize(10)\n .setIdleConnectionTimeout(60000)\n .setPingConnectionInterval(0)\n .setKeepAlive(true)\n .setTcpNoDelay(true)\n\nRedissonClient redisson = Redisson.create(config);\nRBucket\u003cString\u003e bucket = redisson.getBucket(\"key\");\nbucket.set(\"value\");\n\n```\n\ngo-redis\n--------\n\nWe recommend using [go-redis](https://github.com/redis/go-redis), versions\n9.11.0 and later. \n\n### go-redis\n\n```\n\nclusterClient = redis.NewClusterClient(&redis.ClusterOptions{\n Addrs: []string{\"CLUSTER_DISC_EP_ADDR:CLUSTER_DISC_EP_PORT\"},\n // PoolSize applies per cluster node and not for the whole cluster.\n PoolSize: 10,\n ConnMaxIdleTime: 60 * time.Second,\n MinIdleConns: 1,\n})\n\nctx := context.Background()\n\nerr := clusterClient.Set(ctx, \"key\", \"value\", 0).Err()\n\n```"]]