Improve Cluster mode documentation, especially in AWS ElastiCache with TLS

[rarecrumb]

It seems very common to have issues connecting to AWS ElastiCache in Cluster mode, with TLS enabled.

@luin can you provide examples showing how to use ioredis to connect to Redis Cluster? Extra helpful if it can actually be tested against AWS ElastiCache in Cluster mode with TLS enabled/disabled, password auth enabled/disabled.

ioredis version: v5.3.2
Redis Cluster Engine version: 7.0.7 - TLS Enabled, no password

  static async createCluster(
    connectionString: ClusterNode[],
    ioRedisClusterOpts?: ClusterOptions
  ): Promise<RedisClient> {
    const opts: ClusterOptions = {
      ...ioRedisClusterOpts,
      clusterRetryStrategy() {
        const delay = Math.min(10 * 50, 2000);
        return delay;
      },
      scaleReads: "slave",
      lazyConnect: true,
      slotsRefreshInterval: 3000,
      slotsRefreshTimeout: 10000,
      enableOfflineQueue: false,
      dnsLookup: (address, callback) => callback(null, address),
      enableReadyCheck: true,
      redisOptions: {
        family: 6,
        tls: {},
      },
    };

    const client = new IORedis.Cluster(connectionString, opts);

    await client.connect();
    return new RedisClient(client);

Errors:

2023-09-12T20:22:37.357Z ioredis:connection error: Error: getaddrinfo ENOTFOUND x.us-east-1-xxx-xxx.xxx.use1.cache.amazonaws.com    
  at GetAddrInfoReqWrap.onlookup [as oncomplete] (node:dns:108:26)    
  at GetAddrInfoReqWrap.callbackTrampoline (node:internal/async_hooks:130:17) 
{  errno: -3007,  code: 'ENOTFOUND',  syscall: 'getaddrinfo',  hostname: 'xxx.us-east-1-xxx-xxx.xxx.use1.cache.amazonaws.com'}

Related issues:

• Refreshing slots cache infinite loop issue #1766
• ioredis client not able to connect in cluster mode with TLS enabled #1800
• question with redis cluster #1790
• Weird issue with connection dropping with cluster pub/sub #1772
• Cluster client couldn't reconnect because retryStrategy is not set, but the TS types state that's not allowed #1749
• Unhandled error event: Error: getaddrinfo ENOTFOUND - AWS ElasticCache #1576
• Unable to connect #1483
• ClusterAllFailedError: Failed to refresh slots cache. with Cluster mode and TLS enabled #1454
• Cluster retry strategy and reconnection #1437
• Issue related to ClusterAllFailedError #1415
• Dockerized Redis cluster is failing with ElastiCache configuration #1384
• Connect aws cluster reidis failed #1343
• Infinity loop of connect event #1341
• Failed to refresh slot: docker #1337
• How to get an error if connect failed while still retrying on next attempt (cluster mode) #1268
• Can't connect to kubernetes #1151
• [Cluster] TLS-SNI: different servername per node #1114 - [TLS] closes #1114 - add tlsSni option #1137
• ClusterAllFailedError: Failed to refresh slots cache #1103
• Understanding clusterRetryStrategy (after Failed to refresh slots cache) #1062
• ClusterAllFailedError: Failed to refresh slots cache. #1003
  ... and more

[rarecrumb]

These are the options that finally ended up working for me with ElastiCache with Cluster mode and TLS enabled:

      clusterRetryStrategy() {
        const delay = Math.min(10 * 50, 2000);
        return delay;
      },
      scaleReads: 'slave',
      lazyConnect: true,
      slotsRefreshInterval: 60000,
      slotsRefreshTimeout: 10000,
      enableOfflineQueue: true,
      enableReadyCheck: true,
      redisOptions: {
        tls:
          process.env.NODE_ENV === "production"
            ? {
                checkServerIdentity: (/*host, cert*/) => {
                  // skip certificate hostname validation
                  return undefined;
                },
              }
            : undefined,
        commandTimeout: 5000,
        connectTimeout: 10000,
        lazyConnect: true,
        enableReadyCheck: true,
        enableAutoPipelining: true,
        enableOfflineQueue: true,
        noDelay: true,
        keepAlive: 1000
      }

[nrathi]

@ rarecrumb how do you connect to the cluster when developing locally? I'm trying to tunnel via SSH but not sure if that's messing things up.

[blshukla]

We were facing a similar issue. This answer from Stack Overflow worked well for us - https://stackoverflow.com/questions/76409668/how-to-connect-to-amazon-memorydb-for-redis-from-node-running-in-ec2-ecs

[sonthanhdan]

These are the options that finally ended up working for me with ElastiCache with Cluster mode and TLS enabled:

      clusterRetryStrategy() {
        const delay = Math.min(10 * 50, 2000);
        return delay;
      },
      scaleReads: 'slave',
      lazyConnect: true,
      slotsRefreshInterval: 60000,
      slotsRefreshTimeout: 10000,
      enableOfflineQueue: true,
      enableReadyCheck: true,
      redisOptions: {
        tls:
          process.env.NODE_ENV === "production"
            ? {
                checkServerIdentity: (/*host, cert*/) => {
                  // skip certificate hostname validation
                  return undefined;
                },
              }
            : undefined,
        commandTimeout: 5000,
        connectTimeout: 10000,
        lazyConnect: true,
        enableReadyCheck: true,
        enableAutoPipelining: true,
        enableOfflineQueue: true,
        noDelay: true,
        keepAlive: 1000
      }

thanks @rarecrumb I tried with your config it working fine.