RetryStrategy not taking effect in cluster mode

Hello,

I attempted to setup a redis cluster client with exponential back off enabled but failed.

After some deep dive, I found the root cause to be that is the connection pool overrides the retryStrategy option I explicitly set in the redisOptions.

I used the following snippet to connect to the cluster.

const ioredis = require("ioredis")

cluster = new ioredis.Cluster([
        {
                host: "<redacted>",
                port: 6379,
        }
], {
        dnsLookup: (address, callback) => callback(null, address),
        redisOptions: {
                tls: {},
                retryStrategy: function (times) {
                        return times;
                },
                username:"<redacted>",
                password:"<redacted>"
        }

})

but then I get ioredis:connection skip reconnecting because retryStrategy is not a function +0ms in the debug logs.

I was able to reliably confirm that the option is overridden by the connection pool.

The reconnection would happen at the next slot refresh cycle which has a constant cadence, so my attempt to set up exponential back off on disconnection fails.

Questions :

• Is this intentional behavior ?
• What is the recommended way to set up exponential back off retry strategy on a cluster client ? (If supported)
• Should the client documentation be updated to indicate that this option is ignored for cluster clients ? or better yet an assertion can be implemented to get the issue to pop up in the development stage rather in production stages.

Thanks, Hesham