Issue blocking upgrading above 2.2.4

It looks like you’re using WP_REDIS_HOST with a real Redis Cluster.

In that case you must use WP_REDIS_CLUSTER instead. Can you post your full configuration please?

Aha! That makes sense. Here’s what I have:

define( ‘WP_REDIS_HOST’, ‘redacted.h07z7u.clustercfg.use1.cache.amazonaws.com’ );
define( ‘WP_REDIS_PORT’, 6379 );
define( ‘WP_REDIS_DATABASE’, 2 );
define( ‘WP_REDIS_PREFIX’, ‘app-redacted-com’ );

So, is it just that I change WP_REDIS_HOST to WP_REDIS_CLUSTER?

It might take a little bit of fiddling. I see you’re using WP_REDIS_DATABASE=2 which does not work on Redis Cluster only 0 is available.

If you really like like it, I’d say spin up a new cluster with cluster_mode disabled. AWS Elasticache offers that.

Otherwise try this config:

define( 'WP_REDIS_CLIENT', 'phpredis' );

define( 'WP_REDIS_CLUSTER', [
    'tcp://redacted.h07z7u.clustercfg.use1.cache.amazonaws.com:6379?alias=primary1',
] );

Got it. For immediate resolution, I’ve spun up a new one with cluster_mode disabled.

I’m curious on your take for ideal setup. Right now I have 14 WP sites which will probably grow to 20 in the next 12 months. I was setting up multiple databases per redis so that I didn’t need to spin up a separate redis server for each site.

Is it best practice to spin up a separate one for each site? If so, any thoughts on sizing? Trying to avoid unnecessary spending.

• This reply was modified 10 months ago by bradjon.

You can use hundreds of sites with a single Redis instance, just skip the dumb Cluster Mode, because with it you’re limited to 1 database. With it disabled you can give each site its own database and even credentials if you like. That’s the ideal setup IMO.