Group flush not working on AWS Elasticache Serverless

@tillkruess got it. Weird that everything still seems to work properly with WP_REDIS_HOST. I opened a ticket with AWS to figure out the cluster endpoints to use. I’ll follow up once I have a response.

@tillkruess I got back on this today. I finally figured out that the cluster mode doesn’t pass through SSL options and this is why I was having issues.

I updated this block to look like this (bolded parts new):

$args = [
    'cluster' => $this->build_cluster_connection_array(),
    'timeout' => $parameters['timeout'],
    'read_timeout' => $parameters['read_timeout'],
    'persistent' => $parameters['persistent'],
    'password' => null,
    'ssl' => null,
];

if ( isset( $parameters['password'] ) && version_compare( $version, '4.3.0', '>=' ) ) {
    $args['password'] = $parameters['password'];
}

if ( defined('WP_REDIS_SSL_CONTEXT')) {
    $args['ssl'] = WP_REDIS_SSL_CONTEXT;
}

I’m now running this config and can run in cluster mode successfully:

define('WP_REDIS_CLIENT', 'phpredis');
define('WP_REDIS_CLUSTER', [
  'tls://' . getenv('WP_REDIS_HOST') . ':6379?alias=node-01',
]);
define('WP_REDIS_SSL_CONTEXT', [
  'verify_peer' => false,
]);

However wp_cache_flush_group('transient') for instance still returns false.

Aha! Would you be up for opening a PR on GitHub, so it’s merged?