Fix cluster reconnect example

[pendula95]

Motivation: Current example shows bad practice that can lead to dangling connections and concurrent issues.
It should tackle the issue mentioned #255

[pmlopes]

I've cherry picked the commit to master as there's too much noise in the PR. The commit retains full attribution.

790475d

[koby08]

@pmlopes @pendula95 I believe that the fixed version of the example will still suffer from a connection leak. The code attempts to call close() on RedisConnection, which (if I recall correctly) doesn't close the connection pool. You have to call close() on the Redis itself (the one returned from createClient)

See #255

[pendula95]

I think you are right. Looking at my own implementations in projects I see I use ((RedisClusterConnection) redisConnection.getDelegate()).close(); which will call

vertx-redis-client/src/main/java/io/vertx/redis/client/impl/RedisClusterConnection.java

Line 522 in 3290a06

public Future<Void> close() {

The example should be changed to feature this.