Refresh connections when getting JedisMovedDataException during using pipeline

[stillerrr]

Expected behavior

create a JedisCluster instance and only use ClusterPipeline

the slot cache would refresh automatically when cluster node changing (Scale-out or Scale-in)

Actual behavior

the slot cache do not refresh automatically when cluster node changing (Scale-out or Scale-in)

Steps to reproduce:

1. create a JedisCluster instance and only use ClusterPipeline
2. Scale-out or scale-in cluster
3. some responses would get JedisMovedDataException

Redis / Jedis Configuration

Jedis version:

Redis version:

Java version:

[stillerrr]

the root cause for this is that when pipeline do not refresh slot cache when getting JedisMovedDataException

I submit a pr for this: #3699