PR 1249, 1251, 1253, 1256 applied here, build for PulsePoint

Issue #1252:

1. New special exception for “No reachable nodes”
2. Fixed situation when simple connection exception with tryRandomNode=true proceed as just no reachable nodes
3. Don’t try random node with immediate initiate cluster renewal after that if we got ConnectionTimeout from proper slot jedis
4. Fix rewriting real root cause JedisConnectionException with JedisClusterMaxRedirectionsException
(cherry picked from commit c567161)