MXS-756: Retry reads on slave failure

If a slave fails while a non-critical read is being executed, the read is
retried on a different server. This is controlled by the new
`retry_failed_reads` option.

Only selects done that are done outside of a transaction and with
autocommit enabled are retried.

server/modules/routing/readwritesplit/readwritesplit.h

@@ -229,6 +229,7 @@ typedef struct rwsplit_config_st
                                              * to the master after a multistatement query. */
     enum failure_mode rw_master_failure_mode; /**< Master server failure handling mode.
                                                * @see enum failure_mode */
+    bool              rw_retry_failed_reads; /**< Retry failed reads on other servers */
 } rwsplit_config_t;
 
 #if defined(PREP_STMT_CACHING)