Improve Indexer bootstrap behavior

With 5.0, GSI is a topology aware service and needs to interact with ns-server for cluster management operations. This translates to being responsive to ns-server cluster management requests. Right now Indexer bootstrap is a blocking operation and due to that it doesn't respond to any cluster management request during bootstrap.

This can lead to rebalance failures when an Indexer node is booting up e.g. if a failed-over Index node is being rebalanced-in as part of recovery.

{log_entry,{1500,448908,13224},

            'ns_1@172.23.107.210',ns_orchestrator,0,

            [<<"Rebalance exited with reason {service_rebalance_failed,index,\n                                 {linked_process_died,<20283.8167.0>,\n                                     {no_connection,\"index-service_api\"}}}">>],

            [],crit,

            {{2017,7,19},{0,21,48}}},