Since this FSM has no timeout and the remote index FSM has no timeout

by default, we can get stuck with both FSMs partially completed and
holding on to a lot of memory.

This fix adds a timeout to the remote FSM so it should terminate in a
timely fashion, but adds a timeout to the connection request, just in
case the other end stops talking to us for some reason.

src/riak_cs_list_objects_fsm_v2.erl

@@ -331,7 +331,9 @@ make_2i_request(RiakcPid, State=#state{req=?LOREQ{name=BucketName},
                                                       os:timestamp()),
     Opts = [{max_results, BatchSize},
             {start_key, StartKey},
-            {end_key, EndKey}],
+            {end_key, EndKey},
+            {timeout, timer:minutes(1)},
+            {call_timeout, timer:minutes(1) + timer:seconds(10)}],
     FoldResult = riakc_pb_socket:cs_bucket_fold(RiakcPid,
                                                 ManifestBucket,
                                                 Opts),