riak mapreduce worker_limit_reached 错误

1、 

=ERROR REPORT==== 9-Jul-2013::12:26:19 ===

trends_riak_query::find 37 Error={error,

<<"{\"phase\":0,\"error\":\"[worker_limit_reached]\",\"input\":\"{<<\\\"collections\\\">>,<<\\\"1372612349298430\\\">>}\",\"type\":\"result\",\"stack\":\"[]\"}">>}

=INFO REPORT==== 9-Jul-2013::12:26:19 ===

end{<0.24396.0>}

2、调整 worker_limit 参数有额外出现了 下面错误{riak_pipe, [{worker_limit, 100000}]},

因为测试过程gc同学修改过数据源（把CB的78改成80）,所以现在不能确定

=ERROR REPORT==== 9-Jul-2013::12:26:19 ===

trends_riak_query::find 37 Error={error,<<"Error receiving outputs: normal">>}

=INFO REPORT==== 9-Jul-2013::12:26:19 ===

end{<0.24420.0>}

 

3、200进程X2没出现worker_limit错误，只是提示数据库连接池耗尽，CPU使用68%左右

单台服务器

测试代码

F = fun()-> trends_riak_query:find(collection,[]),  trends_riak_query:find(collection,[]),error_logger:info_msg("end~p~n",[{self()}]) end,

 [begin spawn(F),io:format("~p~n",[K]) end||K<-lists:seq(1,200)]. 

%% 

修改配置文件

{riak_pipe, [{worker_limit, 100000},

{worker_queue_limit, 4096000}]},