都是单机,flags(256)是什么意思?我在文档里没查到。已经做了索引。
记录例子
12384013,9,8,1434,"2015-01-06 13:20:56","2015-01-13 11:14:29","1990-01-01 00:00:
00",0,"-",0,"ntp1.cs.wisc.edu","caesar.cs.wisc.edu","128.105.39.11"
12384015,9,8,1434,"2015-01-06 13:22:55","2015-01-13 11:14:29","1990-01-01 00:00:
00",0,"-",0,"www.google.com","","2404:6800:4008:c05::"
12384016,9,8,1434,"2015-01-06 13:22:55","2015-01-13 11:14:29","1990-01-01 00:00:
00",0,"-",0,"www.google.com","","173.194.72.147"
12384019,9,8,1435,"2015-01-06 13:29:22","2015-01-13 11:14:29","1990-01-01 00:00:
00",0,"-",0,"rptuse20120814.gomonetworks.com","","204.138.26.120"
总共有7000w条记录,总共大小10gb左右
我们查询的时候用了聚集和sql两种方法
select hostname,count(hostname) as num from dns_messages group by hostname order by num limit 20
聚集方式:
db.big_data.dns.aggregate({$group:{_id:"$hostname"}}, {$sort:{hostname:1}}, {$limit:20})
查询过程当中,查看cpu开销,只有一个核在跑。查询用了半个小时,而mongodb却只用了11分钟,cpu开销也是满核的。
不知道问题到底出在什么地方