两台 MongoDB 的复制集，怎么做查询负载均衡？

This topic created in 3857 days ago, the information mentioned may be changed or developed.

read_preference 的选项里面，

PRIMARY: Queries are sent to the primary of the replica set.
PRIMARY_PREFERRED: Queries are sent to the primary if available, otherwise a secondary.
SECONDARY: Queries are distributed among secondaries. An error is raised if no secondaries are available.
SECONDARY_PREFERRED: Queries are distributed among secondaries, or the primary if no secondary is available.
NEAREST: Queries are distributed among all members.

最符合的似乎是 NEAREST ，但是实测只是读了本机的 MongoDB ，另外一台机器的完全空闲。。。

难道我要手动建立 2 个 MongoClient 然后随机查询某个库么。。。

queries

PRIMARY

are

Secondary

6 replies • 2015-10-21 14:55:19 +08:00

9hills

Oct 21, 2015 via iPhone

Rep 推荐三台

typcn

Oct 21, 2015

查询负载均衡当然是在程序端做

从连接池取连接的时候根据两台数据库的繁忙程序随机选择

dingyaguang117

Oct 21, 2015

@typcn 根据繁忙这个 mongodb 怎么做？

dingyaguang117

Oct 21, 2015

@9hills 是的，但是现在只有两台，官方并没有提到过这种情况， sf 上还有人直接说不要这样做。。。

不过我想的也许可以从架构方面充分利用这两台机器。比如用 nginx 做分发，跑 2 个 web server,每个都 nearest

9hills

Oct 21, 2015

@dingyaguang117 我感觉你的文档是不是有问题啊，我刚查了下，对 nearest 的定义和你那个不同：

https://docs.mongodb.org/manual/reference/read-preference/

nearest Operations read from member of the replica set with the least network latency, irrespective of the member ’ s type.

继续看： https://docs.mongodb.org/manual/core/read-preference-mechanics/#replica-set-read-preference-behavior-nearest

nearest 是默认从延时<15ms 的节点读取。所以如果两台机器有一台延时>15ms ，那么就会只落在一台了

有个办法是修改这个 15ms 的值，如果两台的延时都落在这个时间内，那么是两台随机读

dingyaguang117

Oct 21, 2015

@9hills 因为 webserver 部署与 primary mongodb 在一台机器上，所以延迟必然是本机最低了