Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

MongoDB分片集群部署实战 #9

Open
chemdemo opened this issue Dec 15, 2014 · 0 comments
Open

MongoDB分片集群部署实战 #9

chemdemo opened this issue Dec 15, 2014 · 0 comments

Comments

@chemdemo
Copy link
Owner

本篇记录MongoDB高可用模式部署步骤,其他部署方式见上一篇

首先准备机器,我这里是在公司云平台创建了三台DB server,ip分别是10.199.144.84,10.199.144.89,10.199.144.90。

分别安装mongodb最新稳定版本:

wget https://fastdl.mongodb.org/linux/mongodb-linux-x86_64-2.4.12.tgz
tar -xzvf mongodb-linux-x86_64-2.4.12.tgz
mv mongodb-linux-x86_64-2.4.12 /usr/lib

做个软连接或者按照官方的做法把mongo shell都添加到环境变量:

ln -s /usr/lib/mongodb-linux-x86_64-2.4.12/bin/mongo /usr/bin/mongo
ln -s /usr/lib/mongodb-linux-x86_64-2.4.12/bin/mongos /usr/bin/mongos
ln -s /usr/lib/mongodb-linux-x86_64-2.4.12/bin/mongod /usr/bin/mongod

分别创建存储数据的目录:

mkdir -p /data/mongodb && cd /data/mongodb/ && mkdir -p conf/data conf/log mongos/log shard{1..3}/data shard{1..3}/log

分别配置启动config服务器:

mongod --configsvr --dbpath /data/mongodb/conf/data --port 27100 --logpath /data/mongodb/conf/confdb.log --fork --directoryperdb

确保config服务都启动之后,启动路由服务器(mongos):

mongos --configdb 10.199.144.84:27100,10.199.144.89:27100,10.199.144.90:27100 --port 27000 --logpath /data/mongodb/mongos/mongos.log --fork

分别配置启动各个分片副本集,这里副本集名分别叫shard1shard2shard3

mongod --shardsvr --replSet shard1 --port 27001 --dbpath /data/mongodb/shard1/data --logpath /data/mongodb/shard1/log/shard1.log --directoryperdb --fork

mongod --shardsvr --replSet shard2 --port 27002 --dbpath /data/mongodb/shard2/data --logpath /data/mongodb/shard2/log/shard2.log --directoryperdb --fork

mongod --shardsvr --replSet shard3 --port 27003 --dbpath /data/mongodb/shard3/data --logpath /data/mongodb/shard3/log/shard3.log --directoryperdb --fork

接下来配置副本集,假设使用如下的架构,每台物理机都有一个主节点,一个副本节点和一个仲裁节点:
mongodb-shard-architecture

配置shard1(登陆84,没有显式指定主节点时,会选择登陆的机器为主节点):

mongo --port 27001
use admin
rs.initiate({
    _id: 'shard1',
    members: [
        {_id: 84, host: '10.199.144.84:27001'},
        {_id: 89, host: '10.199.144.89:27001'},
        {_id: 90, host: '10.199.144.90:27001', arbiterOnly: true}
    ]
});

配置shard2(登陆89):

mongo --port 27001
use admin
rs.initiate({
    _id: 'shard2',
    members: [
        {_id: 84, host: '10.199.144.84:27002', arbiterOnly: true},
        {_id: 89, host: '10.199.144.89:27002'},
        {_id: 90, host: '10.199.144.90:27002'}
    ]
});

配置shard3(登陆90):

mongo --port 27001
use admin
rs.initiate({
    _id: 'shard3',
    members: [
        {_id: 84, host: '10.199.144.84:27002'},
        {_id: 89, host: '10.199.144.89:27002', arbiterOnly: true},
        {_id: 90, host: '10.199.144.90:27002'}
    ]
});

下面设置路由到分片集群配置,随便登陆一台机器,假设是84:

mongo --port 27000
use admin
db.runCommand({addShard: 'shard1/10.199.144.84:27001,10.199.144.89:27001,10.199.144.90:27001'});
db.runCommand({addShard: 'shard2/10.199.144.84:27002,10.199.144.89:27002,10.199.144.90:27002'});
db.runCommand({addShard: 'shard3/10.199.144.84:27003,10.199.144.89:27003,10.199.144.90:27003'});

查看配置好的shard:

mongo --port 27000
use admin
db.runCommand({listshards: 1});

结果:

{
    "shards" : [
        {
            "_id" : "shard1",
            "host" : "shard1/10.199.144.84:27001,10.199.144.89:27001"
        },
        {
            "_id" : "shard2",
            "host" : "shard2/10.199.144.89:27002,10.199.144.90:27002"
        },
        {
            "_id" : "shard3",
            "host" : "shard3/10.199.144.90:27003,10.199.144.84:27003"
        }
    ],
    "ok" : 1
}

其中仲裁(ARBITER)节点没有列出来。

下面测试分片:

mongo --port 27000
use admin
db.runCommand({enablesharding: 'dbtest'});
db.runCommand({shardcollection: 'dbtest.coll1', key: {id: 1}});
use dbtest;
for(var i=0; i<10000; i++) db.coll1.insert({id: i, s: 'str_' + i});

如果dbtest已经存在,需要确保它已经以id建立了索引!

过上一段时间之后,运行db.coll1.stats()显式分片状态:

{
    "sharded" : true,
    "ns" : "dbtest.coll1",
    "count" : 10000,
    ...
    "shards" : {
        "shard1" : {
            "ns" : "dbtest.coll1",
            "count" : 0,
            "size" : 0,
            ...
        },
        "shard2" : {
            "ns" : "dbtest.coll1",
            "count" : 10000,
            "size" : 559200,
            ...
        }
    }
    ...
}

可以看到,这里分片已经生效,只是分配不均匀,所有的数据都存在了shard2中了。分片key的选择策略可以参考官方文档。在2.4版本中,使用hashed shard key算法保证文档均匀分布:

mongo --port 27000
use admin
sh.shardCollection('dbtest.coll1', {id: 'hashed'});

使用hashed算法之后,做同样的测试,插入的数据基本均匀分布:

{
    "sharded" : true,
    "ns" : "dbtest.coll1",
    "count" : 10000,
    ...
    "shards" : {
        "shard1" : {
            "ns" : "dbtest.coll1",
            "count" : 3285,
            "size" : 183672,
            ...
        },
        "shard2" : {
            "ns" : "dbtest.coll1",
            "count" : 3349,
            "size" : 187360,
            ...
        },
        "shard3" : {
            "ns" : "dbtest.coll1",
            "count" : 3366,
            "size" : 188168,
            ...
        }
    }
}

更多资料,请参考MongoDB Sharding

在应用程序里,使用MongoClient创建db连接:

MongoClient.connect('mongodb://10.199.144.84:27000,10.199.144.89:27000,10.199.144.90:27000/dbtest?w=1', function(err, db) {
    ;
});
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant