1. MongoDB分片+副本集
健壮的集群方案
多个配置服务器 多个mongos服务器 每个片都是副本集 正确设置w
架构图
说明： 1. 此实验环境在一台机器上通过不同port和dbpath实现启动不同的mongod实例 2. 总的9个mongod实例，分别做成shard1、shard2、shard3三组副本集，每组1主2从 3. Mongos进程的数量不限，建议把mongos配置在每个应用服务器本机上，这样每个应用服务器就与自身的mongos进行通信，如果服务器不工作了，并不会影响其他的应用服务器与其自己的mongos通信 4. 此实验模拟2台应用服务器（2个mongos服务） 5. 生产环境中每个片都应该是副本集，这样单个服务器坏了，才不会导致片失效 |
部署环境
创建相关目录
~]# mkdir -p /data/mongodb/shard{1,2,3}/node{1,2,3}
~]# mkdir -p /data/mongodb/shard{1,2,3}/logs
~]# ls /data/mongodb/shard*
/data/mongodb/shard1:
logs node1 node2 node3
/data/mongodb/shard2:
logs node1 node2 node3
/data/mongodb/shard3:
logs node1 node2 node3
~]# mkdir -p /data/mongodb/config/logs
~]# mkdir -p /data/mongodb/config/node{1,2,3}
~]# ls /data/mongodb/config/
logs node1 node2 node3
~]# mkdir -p /data/mongodb/mongos/logs
启动配置服务
Config server | /data/mongodb/config/node1 | /data/mongodb/config/logs/node1.log | 10000 |
/data/mongodb/config/node2 | /data/mongodb/config/logs/node2.log | 20000 |
/data/mongodb/config/node3 | /data/mongodb/config/logs/node3.log | 30000 |
#按规划启动3个：跟启动单个配置服务一样，只是重复3次
~]# mongod --dbpath /data/mongodb/config/node1 --logpath /data/mongodb/config/logs/node1.log --logappend --fork --port 10000
~]# mongod --dbpath /data/mongodb/config/node2 --logpath /data/mongodb/config/logs/node2.log --logappend --fork --port 20000
~]# mongod --dbpath /data/mongodb/config/node3 --logpath /data/mongodb/config/logs/node3.log --logappend --fork --port 30000
~]# ps -ef|grep mongod|grep -v grep
root 3983 1 0 11:10 ? 00:00:03 /usr/local/mongodb/bin/mongod --dbpath /data/mongodb/config/node1 --logpath /data/mongodb/config/logs/node1.log --logappend --fork --port 10000
root 4063 1 0 11:13 ? 00:00:02 /usr/local/mongodb/bin/mongod --dbpath /data/mongodb/config/node2 --logpath /data/mongodb/config/logs/node2.log --logappend --fork --port 20000
root 4182 1 0 11:17 ? 00:00:03 /usr/local/mongodb/bin/mongod --dbpath /data/mongodb/config/node3 --logpath /data/mongodb/config/logs/node3.log --logappend --fork --port 30000
启动路由服务
Mongos server | —— | /data/mongodb/mongos/logs/node1.log | 40000 |
—— | /data/mongodb/mongos/logs/node2.log | 50000 |
#mongos的数量不受限制，通常应用一个服务器运行一个mongos
~]#mongos --port 40000 --configdb 192.168.211.217:10000,192.168.211.217:20000,192.168.211.217:30000 --logpath /data/mongodb/mongos/logs/mongos1.log --logappend --fork
~]#mongos --port 50000 --configdb 192.168.211.217:10000,192.168.211.217:20000,192.168.211.217:30000 --logpath /data/mongodb/mongos/logs/mongos2.log --logappend –fork
~]# ps -ef|grep mongos|grep -v grep
root 4421 1 0 11:29 ? 00:00:00 /usr/local/mongodb/bin/mongos --port 40000 --configdb 192.168.211.217:10000,192.168.211.217:20000,192.168.211.217:30000 --logpath /data/mongodb/mongos/logs/mongos1.log --logappend --fork
root 4485 1 0 11:29 ? 00:00:00 /usr/local/mongodb/bin/mongos --port 50000 --configdb 192.168.211.217:10000,192.168.211.217:20000,192.168.211.217:30000 --logpath /data/mongodb/mongos/logs/mongos2.log --logappend --fork
配置副本集
按规划，配置启动shard1、shard2、shard3三组副本集
#此处以shard1为例说明配置方法
#启动三个mongod进程
~]#mongod --replSet shard1 --dbpath /data/mongodb/shard1/node1 --logpath /data/mongodb/shard1/logs/node1.log --logappend --fork --port 10001
~]#mongod --replSet shard1 --dbpath /data/mongodb/shard1/node2 --logpath /data/mongodb/shard1/logs/node2.log --logappend --fork --port 10002
~]#mongod --replSet shard1 --dbpath /data/mongodb/shard1/node3 --logpath /data/mongodb/shard1/logs/node3.log --logappend --fork --port 10003
#初始化Replica Set:shard1
~]# /usr/local/mongodb/bin/mongo --port 10001
MongoDB shell version: 2.6.6
connecting to: 127.0.0.1:10001/test
> use admin
switched to db admin
> rsconf={
... "_id" : "shard1",
... "members" : [
... {
... "_id" : 0,
... "host" : "192.168.211.217:10001"
... }
... ]
... }
{
"_id" : "shard1",
"members" : [
{
"_id" : 0,
"host" : "192.168.211.217:10001"
}
]
}
> rs.initiate(rsconf)
{
"info" : "Config now saved locally. Should come online in about a minute.",
"ok" : 1
}
> rs.add("192.168.211.217:10002")
{ "ok" : 1 }
shard1:PRIMARY> rs.add("192.168.211.217:10003")
{ "ok" : 1 }
shard1:PRIMARY> rs.conf()
{
"_id" : "shard1",
"version" : 3,
"members" : [
{
"_id" : 0,
"host" : "192.168.211.217:10001"
},
{
"_id" : 1,
"host" : "192.168.211.217:10002"
},
{
"_id" : 2,
"host" : "192.168.211.217:10003"
}
]
}
Shard2和shard3同shard1配置副本集
#最终副本集配置如下：
shard3:PRIMARY> rs.conf()
{
"_id" : "shard3",
"version" : 3,
"members" : [
{
"_id" : 0,
"host" : "192.168.211.217:30001"
},
{
"_id" : 1,
"host" : "192.168.211.217:30002"
},
{
"_id" : 2,
"host" : "192.168.211.217:30003"
}
]
}
~]# /usr/local/mongodb/bin/mongo --port 20001
MongoDB shell version: 2.6.6
connecting to: 127.0.0.1:20001/test
shard2:PRIMARY> rs.conf()
{
"_id" : "shard2",
"version" : 3,
"members" : [
{
"_id" : 0,
"host" : "192.168.211.217:20001"
},
{
"_id" : 1,
"host" : "192.168.211.217:20002"
},
{
"_id" : 2,
"host" : "192.168.211.217:20003"
}
]
}
~]# /usr/local/mongodb/bin/mongo --port 10001
MongoDB shell version: 2.6.6
connecting to: 127.0.0.1:10001/test
shard1:PRIMARY> rs.conf()
{
"_id" : "shard1",
"version" : 3,
"members" : [
{
"_id" : 0,
"host" : "192.168.211.217:10001"
},
{
"_id" : 1,
"host" : "192.168.211.217:10002"
},
{
"_id" : 2,
"host" : "192.168.211.217:10003"
}
]
}
目前mongo相关进程端口情况如下：
#此时，刚好与环境规划列表对应
添加（副本集）分片
#连接到mongs，并切换到admin这里必须连接路由节点
~]# /usr/local/mongodb/bin/mongo --port 40000
MongoDB shell version: 2.6.6
connecting to: 127.0.0.1:40000/test
mongos> use admin
switched to db admin
mongos> db
admin
mongos> db.runCommand({"addShard":"shard1/192.168.211.217:10001"})
{ "shardAdded" : "shard1", "ok" : 1 }
mongos> db.runCommand({"addShard":"shard2/192.168.211.217:20001"})
{ "shardAdded" : "shard2", "ok" : 1 }
mongos> db.runCommand({"addShard":"shard3/192.168.211.217:30001"})
{ "shardAdded" : "shard3", "ok" : 1 }
mongos> db.runCommand({listshards:1})
mongos> db.runCommand({listshards:1})
{
"shards" : [
{
"_id" : "shard1",
"host" : "shard1/192.168.211.217:10001,192.168.211.217:10002,192.168.211.217:10003"
},
{
"_id" : "shard2",
"host" : "shard2/192.168.211.217:20001,192.168.211.217:20002,192.168.211.217:20003"
},
{
"_id" : "shard3",
"host" : "shard3/192.168.211.217:30001,192.168.211.217:30002,192.168.211.217:30003"
}
],
"ok" : 1
}
激活db和collections分片
激活数据库分片，命令
> db.runCommand( { enablesharding : “” } );
执行以上命令，可以让数据库跨shard，如果不执行这步，数据库只会存放在一个shard
一旦激活数据库分片，数据库中不同的collection将被存放在不同的shard上
但一个collection仍旧存放在同一个shard上，要使单个collection也分片，还需单独对collection作些操作
#如：激活test数据库分片功能，连接mongos进程
~]# /usr/local/mongodb/bin/mongo --port 50000
MongoDB shell version: 2.6.6
connecting to: 127.0.0.1:50000/test
mongos> use admin
switched to db admin
mongos> db.runCommand({"enablesharding":"test"})
{ "ok" : 1 }
要使单个collection也分片存储，需要给collection指定一个分片key，通过以下命令操作：
> db.runCommand( { shardcollection : “”,key : });
注： a. 分片的collection系统会自动创建一个索引（也可用户提前创建好）
b. 分片的collection只能有一个在分片key上的唯一索引，其它唯一索引不被允许
#对collection：test.yujx分片
mongos> use admin
switched to db admin
mongos> db.runCommand({"shardcollection":"test.yujx","key":{"_id":1}})
{ "collectionsharded" : "test.yujx", "ok" : 1 }
生成测试数据
mongos> use test
#生成测试数据，此处只是为了说明问题，并不一定要生成那么多行
mongos> for(var i=1;i<=888888;i++) db.yujx.save({"id":i,"a":123456789,"b":888888888,"c":100000000})
mongos> db.yujx.count()
271814 #此实验使用了这么多行测试数据
查看集合分片
mongos> db.yujx.stats()
{
"sharded" : true,
"systemFlags" : 1,
"userFlags" : 1,
"ns" : "test.yujx",
"count" : 271814,
"numExtents" : 19,
"size" : 30443168,
"storageSize" : 51773440,
"totalIndexSize" : 8862784,
"indexSizes" : {
"_id_" : 8862784
},
"avgObjSize" : 112,
"nindexes" : 1,
"nchunks" : 4,
"shards" : {
"shard1" : {
"ns" : "test.yujx",
"count" : 85563,
"size" : 9583056,
"avgObjSize" : 112,
"storageSize" : 11182080,
"numExtents" : 6,
"nindexes" : 1,
"lastExtentSize" : 8388608,
"paddingFactor" : 1,
"systemFlags" : 1,
"userFlags" : 1,
"totalIndexSize" : 2796192,
"indexSizes" : {
"_id_" : 2796192
},
"ok" : 1
},
"shard2" : {
"ns" : "test.yujx",
"count" : 180298,
"size" : 20193376,
"avgObjSize" : 112,
"storageSize" : 37797888,
"numExtents" : 8,
"nindexes" : 1,
"lastExtentSize" : 15290368,
"paddingFactor" : 1,
"systemFlags" : 1,
"userFlags" : 1,
"totalIndexSize" : 5862192,
"indexSizes" : {
"_id_" : 5862192
},
"ok" : 1
},
"shard3" : {
"ns" : "test.yujx",
"count" : 5953,
"size" : 666736,
"avgObjSize" : 112,
"storageSize" : 2793472,
"numExtents" : 5,
"nindexes" : 1,
"lastExtentSize" : 2097152,
"paddingFactor" : 1,
"systemFlags" : 1,
"userFlags" : 1,
"totalIndexSize" : 204400,
"indexSizes" : {
"_id_" : 204400
},
"ok" : 1
}
"ok" : 1
}
#此时我们连接各个分片的primary查询自身拥有的记录数据：
可以发现分别与上面的集合分片状态显示的一致
补充提示：secondary节点默认是无法执行查询操作，需要执行setSlaveOK操作
查看数据库分片
mongos> db.printShardingStatus()
#或者连接mongos的config数据库查询
mongos> db.shards.find()
{ "_id" : "shard1", "host" : "shard1/192.168.211.217:10001,192.168.211.217:10002,192.168.211.217:10003" }
{ "_id" : "shard2", "host" : "shard2/192.168.211.217:20001,192.168.211.217:20002,192.168.211.217:20003" }
{ "_id" : "shard3", "host" : "shard3/192.168.211.217:30001,192.168.211.217:30002,192.168.211.217:30003" }
mongos> db.databases.find()
{ "_id" : "admin", "partitioned" : false, "primary" : "config" }
{ "_id" : "test", "partitioned" : true, "primary" : "shard3" }
mongos> db.chunks.find()
{ "_id" : "test.yujx-_id_MinKey", "lastmod" : Timestamp(2, 0), "lastmodEpoch" : ObjectId("54b8b475a13b3af589cffc62"), "ns" : "test.yujx", "min" : { "_id" : { "$minKey" : 1 } }, "max" : { "_id" : ObjectId("54b8b58ddb2797bbb973a718") }, "shard" : "shard1" }
{ "_id" : "test.yujx-_id_ObjectId('54b8b58ddb2797bbb973a718')", "lastmod" : Timestamp(3, 1), "lastmodEpoch" : ObjectId("54b8b475a13b3af589cffc62"), "ns" : "test.yujx", "min" : { "_id" : ObjectId("54b8b58ddb2797bbb973a718") }, "max" : { "_id" : ObjectId("54b8b599db2797bbb973be59") }, "shard" : "shard3" }
{ "_id" : "test.yujx-_id_ObjectId('54b8b599db2797bbb973be59')", "lastmod" : Timestamp(4, 1), "lastmodEpoch" : ObjectId("54b8b475a13b3af589cffc62"), "ns" : "test.yujx", "min" : { "_id" : ObjectId("54b8b599db2797bbb973be59") }, "max" : { "_id" : ObjectId("54b8b6cfdb2797bbb9767ea2") }, "shard" : "shard2" }
{ "_id" : "test.yujx-_id_ObjectId('54b8b6cfdb2797bbb9767ea2')", "lastmod" : Timestamp(4, 0), "lastmodEpoch" : ObjectId("54b8b475a13b3af589cffc62"), "ns" : "test.yujx", "min" : { "_id" : ObjectId("54b8b6cfdb2797bbb9767ea2") }, "max" : { "_id" : { "$maxKey" : 1 } }, "shard" : "shard1" }
单点故障分析
由于这是为了了解入门mongodb做的实验，而故障模拟太浪费时间，所以这里就不一一列出，关于故障场景分析，可以参考：http://blog.itpub.net/27000195/viewspace-1404402/