MongoDB副本集的在线迁移
查看当前集群的状态：
{
"setName" : "gechongrepl",
"setVersion" : 11,
"ismaster" : true,
"secondary" : false,
"hosts" : [
"192.168.91.135:27017",
"192.168.91.133:27017"
],
"arbiters" : [
"192.168.91.144:27017"
],
"primary" : "192.168.91.135:27017",
"me" : "192.168.91.135:27017",
"maxBsonObjectSize" : 16777216,
"maxMessageSizeBytes" : 48000000,
"maxWriteBatchSize" : 1000,
"localTime" : ISODate("2015-06-17T02:29:52.582Z"),
"maxWireVersion" : 2,
"minWireVersion" : 0,
"ok" : 1
}
当前的副本集的状态为：
Primary:192.168.91.135:27017
Arbiters:192.168.91.144:27017
Secondary:192.168.91.133:27017
需求1：在副本集正常工作的情况下再添加一个节点。
1：模拟业务数据
gechongrepl:PRIMARY> db.test.count()
109000
gechongrepl:PRIMARY> for (i=0;i<10000;i++) db.test.insert({uid:i,uname:'testmoveonline'+i});
WriteResult({ "nInserted" : 1 })
gechongrepl:PRIMARY> db.test.count()
119000
gechongrepl:PRIMARY>
查看从节点的状态：
gechongrepl:SECONDARY> rs.slaveOk()
gechongrepl:SECONDARY> db.test.count()
119000
2：把从节点关掉
gechongrepl:SECONDARY> use admin
switched to db admin
gechongrepl:SECONDARY> db.shutdownServer() #查看当前的副本集的状态 gechongrepl:PRIMARY> rs.status() #再次插入数据 gechongrepl:PRIMARY> for (i=0;i<10000;i++) db.test.insert({uid:i,uname:'testmoveonline12345'+i});
WriteResult({ "nInserted" : 1 })
gechongrepl:PRIMARY> db.test.count()
129000
3：根据配置文件的配置，scpdata目录到新节点上。新节点IP为：192.168.91.132
[root@localhost mongodb]# scp -r ./data/ root@192.168.91.132:/root/software/mongodb/
root@192.168.91.132's password:
local.0 100% 64MB 21.3MB/s 00:03
local.1 100% 2047MB 21.8MB/s 01:34
test.0 100% 64MB 21.3MB/s 00:03
test.ns 100% 16MB 16.0MB/s 00:00
mongod.lock 100% 0 0.0KB/s 00:00
local.ns 100% 16MB 16.0MB/s 00:00
[root@localhost mongodb]# ls
bin data GNU-AGPL-3.0 log mongod.conf README THIRD-PARTY-NOTICES
[root@localhost mongodb]# scp -r ./log/ root@192.168.91.132:/root/software/mongodb/
root@192.168.91.132's password:
mongod 100% 1101KB 1.1MB/s 00:00
[root@localhost mongodb]#
查看192.168.91.132主机上的MongoDB目录已经存在scp过来的文件
[root@localhost mongodb]# ls
bin data GNU-AGPL-3.0 log mongod.conf README THIRD-PARTY-NOTICES
4：用scp过来的物理文件启动新节点
[root@localhost mongodb]# cd bin/
[root@localhost bin]# ./mongod --config /root/software/mongodb/mongod.conf
about to fork child process, waiting until server is ready for connections.
forked process: 30944
child process started successfully, parent exiting
5：在集群几点中加入新节点即可
gechongrepl:PRIMARY> rs.add("192.168.91.132:27017")
{
"_id" : 6,
"name" : "192.168.91.132:27017",
"health" : 1,
"state" : 6,
"stateStr" : "UNKNOWN",
"uptime" : 9,
"optime" : Timestamp(0, 0),
"optimeDate" : ISODate("1970-01-01T00:00:00Z"),
"lastHeartbeat" : ISODate("2015-06-17T02:49:43Z"),
"lastHeartbeatRecv" : ISODate("1970-01-01T00:00:00Z"),
"pingMs" : 0,
"lastHeartbeatMessage" : "still initializing"
}
6：remove刚才关掉的节点即可
gechongrepl:PRIMARY> rs.remove("192.168.91.133:27017")
如此再次重复这样的步骤即可把集群慢慢的转移到其他服务器上。