[原创] MongoDB管理与开发精要《红丸出品》21.4.3 管理维护Replica Sets之增减节点
http://blog.chinaunix.net/attachment/201205/26/26785103_1338009683Kvjx.jpg21.4.3增减节点MongoDB Replica Sets不仅提供高可用性的解决方案，它也同时提供负载均衡的解决方案，增减Replica Sets节点在实际应用中非常普遍，例如当应用的读压力暴增时，3台节点的环境已不能满足需求，那么就需要增加一些节点将压力平均分配一下；当应用的压力小时，可以减少一些节点来减少硬件资源的成本；总之这是一个长期且持续的工作。21.4.3.1 增加节点官方给我们提了2个方案用于增加节点，一种是通过oplog来增加节点，一种是通过数据库快照(--fastsync)和oplog来增加节点，下面将分别介绍。21.4.3.1.1 通过oplog增加节点①、配置并启动新节点，启用28013这个端口给新的节点# mkdir -p /data/data/r3# echo "this is rs1 super secret key" > /data/key/r3# chmod 600 /data/key/r3# /Apps/mongo/bin/mongod --replSet rs1 --keyFile /data/key/r3 --fork --port 28013 --dbpath /data/data/r3 --logpath=/data/log/r3.log --logappendall output going to: /data/log/r3.logforked process: 10553#
②、添加此新节点到现有的Replica Sets
rs1:PRIMARY> rs.add("localhost:28013"){ "ok" : 1 }
③、查看Replica Sets我们可以清晰的看到内部是如何添加28013这个新节点的.步骤一: 进行初始化
rs1: PRIMARY > rs.status(){ "set" : "rs1", "date" : ISODate("2012-05-31T12:17:44Z"), "myState" : 1, "members" : […… { "_id" : 3, "name" : "localhost:28013", "health" : 0, "state" : 6, "stateStr" : "(not reachable/healthy)", "uptime" : 0, "optime" : { "t" : 0, "i" : 0 }, "optimeDate" : ISODate("1970-01-01T00:00:00Z"), "lastHeartbeat" : ISODate("2012-05-31T12:17:43Z"), "errmsg" : "still initializing" } ], "ok" : 1}
步骤二: 进行数据同步
rs1:PRIMARY> rs.status(){ "set" : "rs1", "date" : ISODate("2012-05-31T12:18:07Z"), "myState" : 1, "members" : […… { "_id" : 3, "name" : "localhost:28013", "health" : 1, "state" : 3, "stateStr" : "RECOVERING", "uptime" : 16, "optime" : { "t" : 0, "i" : 0 }, "optimeDate" : ISODate("1970-01-01T00:00:00Z"), "lastHeartbeat" : ISODate("2012-05-31T12:18:05Z"), "errmsg" : "initial sync need a member to be primary or secondary to do our initial sync" } ], "ok" : 1}
步骤三: 初始化同步完成
rs1:PRIMARY> rs.status(){ "set" : "rs1", "date" : ISODate("2012-05-31T12:18:08Z"), "myState" : 1, "members" : […… { "_id" : 3, "name" : "localhost:28013", "health" : 1, "state" : 3, "stateStr" : "RECOVERING", "uptime" : 17, "optime" : { "t" : 1338466661000, "i" : 1 }, "optimeDate" : ISODate("2012-05-31T12:17:41Z"), "lastHeartbeat" : ISODate("2012-05-31T12:18:07Z"), "errmsg" : "initial sync done" } ], "ok" : 1}
步骤四: 节点添加完成，状态正常
rs1:PRIMARY> rs.status(){ "set" : "rs1", "date" : ISODate("2012-05-31T12:18:10Z"), "myState" : 1, "members" : […… { "_id" : 3, "name" : "localhost:28013", "health" : 1, "state" : 2, "stateStr" : "SECONDARY", "uptime" : 19, "optime" : { "t" : 1338466661000, "i" : 1 }, "optimeDate" : ISODate("2012-05-31T12:17:41Z"), "lastHeartbeat" : ISODate("2012-05-31T12:18:09Z") } ], "ok" : 1}
④、验证数据已经同步过来了
# /Apps/mongo/bin/mongo -port 28013MongoDB shell version: 1.8.1connecting to: 127.0.0.1:28013/testrs1:SECONDARY> rs.slaveOk()rs1:SECONDARY> db.c1.find(){ "_id" : ObjectId("4fc760d2383ede1dce14ef86"), "age" : 10 }rs1:SECONDARY>
21.4.3.1.2 通过数据库快照(--fastsync)和oplog增加节点通过oplog直接进行增加节点操作简单且无需人工干预过多，但oplog是capped collection，采用循环的方式进行日志处理，所以采用oplog的方式进行增加节点，有可能导致数据的不一致，因为日志中存储的信息有可能已经刷新过了。不过没关系，我们可以通过数据库快照(--fastsync)和oplog结合的方式来增加节点，这种方式的操作流程是，先取某一个复制集成员的物理文件来做为初始化数据，然后剩余的部分用oplog日志来追，最终达到数据一致性
①、取某一个复制集成员的物理文件来做为初始化数据
# scp -r /data/data/r3 /data/data/r4# echo "this is rs1 super secret key" > /data/key/r4# chmod 600 /data/key/r4
②、在取完物理文件后，在c1集中插入一条新文档，用于最后验证此更新也同步了
rs1:PRIMARY> db.c1.find(){ "_id" : ObjectId("4fc760d2383ede1dce14ef86"), "age" : 10 }rs1:PRIMARY> db.c1.insert({age:20})rs1:PRIMARY> db.c1.find() { "_id" : ObjectId("4fc760d2383ede1dce14ef86"), "age" : 10 }{ "_id" : ObjectId("4fc7748f479e007bde6644ef"), "age" : 20 }rs1:PRIMARY>
③、启用28014这个端口给新的节点
/Apps/mongo/bin/mongod --replSet rs1 --keyFile /data/key/r4 --fork --port 28014 --dbpath /data/data/r4 --logpath=/data/log/r4.log --logappend--fastsync
④、添加28014节点
rs1:PRIMARY> rs.add("localhost:28014"){ "ok" : 1 }
⑤、验证数据已经同步过来了
# /Apps/mongo/bin/mongo -port 28014MongoDB shell version: 1.8.1connecting to: 127.0.0.1:28014/testrs1:SECONDARY> rs.slaveOk()rs1:SECONDARY> db.c1.find(){ "_id" : ObjectId("4fc760d2383ede1dce14ef86"), "age" : 10 }{ "_id" : ObjectId("4fc7748f479e007bde6644ef"), "age" : 20 }rs1:SECONDARY>
21.4.3.2 减少节点下面将刚刚添加的两个新节点28013和28014从复制集中去除掉，只需执行rs.remove指令就可以了，具体如下:
rs1:PRIMARY> rs.remove("localhost:28014"){ "ok" : 1 }rs1:PRIMARY> rs.remove("localhost:28013"){ "ok" : 1 }
查看复制集状态，可以看到现在只有28010、28011、28012这三个成员，原来的28013和28014都成功去除了
rs1:PRIMARY> rs.status(){ "set" : "rs1", "date" : ISODate("2012-05-31T14:08:29Z"), "myState" : 1, "members" : [ { "_id" : 0, "name" : "localhost:28010", "health" : 1, "state" : 1, "stateStr" : "PRIMARY", "optime" : { "t" : 1338473273000, "i" : 1 }, "optimeDate" : ISODate("2012-05-31T14:07:53Z"), "self" : true }, { "_id" : 1, "name" : "localhost:28011", "health" : 1, "state" : 2, "stateStr" : "SECONDARY", "uptime" : 34, "optime" : { "t" : 1338473273000, "i" : 1 }, "optimeDate" : ISODate("2012-05-31T14:07:53Z"), "lastHeartbeat" : ISODate("2012-05-31T14:08:29Z") }, { "_id" : 2, "name" : "localhost:28012", "health" : 1, "state" : 2, "stateStr" : "SECONDARY", "uptime" : 34, "optime" : { "t" : 1338473273000, "i" : 1 }, "optimeDate" : ISODate("2012-05-31T14:07:53Z"), "lastHeartbeat" : ISODate("2012-05-31T14:08:29Z") } ], "ok" : 1}rs1:PRIMARY>
-------------------------------------------------------------------免费IT培训（ChinaDBA.net）实践者！学费1元，全捐公益
http://service.t.sina.com.cn/widget/qmd/2446082491/9e768382/1.png
页:
[1]