mongo 叢集配置

weixin_33785972發表於2017-12-05

master-slave

主的配置檔案 /Users/olifer/middle/mongo/master-slave/master/mongod.conf

bind_ip = 127.0.0.1
port = 27017
dbpath = /Users/olifer/middle/mongo/master-slave/master/data/
master = true

從的配置檔案

bind_ip = 127.0.0.1
port = 27018
dbpath = /Users/olifer/middle/mongo/master-slave/slave/data/
slave = true
source = 127.0.0.1:27017

啟動主伺服器,並且看到下面的日誌，說明配置成功。

mongod -f ~/middle/mongo/master-slave/master/mongod.conf

2017-12-05T15:35:22.479+0800 I JOURNAL  [initandlisten] journal dir=/Users/olifer/middle/mongo/master-slave/master/data/journal
2017-12-05T15:35:22.479+0800 I JOURNAL  [initandlisten] recover : no journal files present, no recovery needed
2017-12-05T15:35:22.500+0800 I JOURNAL  [durability] Durability thread started
2017-12-05T15:35:22.500+0800 I JOURNAL  [journal writer] Journal writer thread started
2017-12-05T15:35:22.501+0800 I CONTROL  [initandlisten] MongoDB starting : pid=6368 port=27017 dbpath=/Users/olifer/middle/mongo/master-slave/master/data/ master=1 64-bit host=oliferdeMacBook-Pro.local
2017-12-05T15:35:22.501+0800 I CONTROL  [initandlisten] db version v3.0.7
2017-12-05T15:35:22.501+0800 I CONTROL  [initandlisten] git version: nogitversion
2017-12-05T15:35:22.501+0800 I CONTROL  [initandlisten] build info: Darwin yosemitevm.local 14.5.0 Darwin Kernel Version 14.5.0: Wed Jul 29 02:26:53 PDT 2015; root:xnu-2782.40.9~1/RELEASE_X86_64 x86_64 BOOST_LIB_VERSION=1_49
2017-12-05T15:35:22.501+0800 I CONTROL  [initandlisten] allocator: system
2017-12-05T15:35:22.501+0800 I CONTROL  [initandlisten] options: { config: "/Users/olifer/middle/mongo/master-slave/master/mongod.conf", master: true, net: { bindIp: "127.0.0.1", port: 27017 }, storage: { dbPath: "/Users/olifer/middle/mongo/master-slave/master/data/" } }
2017-12-05T15:35:22.505+0800 I INDEX    [initandlisten] allocating new ns file /Users/olifer/middle/mongo/master-slave/master/data/local.ns, filling with zeroes...
2017-12-05T15:35:22.570+0800 I STORAGE  [FileAllocator] allocating new datafile /Users/olifer/middle/mongo/master-slave/master/data/local.0, filling with zeroes...
2017-12-05T15:35:22.570+0800 I STORAGE  [FileAllocator] creating directory /Users/olifer/middle/mongo/master-slave/master/data/_tmp
2017-12-05T15:35:22.786+0800 I STORAGE  [FileAllocator] done allocating datafile /Users/olifer/middle/mongo/master-slave/master/data/local.0, size: 64MB,  took 0.215 secs
2017-12-05T15:35:22.972+0800 I REPL     [initandlisten] ******
2017-12-05T15:35:22.972+0800 I REPL     [initandlisten] creating replication oplog of size: 192MB...
2017-12-05T15:35:22.972+0800 I STORAGE  [FileAllocator] allocating new datafile /Users/olifer/middle/mongo/master-slave/master/data/local.1, filling with zeroes...
2017-12-05T15:35:24.061+0800 I STORAGE  [FileAllocator] done allocating datafile /Users/olifer/middle/mongo/master-slave/master/data/local.1, size: 256MB,  took 1.088 secs
2017-12-05T15:35:24.117+0800 I REPL     [initandlisten] ******
2017-12-05T15:35:24.119+0800 I NETWORK  [initandlisten] waiting for connections on port 27017

啟動從伺服器,並且看到下面的日誌，說明配置成功。

mongod -f ~/middle/mongo/master-slave/slave/mongod.conf
2017-12-05T15:37:11.518+0800 I JOURNAL  [initandlisten] journal dir=/Users/olifer/middle/mongo/master-slave/slave/data/journal
2017-12-05T15:37:11.518+0800 I JOURNAL  [initandlisten] recover : no journal files present, no recovery needed
2017-12-05T15:37:11.535+0800 I JOURNAL  [durability] Durability thread started
2017-12-05T15:37:11.535+0800 I JOURNAL  [journal writer] Journal writer thread started
2017-12-05T15:37:11.535+0800 I CONTROL  [initandlisten] MongoDB starting : pid=7304 port=27018 dbpath=/Users/olifer/middle/mongo/master-slave/slave/data/ slave=1 64-bit host=oliferdeMacBook-Pro.local
2017-12-05T15:37:11.535+0800 I CONTROL  [initandlisten] db version v3.0.7
2017-12-05T15:37:11.535+0800 I CONTROL  [initandlisten] git version: nogitversion
2017-12-05T15:37:11.535+0800 I CONTROL  [initandlisten] build info: Darwin yosemitevm.local 14.5.0 Darwin Kernel Version 14.5.0: Wed Jul 29 02:26:53 PDT 2015; root:xnu-2782.40.9~1/RELEASE_X86_64 x86_64 BOOST_LIB_VERSION=1_49
2017-12-05T15:37:11.535+0800 I CONTROL  [initandlisten] allocator: system
2017-12-05T15:37:11.535+0800 I CONTROL  [initandlisten] options: { config: "/Users/olifer/middle/mongo/master-slave/slave/mongod.conf", net: { bindIp: "127.0.0.1", port: 27018 }, slave: true, source: "127.0.0.1:27017", storage: { dbPath: "/Users/olifer/middle/mongo/master-slave/slave/data/" } }
2017-12-05T15:37:11.536+0800 I INDEX    [initandlisten] allocating new ns file /Users/olifer/middle/mongo/master-slave/slave/data/local.ns, filling with zeroes...
2017-12-05T15:37:11.605+0800 I STORAGE  [FileAllocator] allocating new datafile /Users/olifer/middle/mongo/master-slave/slave/data/local.0, filling with zeroes...
2017-12-05T15:37:11.605+0800 I STORAGE  [FileAllocator] creating directory /Users/olifer/middle/mongo/master-slave/slave/data/_tmp
2017-12-05T15:37:11.824+0800 I STORAGE  [FileAllocator] done allocating datafile /Users/olifer/middle/mongo/master-slave/slave/data/local.0, size: 64MB,  took 0.218 secs
2017-12-05T15:37:11.890+0800 I NETWORK  [initandlisten] waiting for connections on port 27018
2017-12-05T15:37:12.894+0800 I REPL     [replslave] repl: syncing from host:127.0.0.1:27017
2017-12-05T15:37:17.930+0800 I REPL     [replslave] repl: sleep 2 sec before next pass
2017-12-05T15:37:19.935+0800 I REPL     [replslave] repl: syncing from host:127.0.0.1:27017
2017-12-05T15:37:24.958+0800 I REPL     [replslave] repl: syncing from host:127.0.0.1:27017
2017-12-05T15:37:33.188+0800 I REPL     [replslave] repl: syncing from host:127.0.0.1:27017
2017-12-05T15:37:43.193+0800 I REPL     [replslave] repl: syncing from host:127.0.0.1:27017
2017-12-05T15:37:53.198+0800 I REPL     [replslave] repl: syncing from host:127.0.0.1:27017

我們可以通過slave的日誌中可以看到，slave一直同master保持同步資料的聯絡。

我們通過客戶端連結上服務端

mongo 127.0.0.1:27017  #主
mongo 127.0.0.1:27018  #從

在主上操作

> db.test.insert({"name":"linyang"})
WriteResult({ "nInserted" : 1 })
> db.test.find({});
{ "_id" : ObjectId("5a264e0ecb6f7d3713c516a7"), "name" : "linyang" }

往裡面插入了一條記錄，可以再從的日誌中發現同步的記錄

2017-12-05T15:43:10.387+0800 I INDEX    [replslave] allocating new ns file /Users/olifer/middle/mongo/master-slave/slave/data/test.ns, filling with zeroes...
2017-12-05T15:43:10.450+0800 I STORAGE  [FileAllocator] allocating new datafile /Users/olifer/middle/mongo/master-slave/slave/data/test.0, filling with zeroes...
2017-12-05T15:43:10.646+0800 I STORAGE  [FileAllocator] done allocating datafile /Users/olifer/middle/mongo/master-slave/slave/data/test.0, size: 64MB,  took 0.196 secs
2017-12-05T15:43:10.702+0800 I REPL     [replslave] resync: dropping database test
2017-12-05T15:43:10.709+0800 I JOURNAL  [replslave] journalCleanup...
2017-12-05T15:43:10.710+0800 I JOURNAL  [replslave] removeJournalFiles
2017-12-05T15:43:10.713+0800 I JOURNAL  [replslave] journalCleanup...
2017-12-05T15:43:10.713+0800 I JOURNAL  [replslave] removeJournalFiles
2017-12-05T15:43:10.715+0800 I REPL     [replslave] resync: cloning database test to get an initial copy
2017-12-05T15:43:10.718+0800 I INDEX    [replslave] allocating new ns file /Users/olifer/middle/mongo/master-slave/slave/data/test.ns, filling with zeroes...
2017-12-05T15:43:10.803+0800 I STORAGE  [FileAllocator] allocating new datafile /Users/olifer/middle/mongo/master-slave/slave/data/test.0, filling with zeroes...
2017-12-05T15:43:11.056+0800 I STORAGE  [FileAllocator] done allocating datafile /Users/olifer/middle/mongo/master-slave/slave/data/test.0, size: 64MB,  took 0.252 secs
2017-12-05T15:43:11.107+0800 I INDEX    [replslave] build index on: test.test properties: { v: 1, key: { _id: 1 }, name: "_id_", ns: "test.test" }
2017-12-05T15:43:11.107+0800 I INDEX    [replslave]      building index using bulk method
2017-12-05T15:43:11.107+0800 I INDEX    [replslave] build index done.  scanned 1 total records. 0 secs
2017-12-05T15:43:11.107+0800 I STORAGE  [replslave] copying indexes for: { name: "test", options: {} }
2017-12-05T15:43:11.108+0800 I REPL     [replslave] resync: done with initial clone for db: test

我們再從的client中檢視有沒有資訊。

mongo 127.0.0.1:27018
MongoDB shell version: 3.0.7
connecting to: 127.0.0.1:27018/test
> db.test.find({});
{ "_id" : ObjectId("5a264e0ecb6f7d3713c516a7"), "name" : "linyang" }
>

也確認同步過來了。然後再從上插入記錄

> db.test.insert({"age":34});
WriteResult({ "writeError" : { "code" : undefined, "errmsg" : "not master" } })
>

從上不能插入記錄。如果主掛掉了，從又不能寫資料，那麼是否mongo叢集就掛掉了。答案當然是否定的，因為，mongo還有副本集。

副本集

我們使用三臺mongo來模擬.
第一臺的配置檔案 /Users/olifer/middle/mongo/replica/a/mongod.conf

dbpath = /Users/olifer/middle/mongo/replica/a/data
port = 8001
bind_ip = 127.0.0.1
replSet = child/127.0.0.1:8002

第二臺的配置檔案 /Users/olifer/middle/mongo/replica/b/mongod.conf

dbpath = /Users/olifer/middle/mongo/replica/a/data
port = 8002
bind_ip = 127.0.0.1
replSet = child/127.0.0.1:8003

第三臺的配置檔案 /Users/olifer/middle/mongo/replica/c/mongod.conf

dbpath = /Users/olifer/middle/mongo/replica/a/data
port = 8003
bind_ip = 127.0.0.1
replSet = child/127.0.0.1:8001

啟動伺服器

mongod -f ~/middle/mongo/replica/a/mongod.conf
mongod -f ~/middle/mongo/replica/b/mongod.conf
mongod -f ~/middle/mongo/replica/c/mongod.conf

啟動成功後，進入三個服務的任何一個客戶端

mongo 127.0.0.1:8002

> config = {_id: 'child', members: [
  {
    "_id":1,
    "host":"127.0.0.1:8001"
  },
  {
     "_id":2,
     "host":"127.0.0.1:8002"
  },
  {
     "_id":3,
     "host":"127.0.0.1:8003"
   }
   ]
 }
{
    "_id" : "child",
    "members" : [
        {
            "_id" : 1,
            "host" : "127.0.0.1:8001"
        },
        {
            "_id" : 2,
            "host" : "127.0.0.1:8002"
        },
        {
            "_id" : 3,
            "host" : "127.0.0.1:8003"
        }
    ]
}
> rs.initiate(config);
{ "ok" : 1 }
child:SECONDARY>

配置完後發現字首發生了改變，我們進入另外兩個渠道的客戶端

mongo 127.0.0.1:8001
MongoDB shell version: 3.0.7
connecting to: 127.0.0.1:8001/test
child:PRIMARY> 

mongo 127.0.0.1:8003
MongoDB shell version: 3.0.7
connecting to: 127.0.0.1:8003/test
child:SECONDARY>

其中child:PRIMARY>表示活躍節點。其餘為備份節點。注意:只有活躍節點才能進行查詢資料庫的資訊操作，備份節點不能進行會報錯,在任意客戶端執行rs.status() 來檢視所有狀態

child:PRIMARY> rs.status()
{
    "set" : "child",
    "date" : ISODate("2017-12-05T08:56:44.523Z"),
    "myState" : 1,
    "members" : [
        {
            "_id" : 1,
            "name" : "127.0.0.1:8001",
            "health" : 1,
            "state" : 1,
            "stateStr" : "PRIMARY",
            "uptime" : 620,
            "optime" : Timestamp(1512463818, 1),
            "optimeDate" : ISODate("2017-12-05T08:50:18Z"),
            "electionTime" : Timestamp(1512463821, 1),
            "electionDate" : ISODate("2017-12-05T08:50:21Z"),
            "configVersion" : 1,
            "self" : true
        },
        {
            "_id" : 2,
            "name" : "127.0.0.1:8002",
            "health" : 1,
            "state" : 2,
            "stateStr" : "SECONDARY",
            "uptime" : 387,
            "optime" : Timestamp(1512463818, 1),
            "optimeDate" : ISODate("2017-12-05T08:50:18Z"),
            "lastHeartbeat" : ISODate("2017-12-05T08:56:43.756Z"),
            "lastHeartbeatRecv" : ISODate("2017-12-05T08:56:43.756Z"),
            "pingMs" : 0,
            "configVersion" : 1
        },
        {
            "_id" : 3,
            "name" : "127.0.0.1:8003",
            "health" : 1,
            "state" : 2,
            "stateStr" : "SECONDARY",
            "uptime" : 387,
            "optime" : Timestamp(1512463818, 1),
            "optimeDate" : ISODate("2017-12-05T08:50:18Z"),
            "lastHeartbeat" : ISODate("2017-12-05T08:56:43.756Z"),
            "lastHeartbeatRecv" : ISODate("2017-12-05T08:56:43.756Z"),
            "pingMs" : 0,
            "configVersion" : 1
        }
    ],
    "ok" : 1
}

搭建成功後，我們來驗證一下。

在活躍的節點建立資料,在備份庫拉取資料

child:PRIMARY> db.repl.insert({"name":"123"});
WriteResult({ "nInserted" : 1 })

child:SECONDARY> db.repl.find();
{ "_id" : ObjectId("5a26602f5cefb1fdb377843b"), "name" : "123" }

功能正常。

從伺服器不能寫

child:SECONDARY> db.repl.insert({"age":444});
WriteResult({ "writeError" : { "code" : undefined, "errmsg" : "not master" } })

關閉活躍的節點，從伺服器會通過選舉得到最新的活躍的節點

當我關閉原來活躍的伺服器8001後，通過內部選舉，8002成了最新的活躍節點沒從字首也可以看出

child:SECONDARY>
child:PRIMARY>

現在看一下最新的狀態

child:PRIMARY> rs.status();
{
    "set" : "child",
    "date" : ISODate("2017-12-05T09:09:31.666Z"),
    "myState" : 1,
    "members" : [
        {
            "_id" : 1,
            "name" : "127.0.0.1:8001",
            "health" : 0,
            "state" : 8,
            "stateStr" : "(not reachable/healthy)",
            "uptime" : 0,
            "optime" : Timestamp(0, 0),
            "optimeDate" : ISODate("1970-01-01T00:00:00Z"),
            "lastHeartbeat" : ISODate("2017-12-05T09:09:31.193Z"),
            "lastHeartbeatRecv" : ISODate("2017-12-05T09:07:12.773Z"),
            "pingMs" : 0,
            "lastHeartbeatMessage" : "Failed attempt to connect to 127.0.0.1:8001; couldn't connect to server 127.0.0.1:8001 (127.0.0.1), connection attempt failed",
            "configVersion" : -1
        },
        {
            "_id" : 2,
            "name" : "127.0.0.1:8002",
            "health" : 1,
            "state" : 1,
            "stateStr" : "PRIMARY",
            "uptime" : 1308,
            "optime" : Timestamp(1512464432, 2),
            "optimeDate" : ISODate("2017-12-05T09:00:32Z"),
            "electionTime" : Timestamp(1512464835, 1),
            "electionDate" : ISODate("2017-12-05T09:07:15Z"),
            "configVersion" : 1,
            "self" : true
        },
        {
            "_id" : 3,
            "name" : "127.0.0.1:8003",
            "health" : 1,
            "state" : 2,
            "stateStr" : "SECONDARY",
            "uptime" : 1152,
            "optime" : Timestamp(1512464432, 2),
            "optimeDate" : ISODate("2017-12-05T09:00:32Z"),
            "lastHeartbeat" : ISODate("2017-12-05T09:09:30.963Z"),
            "lastHeartbeatRecv" : ISODate("2017-12-05T09:09:30.963Z"),
            "pingMs" : 0,
            "configVersion" : 1
        }
    ],
    "ok" : 1
}

說明內部的切換確實成功了。還剩下一個分片，我們下次再講。

參考資料

MongoDB 3.0 常見叢集的搭建(主從複製，副本集，分片....)

mongo副本集叢集安裝配置
2017-07-08
Go
redis偽叢集配置Cluster叢集模式
2019-04-09
Redis模式
Kafka叢集配置
2019-02-22
Kafka
MySQL叢集配置
2015-04-30
MySql
【Mongo】mongo配置檔案
2018-07-17
Go
Docker Elasticsearch 叢集配置
2021-02-25
DockerElasticsearch
MySQL叢集配置(轉)
2007-08-15
MySql
【MongoDB】windows平臺搭建Mongo資料庫複製集（相似叢集）（一）
2016-02-01
MongoDBWindows資料庫
MONGO 叢集修改linux主機時間後的影響
2019-04-26
GoLinux
HA叢集heartbeat配置--Nginx
2016-10-18
Nginx
FastDFS 叢集安裝配置
2017-11-01
AST
xCAT叢集管理配置工具
2011-05-14
【Mongodb】Sharding 叢集配置
2011-11-03
MongoDB
39_配置redis叢集
2024-03-13
Redis
linux 怎麼配置叢集
2020-04-20
Linux
spark叢集的配置檔案
2020-12-05
Spark
docker 配置 Mysql主從叢集
2023-04-14
DockerMySql
Vert.x 叢集配置 TCP
2017-12-13
TCP
redis sentinel 叢集監控配置
2016-06-01
Redis
weblogic8.1叢集配置
2008-08-21
Web
Weblogic10 叢集配置
2011-06-07
Web
Elastic Search 7.x 叢集配置
2020-05-16
AST
redis原理及叢集主從配置
2019-06-12
Redis
Hadoop完全分散式叢集配置
2021-05-01
Hadoop分散式
大資料Spark叢集模式配置
2020-10-16
大資料Spark模式
Elasticsearch分散式搜尋叢集配置
2016-05-01
Elasticsearch分散式
Quartz叢集原理及配置應用
2015-08-25
quartz
【Mongo】mongo分片加複製集的備份恢復
2018-05-03
Go
Hadoop叢集是配置時間同步！
2020-10-27
Hadoop
CentOS上zookeeper叢集模式安裝配置
2018-01-23
CentOS模式
搭建高可用MongoDB叢集（一）：配置MongoDB
2014-07-02
MongoDB
mongodb叢集搭建及配置安全認證
2016-02-18
MongoDB
Weblogic 10.3.6叢集配置手冊
2015-07-08
Web
如何配置一個 Docker Swarm 原生叢集
2015-07-30
DockerSwarm
3 安裝配置oracle叢集和RAC
2010-11-10
Oracle
Apache實現weblogic叢集配置（轉）
2012-05-10
ApacheWeb
Kafka詳解二、如何配置Kafka叢集
2015-11-17
Kafka
CentOS下torque叢集配置（二）叢集中配置NFS實現檔案共享
2017-05-04
CentOSNFS