MongoDB Sharding

最新推荐文章于 2024-08-30 13:17:57 发布

java_4_ever

最新推荐文章于 2024-08-30 13:17:57 发布

阅读量1.2k

点赞数

分类专栏： mongoDB

本文链接：https://blog.csdn.net/java_4_ever/article/details/41117837

版权

mongoDB 专栏收录该内容

3 篇文章 0 订阅

订阅专栏

为什么需要Shard

MongoDB的Shard是一种水平分片，主要目的无非就是提高总存储空间，减少单击压力这些，提高总负载这样。

Shard的基本介绍可以参照官方文档

建立Shard集群

MongoDB Shard集群的物理结构如上图所示，主要有三种节点（这几个中文名字是本人自己定义的）：路由节点（Routers），配置节点（config server），数据节点（shard）。三种节点的具体功能就不多少了，看完官方文档也差不多理解了。

上图中推荐的机器数目并非强制要求，主要是为了热备，提升性能，数据备份等。

Shard这一部分，上图所示是一个replica set，就是多机数据备份集合，结构是这样的：一个primary node，多个secondary node，还有一个arbiter(裁决节点)，所有的读写操作都集中在primary node节点上，然后primary和 secondary之间会做数据同步。当primary宕机或者不可能之后，arbiter会从secondary中挑选一个成为新的primary node。arbiter不是必须的，甚至secondary也不是强制的。

Shard可以只是单台机器而非replica set

下面是本人操作的一些步骤，所有操作在一台Ubuntu虚拟机上完成：

1 启动数据节点

./bin/mongod --port 27021 --dbpath data_27021 --fork --logpath logs/27021.log
./bin/mongod --port 27022 --dbpath data_27022 --replSet test27022 --fork --logpath logs/27021.log
./bin/mongod --port 27023 --dbpath data_27023 --replSet test27023 --fork --logpath logs/27021.log

这边我启动了三个数据节点，--fork表示已后台进程的方式运行， --replSet指定了这个节点所属replica set的名称，这里有test27022和test27023两个replica set，并且每个set只有一个节点。端口27021对应的实例留作演示shard为单台机器用。

2 启动config server

./bin/mongod --configsvr --port 27024 --dbpath data_27024 --fork --logpath logs/27024.log

3 启动router

./bin/mongos --configdb localhost:27024 --port 27025 --fork --logpath logs/27021.log

启动的时候指明config server的地址，如果有多个以,分隔。

4 配置replica set

./bin/mongo localhost:27022 #登录至replica set中任一机器

> use admin
switched to db admin

> rsconfig={_id:"test27022", members:[{_id:0, host:"localhost:27022",priority:1}]}
{
    "_id" : "test27022",
    "members" : [
        {
            "_id" : 0,
            "host" : "localhost:27022",
            "priority" : 1
        }
    ]
}

> rs.initiate(rsconfig)  #设置replica set
{
    "info" : "Config now saved locally.  Should come online in about a minute.",
    "ok" : 1
}

> rs.status() # 查看设置结果
{
    "set" : "test27022",
    "date" : ISODate("2014-11-10T14:29:52Z"),
    "myState" : 1,
    "members" : [
        {
            "_id" : 0,
            "name" : "localhost:27022",
            "health" : 1,
            "state" : 1,
            "stateStr" : "PRIMARY",
            "uptime" : 507,
            "optime" : Timestamp(1415629406, 1),
            "optimeDate" : ISODate("2014-11-10T14:23:26Z"),
            "electionTime" : Timestamp(1415629407, 1),
            "electionDate" : ISODate("2014-11-10T14:23:27Z"),
            "self" : true
        }
    ],
    "ok" : 1
}

重复上面的操作建立replica set "test27023"。

5 配置Sharding

./bin/mongo localhost:27025 #登录至router

mongos> use admin #所有操作必须在admin下进行
switched to db admin

mongos> sh.addShard("test27022/localhost:27022"); #test27022是replica set的名字，localhost:27022是该set中primary node的地址
{ "shardAdded" : "test27022", "ok" : 1 }

mongos> sh.addShard("test27023/localhost:27023");
{ "shardAdded" : "test27023", "ok" : 1 }

mongos> db.runCommand({enableSharding:"uus_test"}) #设置允许shard的DB
{ "ok" : 1 }

mongos> db.runCommand({shardCollection:"uus_test.user", key:{"userId":1}}); 
#设置允许shard的collection，及其对应的shard key，这里"userId":1 表示选择userId作为shard key，值为1表示按范围分片。
{ "collectionsharded" : "uus_test.user", "ok" : 1 }

mongos> sh.status() 
--- Sharding Status --- 
  sharding version: {
    "_id" : 1,
    "version" : 4,
    "minCompatibleVersion" : 4,
    "currentVersion" : 5,
    "clusterId" : ObjectId("5460c9f220ab13caa089c97c")
}
  shards:
    {  "_id" : "test27022",  "host" : "test27022/localhost:27022" }
    {  "_id" : "test27023",  "host" : "test27023/localhost:27023" }
  databases:
    {  "_id" : "admin",  "partitioned" : false,  "primary" : "config" }
    {  "_id" : "uus_test",  "partitioned" : true,  "primary" : "test27022" }
        uus_test.user
            shard key: { "userId" : 1 }
            chunks:
                test27022   1
            { "userId" : { "$minKey" : 1 } } -->> { "userId" : { "$maxKey" : 1 } } on : test27022 Timestamp(1, 0)

添加shard还可以用另外的命令db.runCommand({addShard:""})，至此MongoDB Sharding集群搭建完毕。

关于Shard key的选择，MongoDB会为shard key这个字段建立索引，shard key的选择很关键，对于性能会有比较大的影响，尽量使用那些选择性比较大的字段（有点类似于关系型数据库建立索引）。如果查询包含shard key则只在一个shard里面查询即可，如果不包含则需要到所有shard里面查询（大多数时候。这样会比前一种情况慢）。这方面了解的不多，可以参考这里和这里

修改chunk size

这是一个全局的参数。默认是64MB。

小的chunk会让不同的shard数据量更均衡。但会导致更多的Migration。

大的chunk会减少migration。不同的shard数据量不均衡。

这样修改chunk size。先连接上任意mongos

db.settings.save( { _id:"chunksize", value: 1 } )

单位是MB，这里修改为1M，方便插入数据之后看到多个chunk。另外可以通过在启动router的时候用参数 --chunkSize=1来指定chunk大小。

插入数据

./bin/mongo localhost:27025 #登录到router

mongos> use uus_test;
switched to db uus_test

mongos> db.user.count();
0

mongos> for(i=0;i<100000;i++){db.user.insert({"userId":i,"name":"francis", "createDate": new Date()});}  # 插入10W数据
WriteResult({ "nInserted" : 1 })

mongos> db.user.count();
100000

mongos> db.user.stats() #查看user状态
{
    "sharded" : true,
    "systemFlags" : 1,
    "userFlags" : 1,
    "ns" : "uus_test.user",
    "count" : 100000,
    "numExtents" : 12,
    "size" : 11200000,
    "storageSize" : 22364160,
    "totalIndexSize" : 6066592,
    "indexSizes" : {
        "_id_" : 3254048,
        "userId_1" : 2812544
    },
    "avgObjSize" : 112,
    "nindexes" : 2,
    "nchunks" : 11,
    "shards" : {                            # sharding信息
        "test27022" : {
            "ns" : "uus_test.user",
            "count" : 36378,
            "size" : 4074336,
            "avgObjSize" : 112,
            "storageSize" : 11182080,
            "numExtents" : 6,
            "nindexes" : 2,
            "lastExtentSize" : 8388608,
            "paddingFactor" : 1,
            "systemFlags" : 1,
            "userFlags" : 1,
            "totalIndexSize" : 2207520,
            "indexSizes" : {
                "_id_" : 1185520,
                "userId_1" : 1022000
            },
            "ok" : 1
        },
        "test27023" : {
            "ns" : "uus_test.user",
            "count" : 63622,
            "size" : 7125664,
            "avgObjSize" : 112,
            "storageSize" : 11182080,
            "numExtents" : 6,
            "nindexes" : 2,
            "lastExtentSize" : 8388608,
            "paddingFactor" : 1,
            "systemFlags" : 1,
            "userFlags" : 1,
            "totalIndexSize" : 3859072,
            "indexSizes" : {
                "_id_" : 2068528,
                "userId_1" : 1790544
            },
            "ok" : 1
        }
    },
    "ok" : 1
}

从db.user.stats()的结果可以看出来，user这个collection被sharded，数据分布在test27022和test27023两个shard上面，test27022上面有36378条数据，大小为4.07MB；test27023上面有63622条数据，大小为7.12MB。

mongos> use admin
switched to db admin

mongos> sh.status()  #查看整个shard的信息
--- Sharding Status --- 
  sharding version: {
    "_id" : 1,
    "version" : 4,
    "minCompatibleVersion" : 4,
    "currentVersion" : 5,
    "clusterId" : ObjectId("5460c9f220ab13caa089c97c")
}
  shards:
    {  "_id" : "test27022",  "host" : "test27022/localhost:27022" }
    {  "_id" : "test27023",  "host" : "test27023/localhost:27023" }
  databases:
    {  "_id" : "admin",  "partitioned" : false,  "primary" : "config" }
    {  "_id" : "uus_test",  "partitioned" : true,  "primary" : "test27022" }
        uus_test.user
            shard key: { "userId" : 1 }
            chunks:
                test27022   5
                test27023   6
            { "userId" : { "$minKey" : 1 } } -->> { "userId" : 0 } on : test27022 Timestamp(2, 1) 
            { "userId" : 0 } -->> { "userId" : 4978 } on : test27022 Timestamp(1, 3) 
            { "userId" : 4978 } -->> { "userId" : 15714 } on : test27022 Timestamp(3, 0) 
            { "userId" : 15714 } -->> { "userId" : 26181 } on : test27022 Timestamp(4, 0) 
            { "userId" : 26181 } -->> { "userId" : 36378 } on : test27022 Timestamp(5, 0) 
            { "userId" : 36378 } -->> { "userId" : 47401 } on : test27023 Timestamp(5, 1) 
            { "userId" : 47401 } -->> { "userId" : 57463 } on : test27023 Timestamp(3, 4) 
            { "userId" : 57463 } -->> { "userId" : 69238 } on : test27023 Timestamp(4, 2) 
            { "userId" : 69238 } -->> { "userId" : 80487 } on : test27023 Timestamp(4, 4) 
            { "userId" : 80487 } -->> { "userId" : 90351 } on : test27023 Timestamp(5, 2)

上面的结果中： "shards"表示当前集群有哪些shards；"databases"列出了所有的DB，"partitioned"表示是否可以shard， "primary"表示该DB的主分区，未被sharded的collection会存储在主分区上。

然后chunks这一部分是user这个collection中数据分布的情况，从上面可以看出 user的数据在test27022中有5个chunk，在test27023中有6个chunk。

{ "userId" : 0 } -->> { "userId" : 4978 } on : test27022 Timestamp(1, 3)表示userId在[0，4987)这个区间的，存储在test27022上面， Timestamp表示什么意思我也还没搞清楚。

{ "$minKey" : 1 }和{ "$maxKey" : 1 }表示极小值和极大值。

数据分布的并不是十分平均，range shard这种方式具体是怎么样的原理，还没搞清楚。

新增shard

新增的shard可以不为空，但是不允许包含现有集群中已有的DB。这点没有具体尝试，本例中添加的是一个空的shard。

上面已经建立了一个shard集群，现在把最开始启动的localhost:27021这个实例添加进来，作为单台机器添加进来，至于replica set操作步骤一致。

./bin/mongo localhost:27025

mongos> use admin
switched to db admin

mongos> sh.addShard("localhost:27021");
{ "shardAdded" : "shard0000", "ok" : 1 }

mongos> sh.status();
--- Sharding Status --- 
  sharding version: {
    "_id" : 1,
    "version" : 4,
    "minCompatibleVersion" : 4,
    "currentVersion" : 5,
    "clusterId" : ObjectId("5460c9f220ab13caa089c97c")
}
  shards:
    {  "_id" : "shard0000",  "host" : "localhost:27021" }
    {  "_id" : "test27022",  "host" : "test27022/localhost:27022" }
    {  "_id" : "test27023",  "host" : "test27023/localhost:27023" }
  databases:
    {  "_id" : "admin",  "partitioned" : false,  "primary" : "config" }
    {  "_id" : "uus_test",  "partitioned" : true,  "primary" : "test27022" }
        uus_test.user
            shard key: { "userId" : 1 }
            chunks:
                shard0000   3
                test27022   4
                test27023   4
            { "userId" : { "$minKey" : 1 } } -->> { "userId" : 0 } on : shard0000 Timestamp(7, 0) 
            { "userId" : 0 } -->> { "userId" : 4978 } on : test27022 Timestamp(7, 1) 
            { "userId" : 4978 } -->> { "userId" : 15714 } on : test27022 Timestamp(3, 0) 
            { "userId" : 15714 } -->> { "userId" : 26181 } on : test27022 Timestamp(4, 0) 
            { "userId" : 26181 } -->> { "userId" : 36378 } on : test27022 Timestamp(5, 0) 
            { "userId" : 36378 } -->> { "userId" : 47401 } on : shard0000 Timestamp(6, 0) 
            { "userId" : 47401 } -->> { "userId" : 57463 } on : shard0000 Timestamp(8, 0) 
            { "userId" : 57463 } -->> { "userId" : 69238 } on : test27023 Timestamp(8, 1) 
            { "userId" : 69238 } -->> { "userId" : 80487 } on : test27023 Timestamp(4, 4) 
            { "userId" : 80487 } -->> { "userId" : 90351 } on : test27023 Timestamp(5, 2) 
            { "userId" : 90351 } -->> { "userId" : { "$maxKey" : 1 } } on : test27023 Timestamp(5, 3)

新增shard很简单，另外也可以看到新增shard之后mongoDB自动就开始同步数据了。

当chunk过多的时候（大概是超过20个）, sh.status()的结果不会显示每个chunk的信息，若需要强制显示：sh.status(true);

之所以会自动同步数据是因为MongoDB有个balancer在运行，查看balancer信息：

mongos> use config
switched to db config

mongos> db.locks.find({_id: "balancer"}).pretty();
{
    "_id" : "balancer",
    "state" : 0,
    "who" : "user-VirtualBox:27025:1415629297:1804289383:Balancer:1681692777",
    "ts" : ObjectId("5460ddef20ab13caa089ccd7"),
    "process" : "user-VirtualBox:27025:1415629297:1804289383",
    "when" : ISODate("2014-11-10T15:46:55.249Z"),
    "why" : "doing balance round"
}

可以通过命令来设置balancer是否运行: sh.startBalancer()， sh.stopBalancer()。更多命令请参考官方文档

config数据库下面有这些collection：

mongos> show collections;
changelog
chunks
collections
databases
lockpings
locks
mongos
settings
shards
system.indexes
tags
version

chunks可查看每个chunk的详细信息， settings可查看系统参数，如chunkSize

移除shard

移除shard需要等待balancer将需要移除的shard上面的数据迁移到其他shards。如果移除的shard是某个DB的primary shard，则需要为该DB设置新的primary shard。

mongos> use admin
switched to db admin

mongos> db.runCommand({removeShard:"localhost:27021"})
{
    "msg" : "draining started successfully",
    "state" : "started",
    "shard" : "shard0000",
    "ok" : 1
}

mongos> sh.status()
--- Sharding Status --- 
  sharding version: {
    "_id" : 1,
    "version" : 4,
    "minCompatibleVersion" : 4,
    "currentVersion" : 5,
    "clusterId" : ObjectId("5460c9f220ab13caa089c97c")
}
  shards:
    {  "_id" : "shard0000",  "host" : "localhost:27021",  "draining" : true }
    {  "_id" : "test27022",  "host" : "test27022/localhost:27022" }
    {  "_id" : "test27023",  "host" : "test27023/localhost:27023" }
  databases:
    {  "_id" : "admin",  "partitioned" : false,  "primary" : "config" }
    {  "_id" : "uus_test",  "partitioned" : true,  "primary" : "test27022" }
        uus_test.user
            shard key: { "userId" : 1 }
            chunks:
                test27022   6
                test27023   5
            { "userId" : { "$minKey" : 1 } } -->> { "userId" : 0 } on : test27022 Timestamp(9, 0) 
            { "userId" : 0 } -->> { "userId" : 4978 } on : test27022 Timestamp(7, 1) 
            { "userId" : 4978 } -->> { "userId" : 15714 } on : test27022 Timestamp(3, 0) 
            { "userId" : 15714 } -->> { "userId" : 26181 } on : test27022 Timestamp(4, 0) 
            { "userId" : 26181 } -->> { "userId" : 36378 } on : test27022 Timestamp(5, 0) 
            { "userId" : 36378 } -->> { "userId" : 47401 } on : test27023 Timestamp(10, 0) 
            { "userId" : 47401 } -->> { "userId" : 57463 } on : test27022 Timestamp(11, 0) 
            { "userId" : 57463 } -->> { "userId" : 69238 } on : test27023 Timestamp(8, 1) 
            { "userId" : 69238 } -->> { "userId" : 80487 } on : test27023 Timestamp(4, 4) 
            { "userId" : 80487 } -->> { "userId" : 90351 } on : test27023 Timestamp(5, 2) 
            { "userId" : 90351 } -->> { "userId" : { "$maxKey" : 1 } } on : test27023 Timestamp(5, 3) 

mongos> db.runCommand({removeShard:"localhost:27021"})
{
    "msg" : "removeshard completed successfully",
    "state" : "completed",
    "shard" : "shard0000",
    "ok" : 1
}

mongos> sh.status()
--- Sharding Status --- 
  sharding version: {
    "_id" : 1,
    "version" : 4,
    "minCompatibleVersion" : 4,
    "currentVersion" : 5,
    "clusterId" : ObjectId("5460c9f220ab13caa089c97c")
}
  shards:
    {  "_id" : "test27022",  "host" : "test27022/localhost:27022" }
    {  "_id" : "test27023",  "host" : "test27023/localhost:27023" }
  databases:
    {  "_id" : "admin",  "partitioned" : false,  "primary" : "config" }
    {  "_id" : "uus_test",  "partitioned" : true,  "primary" : "test27022" }
        uus_test.user
            shard key: { "userId" : 1 }
            chunks:
                test27022   6
                test27023   5
            { "userId" : { "$minKey" : 1 } } -->> { "userId" : 0 } on : test27022 Timestamp(9, 0) 
            { "userId" : 0 } -->> { "userId" : 4978 } on : test27022 Timestamp(7, 1) 
            { "userId" : 4978 } -->> { "userId" : 15714 } on : test27022 Timestamp(3, 0) 
            { "userId" : 15714 } -->> { "userId" : 26181 } on : test27022 Timestamp(4, 0) 
            { "userId" : 26181 } -->> { "userId" : 36378 } on : test27022 Timestamp(5, 0) 
            { "userId" : 36378 } -->> { "userId" : 47401 } on : test27023 Timestamp(10, 0) 
            { "userId" : 47401 } -->> { "userId" : 57463 } on : test27022 Timestamp(11, 0) 
            { "userId" : 57463 } -->> { "userId" : 69238 } on : test27023 Timestamp(8, 1) 
            { "userId" : 69238 } -->> { "userId" : 80487 } on : test27023 Timestamp(4, 4) 
            { "userId" : 80487 } -->> { "userId" : 90351 } on : test27023 Timestamp(5, 2) 
            { "userId" : 90351 } -->> { "userId" : { "$maxKey" : 1 } } on : test27023 Timestamp(5, 3)

首先db.runCommand({removeShard:"localhost:27021"})，提示"draining started successfully"，这个时候开始迁移localhost:27021上面的数据。

这个时候查看shard信息会看到localhost:27021这个shard显示"draining" : true，表明正在迁移数据。当user只存在test27022和test27023这两个shard上面的时候，表明27021以完成数据迁移。

此时再次运行db.runCommand({removeShard:"localhost:27021"})，完成shard的移除。

下面尝试移除test27022，该分区是uus_test的primary shard：

mongos> db.runCommand({removeShard:"localhost:27022"})
{
    "msg" : "draining started successfully",
    "state" : "started",
    "shard" : "test27022",
    "note" : "you need to drop or movePrimary these databases",
    "dbsToMove" : [
        "uus_test"
    ],
    "ok" : 1
}

mongos> sh.status()
--- Sharding Status --- 
  sharding version: {
    "_id" : 1,
    "version" : 4,
    "minCompatibleVersion" : 4,
    "currentVersion" : 5,
    "clusterId" : ObjectId("5460c9f220ab13caa089c97c")
  }
  shards:
    {  "_id" : "test27022",  "host" : "test27022/localhost:27022",  "draining" : true }
    {  "_id" : "test27023",  "host" : "test27023/localhost:27023" }
  databases:
    {  "_id" : "admin",  "partitioned" : false,  "primary" : "config" }
    {  "_id" : "uus_test",  "partitioned" : true,  "primary" : "test27022" }
        uus_test.user
            shard key: { "userId" : 1 }
            chunks:
                test27023   7
                test27022   4
            { "userId" : { "$minKey" : 1 } } -->> { "userId" : 0 } on : test27023 Timestamp(12, 0) 
            { "userId" : 0 } -->> { "userId" : 4978 } on : test27023 Timestamp(13, 0) 
            { "userId" : 4978 } -->> { "userId" : 15714 } on : test27022 Timestamp(13, 1) 
            { "userId" : 15714 } -->> { "userId" : 26181 } on : test27022 Timestamp(4, 0) 
            { "userId" : 26181 } -->> { "userId" : 36378 } on : test27022 Timestamp(5, 0) 
            { "userId" : 36378 } -->> { "userId" : 47401 } on : test27023 Timestamp(10, 0) 
            { "userId" : 47401 } -->> { "userId" : 57463 } on : test27022 Timestamp(11, 0) 
            { "userId" : 57463 } -->> { "userId" : 69238 } on : test27023 Timestamp(8, 1) 
            { "userId" : 69238 } -->> { "userId" : 80487 } on : test27023 Timestamp(4, 4) 
            { "userId" : 80487 } -->> { "userId" : 90351 } on : test27023 Timestamp(5, 2) 
            { "userId" : 90351 } -->> { "userId" : { "$maxKey" : 1 } } on : test27023 Timestamp(5, 3) 

mongos> sh.status()
--- Sharding Status --- 
  sharding version: {
    "_id" : 1,
    "version" : 4,
    "minCompatibleVersion" : 4,
    "currentVersion" : 5,
    "clusterId" : ObjectId("5460c9f220ab13caa089c97c")
  }
  shards:
    {  "_id" : "test27022",  "host" : "test27022/localhost:27022",  "draining" : true }
    {  "_id" : "test27023",  "host" : "test27023/localhost:27023" }
  databases:
    {  "_id" : "admin",  "partitioned" : false,  "primary" : "config" }
    {  "_id" : "uus_test",  "partitioned" : true,  "primary" : "test27022" }
        uus_test.user
            shard key: { "userId" : 1 }
            chunks:
                test27023   11
            { "userId" : { "$minKey" : 1 } } -->> { "userId" : 0 } on : test27023 Timestamp(12, 0) 
            { "userId" : 0 } -->> { "userId" : 4978 } on : test27023 Timestamp(13, 0) 
            { "userId" : 4978 } -->> { "userId" : 15714 } on : test27023 Timestamp(14, 0) 
            { "userId" : 15714 } -->> { "userId" : 26181 } on : test27023 Timestamp(15, 0) 
            { "userId" : 26181 } -->> { "userId" : 36378 } on : test27023 Timestamp(16, 0) 
            { "userId" : 36378 } -->> { "userId" : 47401 } on : test27023 Timestamp(10, 0) 
            { "userId" : 47401 } -->> { "userId" : 57463 } on : test27023 Timestamp(17, 0) 
            { "userId" : 57463 } -->> { "userId" : 69238 } on : test27023 Timestamp(8, 1) 
            { "userId" : 69238 } -->> { "userId" : 80487 } on : test27023 Timestamp(4, 4) 
            { "userId" : 80487 } -->> { "userId" : 90351 } on : test27023 Timestamp(5, 2) 
            { "userId" : 90351 } -->> { "userId" : { "$maxKey" : 1 } } on : test27023 Timestamp(5, 3) 

mongos> db.runCommand({moveprimary:"uus_test", to:"localhost:27023"})  #迁移至新的primary shard
{ "primary " : "test27023:test27023/localhost:27023", "ok" : 1 }

mongos> db.runCommand({removeShard:"localhost:27022"})
{
    "msg" : "removeshard completed successfully",
    "state" : "completed",
    "shard" : "test27022",
    "ok" : 1
}

mongos> sh.status();
--- Sharding Status --- 
  sharding version: {
    "_id" : 1,
    "version" : 4,
    "minCompatibleVersion" : 4,
    "currentVersion" : 5,
    "clusterId" : ObjectId("5460c9f220ab13caa089c97c")
}
shards:
    {  "_id" : "test27023",  "host" : "test27023/localhost:27023" }
  databases:
    {  "_id" : "admin",  "partitioned" : false,  "primary" : "config" }
    {  "_id" : "uus_test",  "partitioned" : true,  "primary" : "test27023" }
        uus_test.user
            shard key: { "userId" : 1 }
            chunks:
                test27023   11
            { "userId" : { "$minKey" : 1 } } -->> { "userId" : 0 } on : test27023 Timestamp(12, 0) 
            { "userId" : 0 } -->> { "userId" : 4978 } on : test27023 Timestamp(13, 0) 
            { "userId" : 4978 } -->> { "userId" : 15714 } on : test27023 Timestamp(14, 0) 
            { "userId" : 15714 } -->> { "userId" : 26181 } on : test27023 Timestamp(15, 0) 
            { "userId" : 26181 } -->> { "userId" : 36378 } on : test27023 Timestamp(16, 0) 
            { "userId" : 36378 } -->> { "userId" : 47401 } on : test27023 Timestamp(10, 0) 
            { "userId" : 47401 } -->> { "userId" : 57463 } on : test27023 Timestamp(17, 0) 
            { "userId" : 57463 } -->> { "userId" : 69238 } on : test27023 Timestamp(8, 1) 
            { "userId" : 69238 } -->> { "userId" : 80487 } on : test27023 Timestamp(4, 4) 
            { "userId" : 80487 } -->> { "userId" : 90351 } on : test27023 Timestamp(5, 2) 
            { "userId" : 90351 } -->> { "userId" : { "$maxKey" : 1 } } on : test27023 Timestamp(5, 3)

操作步骤同移除普通的shard一样，多了移动uustest到新的primary shard的操作`db.runCommand({moveprimary:"uustest", to:"localhost:27023"})`。

其他

手动移动chunk

手动移动chunk的机会很少

mongos> db.adminCommand({moveChunk: "uus_test.user", find:{userId:10}, to:"localhost:27021"})

故障恢复当集群中某个分片宕掉以后，只要不涉及到该节点的操纵仍然能进行。当宕掉的节点重启后，集群能自动从故障中恢复过来。

java_4_ever

关注

0
点赞
踩
0

收藏

觉得还不错? 一键收藏
0
评论
复制链接

分享到 QQ

分享到新浪微博

扫一扫

专栏目录