Segmenting Data by Application or Customer按应用程序或客户划分数据

On this page本页内容

In sharded clusters, you can create zones of sharded data based on the shard key. 在分片集群中,可以基于分片键创建分片数据区域You can associate each zone with one or more shards in the cluster. 您可以将每个分区与集群中的一个或多个分片相关联。A shard can associate with any number of zones. 分片可以与任意数量的区域关联。In a balanced cluster, MongoDB migrates chunks covered by a zone only to those shards associated with the zone.在平衡集群中,MongoDB只将区域覆盖的迁移到与该区域关联的分片。

Tip提示

Changed in version 4.0.3.在版本4.0.3中更改

By defining the zones and the zone ranges before sharding an empty or a non-existing collection, the shard collection operation creates chunks for the defined zone ranges as well as any additional chunks to cover the entire range of the shard key values and performs an initial chunk distribution based on the zone ranges. 通过在对空集合或不存在的集合进行分片之前定义区域和区域范围,分片集合操作为定义的区域范围以及覆盖整个分片键值范围的任何附加块创建块,并基于区域范围执行初始块分布。This initial creation and distribution of chunks allows for faster setup of zoned sharding. 块的初始创建和分布允许更快地设置分区分片。After the initial distribution, the balancer manages the chunk distribution going forward.在初始分发之后,平衡器将继续管理块分发。See Pre-Define Zones and Zone Ranges for an Empty or Non-Existing Collection for an example.有关示例,请参阅为空集合或不存在集合预定义区域和区域范围

This tutorial shows you how to segment data using Zones.本教程介绍如何使用分区分割数据。

Consider the following scenarios where segmenting data by application or customer may be necessary:考虑以下场景,其中可能需要按应用程序或客户划分数据:

This diagram illustrates a sharded cluster using zones to segment data based on application or customer. 此图说明了一个分片集群,它使用区域根据应用程序或客户划分数据。This allows for data to be isolated to specific shards. 这允许将数据隔离到特定的分片。Additionally, each shard can have specific hardware allocated to fit the performance requirement of the data stored on that shard.此外,每个分片可以分配特定的硬件,以满足存储在该分片上的数据的性能要求。

Overview of zones used for supporting data segmentation

Scenario情形

An application tracks the score of a user along with a client field, storing scores in the gamify database under the users collection. 应用程序跟踪用户的分数以及client字段,将分数存储在gamify数据库中的users集合下。Each possible value of client requires its own zone to allow for data segmentation. client的每个可能值都需要自己的区域,以允许数据分段。It also allows the administrator to optimize the hardware for each shard associated to a client for performance and cost.它还允许管理员优化与client关联的每个分片的硬件,以获得性能和成本。

The following documents represent a partial view of two users:以下文档表示两个用户的部分视图:

{
  "_id" : ObjectId("56f08c447fe58b2e96f595fa"),
  "client" : "robot",
  "userid" : 123,
  "high_score" : 181,
  ...,
}
{
  "_id" : ObjectId("56f08c447fe58b2e96f595fb"),
  "client" : "fruitos",
  "userid" : 456,
  "high_score" : 210,
  ...,
}

Shard Key分片键

The users collection uses the { client : 1, userid : 1 } compound index as the shard key.users集合使用{ client : 1, userid : 1 }复合索引作为分片键。

The client field in each document allows creating a zone for each distinct client value.每个文档中的client字段允许为每个不同的客户机值创建一个区域。

The userid field provides a high cardinality and low frequency component to the shard key relative to country.userid字段为分片键提供了相对于国家的高基数和低频率分量。

See Choosing a Shard Key for more general instructions on selecting a shard key.有关选择分片键的更多一般说明,请参阅选择分片键

Architecture策略

The application requires adding shard to a zone associated to a specific client.应用程序需要将分片添加到与特定client关联的区域。

The sharded cluster deployment currently consists of four shards.分片集群部署目前由四个分片组成。

Diagram of Data Segmentation Architecture using zones

Zones区域

For this application, there are two client zones.对于此应用程序,有两个客户端区域。

Diagram of zones used for supporting data segmentation
Robot client ("robot")
This zone represents all documents where client : robot.此区域表示client : robot的所有文档。
FruitOS client ("fruitos")
This zone represents all documents where client : fruitos.此区域表示client : fruitos的所有文档。

Write Operations写操作

With zones, if an inserted or updated document matches a configured zone, it can only be written to a shard inside that zone.对于区域,如果插入或更新的文档与配置的区域匹配,则只能将其写入该区域内的分片。

MongoDB can write documents that do not match a configured zone to any shard in the cluster.MongoDB可以编写与集群中任何分片的配置区域不匹配的文档。

Note注意

The behavior described above requires the cluster to be in a steady state with no chunks violating a configured zone. 上述行为要求集群处于稳定状态,没有块违反配置区域。See the following section on the balancer for more information.有关更多信息,请参阅以下关于平衡器的章节。

Read Operations读操作

MongoDB can route queries to a specific shard if the query includes at least the client field.如果查询至少包括client字段,MongoDB可以将查询路由到特定的分片。

For example, MongoDB can attempt a targeted read operation on the following query:例如,MongoDB可以尝试对以下查询执行目标读取操作

chatDB = db.getSiblingDB("gamify")
chatDB.users.find( { "client" : "robot" , "userid" : "123" } )

Queries without the client field perform broadcast operations.没有client字段的查询执行广播操作

Balancer平衡器

The balancer migrates chunks to the appropriate shard respecting any configured zones. 平衡器将块迁移到与任何配置区域相关的适当分片。Until the migration, shards may contain chunks that violate configured zones. Once balancing completes, shards should only contain chunks whose ranges do not violate its assigned zones.在迁移之前,分片可能包含违反配置区域的块。一旦平衡完成,分片应该只包含其范围不违反其分配区域的块。

Adding or removing zones or zone ranges can result in chunk migrations. 添加或删除区域或区域范围可能导致块迁移。Depending on the size of your data set and the number of chunks a zone or zone range affects, these migrations may impact cluster performance. 根据数据集的大小以及区域或区域范围影响的块数,这些迁移可能会影响群集性能。Consider running your balancer during specific scheduled windows. 考虑在特定的计划窗口期间运行平衡器See Schedule the Balancing Window for a tutorial on how to set a scheduling window.有关如何设置计划窗口的教程,请参阅计划平衡窗口

Security安全

For sharded clusters running with Role-Based Access Control, authenticate as a user with at least the clusterManager role on the admin database.对于使用基于角色的访问控制运行的分片集群,至少使用admin数据库上的clusterManager角色作为用户进行身份验证。

Procedure过程

You must be connected to a mongos associated to the target sharded cluster to proceed. 您必须连接到与目标分片集群关联的mongos才能继续。You cannot create zones or zone ranges by connecting directly to a shard.无法通过直接连接到分片来创建分区或分区范围。

1

Disable the Balancer禁用平衡器

The balancer must be disabled on the collection to ensure no migrations take place while configuring the new zones.必须在集合上禁用平衡器,以确保在配置新区域时不会发生迁移。

Use sh.disableBalancing(), specifying the namespace of the collection, to stop the balancer.使用sh.disableBalancing(),指定集合的命名空间,以停止平衡器。

sh.disableBalancing("chat.message")

Use sh.isBalancerRunning() to check if the balancer process is currently running. 使用sh.isBalancerRunning()检查平衡器进程当前是否正在运行。Wait until any current balancing rounds have completed before proceeding.在继续之前,请等待所有当前平衡轮完成。

2

Add each shard to the appropriate zone将每个分片添加到相应的分区

Add shard0000 to the robot zone.shard0000添加到robot区域。

sh.addShardTag("shard0000", "robot")

Add shard0001 to the robot zone.shard0001添加到robot区域。

sh.addShardTag("shard0001", "robot")

Add shard0002 to the fruitos zone.shard0002添加到fruitos区域。

sh.addShardTag("shard0002", "fruitos")

Add shard0003 to the fruitos zone.shard0003添加到fruitos区域。

sh.addShardTag("shard0003", "fruitos")

Run sh.status() to review the zone configured for the sharded cluster.运行sh.status()查看为分片集群配置的区域。

3

Define ranges for each zone定义每个区域的范围

Define range for the robot client and associate it to the robot zone using the sh.addTagRange() method.定义robot客户端的范围,并使用sh.addTagRange()方法将其与robot区域关联。

This method requires:该方法要求:

  • The full namespace of the target collection目标集合的完整命名空间
  • The inclusive lower bound of the range范围的包含下限
  • The exclusive upper bound of the range范围的唯一上界
  • The name of the zone区域名称
sh.addTagRange(
  "gamify.users",
  { "client" : "robot", "userid" : MinKey },
  { "client" : "robot", "userid" : MaxKey }, 
  "robot"
)

Define range for the fruitos client and associate it to the fruitos zone using the sh.addTagRange() method.定义fruitos客户端的范围,并使用sh.addTagRange()方法将其与fruitos区域关联。

This method requires:该方法要求:

  • The full namespace of the target collection目标集合的完整命名空间
  • The inclusive lower bound of the range范围的包含下限
  • The exclusive upper bound of the range范围的唯一上界
  • The name of the zone区域名称
sh.addTagRange(
  "gamify.users",
  { "client" : "fruitos", "userid" : MinKey },
  { "client" : "fruitos", "userid" : MaxKey }, 
  "fruitos"
)

The MinKey and MaxKey values are reserved special values for comparisons. MinKeyMaxKey值是为比较保留的特殊值。MinKey always compares as lower than every other possible value, while 总是比其他可能的值低,而MaxKey always compares as higher than every other possible value. 总是比任何其他可能的值都高。The configured ranges captures every user for each client.配置的范围捕获每个client的每个用户。

4

Enable the Balancer启用平衡器

Re-enable the balancer to rebalance the cluster.重新启用平衡器以重新平衡群集。

Use sh.enableBalancing(), specifying the namespace of the collection, to start the balancer.使用sh.enableBalancing(),指定集合的命名空间,以启动平衡器。

sh.enableBalancing("chat.message")

Use sh.isBalancerRunning() to check if the balancer process is currently running.使用sh.isBalancerRunning()检查平衡器进程当前是否正在运行。

5

Review the changes回顾变化

The next time the balancer runs, it splits and migrates chunks across the shards respecting the configured zones.下一次运行平衡器时,它会根据配置的分区在分片之间分割迁移块。

Once balancing finishes, the shards in the robot zone only contain documents with client : robot, while shards in the fruitos zone only contain documents with client : fruitos.平衡完成后,robot区域中的分片仅包含client : robot的文档,而fruitos区域中的片段仅包含client : fruitos的文档。

You can confirm the chunk distribution by running sh.status().您可以通过运行sh.status()来确认区块分布。

←  Tiered Hardware for Varying SLA or SLODistributed Local Writes for Insert Only Workloads →