TiDB - 版本历史

imported>Yumeto：使用DisamAssist清理消歧義連結：Go（改連結至Go語言）。

2026-03-02T15:26:44Z

使用DisamAssist清理消歧義連結：Go（改連結至Go語言）。

新页面

{{NoteTA|G1=IT}}
{{Infobox software
| name = TiDB
| developer = 北京平凯星辰科技发展有限公司（PingCAP）
| released = {{Start date and age|2017|10|15}}<ref>{{cite web|title=1.0 GA release notes|url=https://github.com/pingcap/tidb/releases/tag/v1.0.0}}</ref>
| latest release version = {{LSR/wikidata}}
| programming language = [[Go語言|Go]] (TiDB), [[Rust (编程语言)|Rust]] (TiKV)
| language = 英文、中文
| genre = [[NewSQL]]
| license = [[Apache许可证|Apache 2.0]]
}}

'''TiDB'''是一个开源的[[NewSQL]]数据库，支持混合事务和分析处理（[[混合事务分析处理|HTAP]]）工作负载<ref>{{cite web|title=How TiDB combines OLTP and OLAP in a distributed database|url=https://www.infoworld.com/article/3313327/how-tidb-combines-oltp-and-olap-in-a-distributed-database.html|accessdate=2019-12-07|archive-date=2019-05-02|archive-url=https://web.archive.org/web/20190502001138/https://www.infoworld.com/article/3313327/how-tidb-combines-oltp-and-olap-in-a-distributed-database.html|dead-url=no}}</ref>。它与[[MySQL]]兼容，并且可以提供水平可扩展性、强一致性和高可用性。它主要由PingCAP公司开发和支持，并在Apache 2.0下授权。TiDB从Google的Spanner<ref>{{cite web|title=Spanner: Google's Globally-Distributed Database|url=https://ai.google/research/pubs/pub39966|accessdate=2019-12-07|archive-date=2018-12-30|archive-url=https://web.archive.org/web/20181230181056/https://ai.google/research/pubs/pub39966|dead-url=no}}</ref>和F1<ref>{{cite web |title=F1: A Distributed SQL Database That Scales |url=https://ai.google/research/pubs/pub41344 |accessdate=2019-12-07 |archive-date=2019-12-02 |archive-url=https://web.archive.org/web/20191202100632/https://ai.google/research/pubs/pub41344 |dead-url=no }}</ref>论文中汲取了最初的设计灵感。<ref>{{cite web |title=TiDB Brings Distributed Scalability to SQL |url=https://thenewstack.io/tidb-brings-distributed-scalability-sql/ |accessdate=2019-12-07 |archive-date=2019-05-02 |archive-url=https://web.archive.org/web/20190502001139/https://thenewstack.io/tidb-brings-distributed-scalability-sql/ |dead-url=no }}</ref>

TiDB被InfoWorld 2018 Bossie Award评选为最佳数据存储和分析开源软件之一。<ref>{{cite web |title=The best open source software for data storage and analytics |url=https://www.infoworld.com/article/3306454/the-best-open-source-software-for-data-storage-and-analytics.html |accessdate=2019-12-07 |archive-date=2019-09-02 |archive-url=https://web.archive.org/web/20190902155054/https://www.infoworld.com/article/3306454/the-best-open-source-software-for-data-storage-and-analytics.html |dead-url=no }}</ref>

==历史==
PingCAP Inc.是一家软件公司，成立于2015年4月，成立后便开始开发TiDB。该公司是TiDB及其关联的开源社区的主要开发人员，维护人员和驱动者。<ref>{{cite web |title=TiDB developer PingCAP wants to expand in North America after raising $50M Series C |url=https://techcrunch.com/2018/09/11/tidb-developer-pingcap-wants-to-expand-in-north-america-after-raising-50m-series-c/ |accessdate=2019-12-07 |archive-date=2019-05-02 |archive-url=https://web.archive.org/web/20190502001137/https://techcrunch.com/2018/09/11/tidb-developer-pingcap-wants-to-expand-in-north-america-after-raising-50m-series-c/ |dead-url=no }}</ref>

== 主要特性 ==

=== 水平可扩展性 ===
TiDB可以通过添加新节点来扩展SQL处理和存储容量。这使得基础设施容量的扩展更加容易和灵活，而传统的[[关系数据库]]只能垂直扩展。

=== MySQL兼容性 ===
对于应用程序来说，TiDB的行为就如同一个[[MySQL]] 5.7服务器。用户可以继续使用所有现有的MySQL客户端库。因为TiDB的SQL处理层是从头开始构建的，而不是一个MySQL分支，所以它的兼容性不是100%，<ref>{{cite web |title=Compatibility with MySQL |url=https://www.pingcap.com/docs/sql/mysql-compatibility/ |access-date=2019-12-07 |archive-url=https://web.archive.org/web/20190502001148/https://pingcap.com/docs/sql/mysql-compatibility/ |archive-date=2019-05-02 |dead-url=yes }}</ref>而且MySQL和TiDB之间存在已知的行为差异。<ref>{{cite web |title=Meet TiDB: An open source NewSQL database |url=https://opensource.com/article/18/11/key-differences-between-mysql-and-tidb |accessdate=2019-12-07 |archive-date=2019-07-31 |archive-url=https://web.archive.org/web/20190731164052/https://opensource.com/article/18/11/key-differences-between-mysql-and-tidb |dead-url=no }}</ref>

=== 具有强一致性的分布式事务 ===
TiDB在内部基于范围将表分片为的小块，这些块称为“区域”。<ref>{{cite web |title=TiKV Architecture |url=https://tikv.org/docs/concepts/architecture/#region |access-date=2019-12-07 |archive-url=https://web.archive.org/web/20190502001201/https://tikv.org/docs/concepts/architecture/#region |archive-date=2019-05-02 |dead-url=yes }}</ref>每个区域的默认大小约为100 MB，TiDB在内部使用[[二阶段提交]]来确保区域以事务上一致的方式进行维护。

=== 云原生 ===
为了使部署、配置、操作和维护更加灵活，TiDB被设计成可在云中工作。TiDB的存储层称为TiKV，于2018年8月成为[[Linux基金會|云原生计算基金会]]的沙盒级别项目。<ref>{{cite web |title=CNCF to Host TiKV in the Sandbox |url=https://www.cncf.io/blog/2018/08/28/cncf-to-host-tikv-in-the-sandbox/ |accessdate=2019-12-07 |archive-date=2019-09-03 |archive-url=https://web.archive.org/web/20190903065931/https://www.cncf.io/blog/2018/08/28/cncf-to-host-tikv-in-the-sandbox/ |dead-url=no }}</ref>TiDB平台的体系结构还允许分别独立扩展SQL处理和存储。

=== 更精简的ETL流程 ===
TiDB可以支持联机事务处理（[[線上交易處理|OLTP]]）和联机分析处理（[[線上分析處理|OLAP]]）工作负载。这意味着，虽然用户可能传统上在MySQL上进行交易，然后将数据抽取、转换和加载（[[ETL]]）到一个[[列式数据库|列存储]]中进行分析处理，但这一步骤不再需要。

=== 高可用性 ===
TiDB使用[[Raft|Raft一致性算法]]<ref>{{cite web |title=The Raft Consensus Algorithm |url=https://raft.github.io/ |accessdate=2019-12-07 |archive-date=2018-01-20 |archive-url=https://web.archive.org/web/20180120122313/https://raft.github.io/ |dead-url=no }}</ref>确保数据在整个Raft组的存储过程中高度可用和安全复制。在发生故障时，Raft组将自动为失败的成员选择一个新的领导者，并自我修复TiDB集群，而不需要任何必要的人工干预。失败和自愈操作对应用程序是透明的。

== 部署方法 ==

=== Kubernetes Operator ===
可以使用TiDB Operator将TiDB部署到支持[[Kubernetes]]的云环境中。<ref>{{cite web |title=Database Operators Bring Stateful Workloads to Kubernetes |url=https://thenewstack.io/databases-operators-bring-stateful-workloads-to-kubernetes/ |accessdate=2019-12-07 |archive-date=2019-12-29 |archive-url=https://web.archive.org/web/20191229101845/https://thenewstack.io/databases-operators-bring-stateful-workloads-to-kubernetes/ |dead-url=no }}</ref>Operator是打包、部署和管理Kubernetes应用程序的方法。它是为运行有状态工作负载而设计的，最早由[[CoreOS]]在2016年引入。<ref>{{cite web |title=Introducing Operators: Putting Operational Knowledge into Software |url=https://coreos.com/blog/introducing-operators.html |accessdate=2019-12-07 |archive-date=2019-11-25 |archive-url=https://web.archive.org/web/20191125171801/https://coreos.com/blog/introducing-operators.html |dead-url=no }}</ref>TiDB Operator<ref>{{cite web|title=TiDB Operator GitHub repo|url=https://github.com/pingcap/tidb-operator|accessdate=2019-12-07|archive-date=2019-07-16|archive-url=https://web.archive.org/web/20190716040257/https://github.com/pingcap/tidb-operator|dead-url=no}}</ref>最初由PingCAP开发，并于2018年8月开源。<ref>{{cite web |title=Introducing the Kubernetes Operator for TiDB |url=https://www.infoworld.com/article/3297700/introducing-the-kubernetes-operator-for-tidb.html |accessdate=2019-12-07 |archive-date=2019-05-02 |archive-url=https://web.archive.org/web/20190502212824/https://www.infoworld.com/article/3297700/introducing-the-kubernetes-operator-for-tidb.html |dead-url=no }}</ref>TiDB Operator可以用于在笔记本电脑<ref>{{cite web |title=Deploy TiDB to Kubernetes on Your Laptop |url=https://github.com/pingcap/tidb-operator/blob/master/docs/local-dind-tutorial.md |accessdate=2019-12-07 |archive-date=2018-08-17 |archive-url=https://web.archive.org/web/20180817160200/https://github.com/pingcap/tidb-operator/blob/master/docs/local-dind-tutorial.md |dead-url=no }}</ref>、Google Cloud Platform的Google Kubernetes引擎<ref>{{cite web|title=Deploy TiDB, a distributed MySQL compatible database, to Kubernetes on Google Cloud |url=https://github.com/pingcap/tidb-operator/blob/master/docs/google-kubernetes-tutorial.md}}</ref>和Amazon Web Services的Kubernetes弹性容器服务上部署TiDB。<ref>{{cite web|title=Deploy TiDB, a distributed MySQL compatible database, on Kubernetes via AWS EKS |url=https://github.com/pingcap/tidb-operator/blob/master/docs/aws-eks-tutorial.md}}</ref>

=== Ansible ===
通过使用TiDB-Ansible playbook，可以使用[[Ansible (軟體)]]部署TiDB。<ref>{{cite web |title=Ansible Playbook for TiDB |url=https://github.com/pingcap/tidb-ansible |accessdate=2019-12-07 |archive-date=2020-08-25 |archive-url=https://web.archive.org/web/20200825123319/https://github.com/pingcap/tidb-ansible |dead-url=no }}</ref>

=== Docker ===
可以使用[[Docker]]在多个节点和多台机器上的容器化环境中部署TiDB，而Docker Compose可以使用单个命令部署TiDB以进行测试。<ref>{{cite web |title=How to Spin Up an HTAP Database in 5 Minutes With TiDB + TiSpark |url=https://dzone.com/articles/how-to-spin-up-an-htap-database-in-5-minutes-with |accessdate=2019-12-07 |archive-date=2019-05-02 |archive-url=https://web.archive.org/web/20190502220705/https://dzone.com/articles/how-to-spin-up-an-htap-database-in-5-minutes-with |dead-url=no }}</ref>

== 工具 ==

TiDB提供了一系列开源工具来帮助现有[[MySQL]]和[[MariaDB]]用户进行数据复制和迁移。

=== Syncer与Data Migration ===
Syncer是一个工具，它支持从MySQL或MariaDB实例到TiDB集群的完整数据迁移或增量数据复制。<ref>{{cite web |title=Syncer User Guide |url=https://pingcap.com/docs/tools/syncer/#syncer-user-guide |accessdate=2019-12-07 |archive-date=2019-11-21 |archive-url=https://web.archive.org/web/20191121073406/https://pingcap.com/docs/tools/syncer/#syncer-user-guide |dead-url=no }}</ref>Data Migration（DM）是 Syncer的第二代迭代，它适合将已分片的MySQL或MariaDB表中的数据复制到TiDB。<ref>{{cite web |title=DM GitHub Repo |url=https://github.com/pingcap/dm |accessdate=2019-12-07 |archive-date=2020-08-25 |archive-url=https://web.archive.org/web/20200825122824/https://github.com/pingcap/dm |dead-url=no }}</ref>Syncer/DM的一个常见用例是将MySQL或MariaDB表连接到TiDB，几乎将TiDB视为从属，然后在该TiDB集群上直接实时运行分析工作负载。

=== Lightning ===
Lightning是一个可将大型MySQL转储高速完全导入到新的TiDB群集中的工具，它提供了比执行每个SQL语句更快的导入体验。该工具用于把大量数据快速填充到一个最初为空的TiDB集群，以便加快测试或生产迁移。之所以能提升导入速度，是因为TiDB通过将SQL语句解析为键值对，然后直接生成'''有序字符串表'''（SSTable）文件到[[RocksDB]]。<ref>{{cite web |title=Introducing TiDB Lightning |url=https://www.pingcap.com/blog/introducing-tidb-lightning/ |accessdate=2019-12-07 |archive-date=2019-05-08 |archive-url=https://web.archive.org/web/20190508184805/https://pingcap.com/blog/introducing-tidb-lightning/ |dead-url=no }}</ref>

=== TiDB-Binlog ===
TiDB-Binlog是一个用于收集对TiDB集群所做逻辑更改的工具。它用于在两个TiDB集群之间或从一个TiDB集群到另一个下游平台之间提供增量备份和复制。

它的功能类似于MySQL主从复制。主要的区别在于，由于TiDB是一个分布式数据库文件管理器，因此每个TiDB实例生成的binlog需要在被下游使用之前，根据事务提交的时间进行合并和排序。<ref>{{cite web|title=TiDB-Binlog Architecture Evolution and Implementation Principles|url=https://www.pingcap.com/blog/tidb-binlog-architecture-evolution-and-implementation-principles/|accessdate=2019-12-07|archive-date=2019-06-16|archive-url=https://web.archive.org/web/20190616125703/https://pingcap.com/blog/tidb-binlog-architecture-evolution-and-implementation-principles/|dead-url=no}}</ref>

== 用户案例 ==
目前，已有近3000家公司在使用 TiDB，其中包括Mastercard、AirBNB、Square、Paypay、中国银行、建设银行、美团、Bilibili、Shopee<ref>{{cite web|url=https://pingcap.com/success-stories/tidb-in-shopee/|title=Shopping on Shopee, the TiDB Way - TiDB|website=pingcap.com|accessdate=2019-12-07|archive-date=2019-11-06|archive-url=https://web.archive.org/web/20191106031810/https://pingcap.com/success-stories/tidb-in-shopee/|dead-url=no}}</ref>、BookMyShow<ref>{{cite web|url=https://pingcap.com/success-stories/tidb-in-bookmyshow/|title=BookMyShow.com: More Uptime, 30% Less Operational Cost with TiDB - TiDB|website=pingcap.com|accessdate=2019-12-07|archive-date=2019-11-06|archive-url=https://web.archive.org/web/20191106031814/https://pingcap.com/success-stories/tidb-in-bookmyshow/|dead-url=no}}</ref>、小米<ref>{{cite web|url=https://pingcap.com/success-stories/tidb-in-xiaomi/|title=Powering the Xiaomi Mobile Lifestyle with TiDB - TiDB|website=pingcap.com|accessdate=2019-12-07|archive-date=2019-11-06|archive-url=https://web.archive.org/web/20191106031813/https://pingcap.com/success-stories/tidb-in-xiaomi/|dead-url=no}}</ref>、知乎<ref>{{cite web|url=https://pingcap.com/success-stories/lesson-learned-from-queries-over-1.3-trillion-rows-of-data-within-milliseconds-of-response-time-at-zhihu/|title=Lesson Learned from Queries over 1.3 Trillion Rows of Data Within Milliseconds of Response Time at Zhihu.com - TiDB|website=pingcap.com|accessdate=2019-12-07|archive-date=2019-11-06|archive-url=https://web.archive.org/web/20191106031817/https://pingcap.com/success-stories/lesson-learned-from-queries-over-1.3-trillion-rows-of-data-within-milliseconds-of-response-time-at-zhihu/|dead-url=no}}</ref>、美团点评<ref>{{cite web|url=https://pingcap.com/success-stories/tidb-in-meituan-dianping/|title=TiDB, the Key to a Better Life for Meituan-Dianping’s 290 Million Monthly Users - TiDB|website=pingcap.com|accessdate=2019-12-07|archive-date=2019-11-21|archive-url=https://web.archive.org/web/20191121032032/https://pingcap.com/success-stories/tidb-in-meituan-dianping/|dead-url=no}}</ref>、爱奇艺<ref>{{cite web|url=https://pingcap.com/success-stories/tidb-in-iqiyi/|title=Always Fun, Always On: How TiDB Helps iQiyi Deliver Streaming Videos - TiDB|website=pingcap.com|accessdate=2019-12-07|archive-date=2019-11-06|archive-url=https://web.archive.org/web/20191106031811/https://pingcap.com/success-stories/tidb-in-iqiyi/|dead-url=no}}</ref>, 转转<ref>{{cite web|url=https://pingcap.com/success-stories/tidb-in-zhuanzhuan/|title=Managing the Surging Data Volume of a Fast-Growing Marketplace with TiDB - TiDB|website=pingcap.com|accessdate=2019-12-07|archive-date=2019-11-06|archive-url=https://web.archive.org/web/20191106031812/https://pingcap.com/success-stories/tidb-in-zhuanzhuan/|dead-url=no}}</ref>、摩拜<ref>{{cite web|url=https://pingcap.com/success-stories/tidb-in-mobike/|title=Blitzscaling the Largest Dockless Bikesharing Platform with TiDB’s Help - TiDB|website=pingcap.com|accessdate=2019-12-07|archive-date=2019-05-02|archive-url=https://web.archive.org/web/20190502221343/https://pingcap.com/success-stories/tidb-in-mobike/|dead-url=no}}</ref>、易果<ref>{{cite web|url=https://www.datanami.com/2018/02/22/hybrid-database-capturing-perishable-insights-yiguo/|title=The Hybrid Database Capturing Perishable Insights at Yiguo|first=Austin TX United|last=States|date=2018-02-22|website=Datanami|accessdate=2019-12-07|archive-date=2019-11-06|archive-url=https://web.archive.org/web/20191106031819/https://www.datanami.com/2018/02/22/hybrid-database-capturing-perishable-insights-yiguo/|dead-url=no}}</ref>，以及猿辅导<ref>{{cite web|url=https://pingcap.com/success-stories/tidb-in-yuanfudao/|title=How TiDB tackles fast data growth and complex queries for yuanfudao.com - TiDB|website=pingcap.com|accessdate=2019-12-07|archive-date=2019-11-06|archive-url=https://web.archive.org/web/20191106031811/https://pingcap.com/success-stories/tidb-in-yuanfudao/|dead-url=no}}</ref>。

== 参考文献 ==
{{reflist|30em}}

{{Database}}

[[Category:开源软件]]
[[Category:数据库]]
[[Category:NewSQL]]