怎么从0到1部署一套TiDB本地集群

2025-03-16 技术教程

这篇文章主要讲解了“怎么从0到1部署一套TiDB本地集群”，文中的讲解内容简单清晰，易于学习与理解，下面请大家跟着小编的思路慢慢深入，一起来研究和学习“怎么从0到1部署一套TiDB本地集群”吧！

TiDB是一款开源的NewSQL数据库，我们看一下官方的描述：

TiDB 是 PingCAP 公司自主设计、研发的开源分布式关系型数据库，是一款同时支持在线事务处理与在线分析处理 (Hybrid Transactional and Analytical Processing, HTAP)的融合型分布式数据库产品，具备水平扩容或者缩容、金融级高可用、实时 HTAP、云原生的分布式数据库、兼容 MySQL 5.7 协议和 MySQL 生态等重要特性。目标是为用户提供一站式 OLTP (Online Transactional Processing)、OLAP (Online Analytical Processing)、HTAP 解决方案。TiDB 适合高可用、强一致要求较高、数据规模较大等各种应用场景。

这里面有几个关键点：

分布式关系型数据库

兼容MySQL5.7

支持HTAP(在线事务处理和在线分析处理)

对金融行业支持很好，支持高可用、强一致和大数据场景

基本概念

这里介绍一下TiDB中的几个重要概念：

PD：Placement Driver，是TiDB的一个总控节点，负责集群的整体调度外，全局ID生成，以及全局时间戳TSO(中心化授时)的生成。也就是说全局时钟在这个节点实现。

TiKV：TiDB 的存储层，是一个分布式事务型的键值数据库，满足ACID事务，使用Raft协议保证多副本一致性，还存储统计数据，

TiFlash：HTAP形态的关键组件，它是TiKV的列存扩展，在提供了良好的隔离性的同时，也兼顾了强一致性。

Monitor：TiDB监控组件

实验环境

由于我本地资源的限制，我们使用快速部署的方式。

TiDB快速部署的方式有2种：

第一种：使用 TiUP Playground 快速部署本地测试环境

适用场景：利用本地 Mac 或者单机 Linux 环境快速部署 TiDB 集群。可以体验 TiDB 集群的基本架构，以及 TiDB、TiKV、PD、监控等基础组件的运行。

第二种：使用TiUP cluster在单机上模拟生产环境部署步骤

希望用单台Linux服务器，体验TiDB最小的完整拓扑的集群，并模拟生产的部署步骤。

这里我采用第二种方式。

据官方描述，TiDB在CentOS 7.3做过大量的测试，建议在CentOS 7.3以上部署。

本地环境：VMware虚拟机，操作系统CentOS7.6

开始部署

我们按照官方的步骤来安装

1.关闭防火墙

systemctlstopfirewalldserviceiptablesstop

2.下载并安装 TiUP，命令和结果如下

[root@master~]#curl--proto'=https'--tlsv1.2-sSfhttps://tiup-mirrors.pingcap.com/install.sh|sh%Total%Received%XferdAverageSpeedTimeTimeTimeCurrentDloadUploadTotalSpentLeftSpeed1008697k1008697k004316k00:00:020:00:02--:--:--4318kWARN:addingrootcertificateviainternet:https://tiup-mirrors.pingcap.com/root.jsonYoucanrevokethisbyremove/root/.tiup/bin/7b8e153f2e2d0928.root.jsonSetmirrortohttps://tiup-mirrors.pingcap.comsuccessDetectedshell:bashShellprofile:/root/.bash_profile/root/.bash_profilehasbeenmodifiedtoaddtiuptoPATHopenanewterminalorsource/root/.bash_profiletouseitInstalledpath:/root/.tiup/bin/tiup===============================================Haveatry:tiupplayground===============================================

3.安装 TiUP 的 cluster 组件

首先声明全局的环境变量，不然找不到tiup命令：

source.bash_profile

执行安装cluster命令：

tiupcluster

输出如下：

[root@master~]#tiupclusterThecomponent`cluster`isnotinstalled;downloadingfromrepository.downloadhttps://tiup-mirrors.pingcap.com/cluster-v1.3.1-linux-amd64.tar.gz10.05MiB/10.05MiB100.00%13.05MiBp/sStartingcomponent`cluster`:/root/.tiup/components/cluster/v1.3.1/tiup-clusterDeployaTiDBclusterforproductionUsage:tiupcluster[command]AvailableCommands:checkPerformpreflightchecksforthecluster.deployDeployaclusterforproductionstartStartaTiDBclusterstopStopaTiDBclusterrestartRestartaTiDBclusterscale-inScaleinaTiDBclusterscale-outScaleoutaTiDBclusterdestroyDestroyaspecifiedclusterclean(EXPERIMENTAL)CleanupaspecifiedclusterupgradeUpgradeaspecifiedTiDBclusterexecRunshellcommandonhostinthetidbclusterdisplayDisplayinformationofaTiDBclusterpruneDestroyandremoveinstancesthatisintombstonestatelistListallclustersauditShowauditlogofclusteroperationimportImportanexistTiDBclusterfromTiDB-Ansibleedit-configEditTiDBclusterconfig.Willuseeditorfromenvironmentvariable`EDITOR`,defaultusevireloadReloadaTiDBcluster'sconfigandrestartifneededpatchReplacetheremotepackagewithaspecifiedpackageandrestarttheservicerenameRenametheclusterenableEnableaTiDBclusterautomaticallyatbootdisableDisablestartingaTiDBclusterautomaticallyatboothelpHelpaboutanycommandFlags:-h,--helphelpfortiup--sshstring(EXPERIMENTAL)Theexecutortype:'builtin','system','none'.--ssh-timeoutuintTimeoutinsecondstoconnecthostviaSSH,ignoredforoperationsthatdon'tneedanSSHconnection.(default5)-v,--versionversionfortiup--wait-timeoutuintTimeoutinsecondstowaitforanoperationtocomplete,ignoredforoperationsthatdon'tfit.(default120)-y,--yesSkipallconfirmationsandassumes'yes'Use"tiupclusterhelp[command]"formoreinformationaboutacommand.

4.调大sshd服务的连接数限制

这里需要使用root权限，具体修改/etc/ssh/sshd_config文件下面的参数配置：

MaxSessions20

改完后重启sshd：

[root@master~]#servicesshdrestartRedirectingto/bin/systemctlrestartsshd.service

5.编辑集群配置模板文件

这个文件我们命名为topo.yaml，内容如下：

##Globalvariablesareappliedtoalldeploymentsandusedasthedefaultvalueof##thedeploymentsifaspecificdeploymentvalueismissing.global:user:"tidb"ssh_port:22deploy_dir:"/tidb-deploy"data_dir:"/tidb-data"##Monitoredvariablesareappliedtoallthemachines.monitored:node_exporter_port:9100blackbox_exporter_port:9115server_configs:tidb:log.slow-threshold:300tikv:readpool.storage.use-unified-pool:falsereadpool.coprocessor.use-unified-pool:truepd:replication.enable-placement-rules:truereplication.location-labels:["host"]tiflash:logger.level:"info"pd_servers:-host:192.168.59.146tidb_servers:-host:192.168.59.146tikv_servers:-host:192.168.59.146port:20160status_port:20180config:server.labels:{host:"logic-host-1"}#-host:192.168.59.146#port:20161#status_port:20181#config:#server.labels:{host:"logic-host-2"}#-host:192.168.59.146#port:20162#status_port:20182#config:#server.labels:{host:"logic-host-3"}tiflash_servers:-host:192.168.59.146

这里有2点需要注意:

文件中的host是部署TiDB的服务器ip

ssh_port默认是22

官方文件的tikv_servers是3个节点，我这里设置成了只有1个节点，原因是本地配置多个节点时只有1个节点能启动成功

6.部署集群

部署集群的命令如下：

tiupclusterdeploy<cluster-name><tidb-version>./topo.yaml--userroot-p

上面的cluster-name是集群名称，tidb-version是指TiDB版本号，可以通过tiup list tidb这个命令来查看，这里使用v3.1.2，集群名称叫mytidb-cluster，命令如下：

tiupclusterdeploymytidb-clusterv3.1.2./topo.yaml--userroot-p

下面是部署时输出的日志：

[root@master~]#tiupclusterdeploymytidb-clusterv3.1.2./topo.yaml--userroot-pStartingcomponent`cluster`:/root/.tiup/components/cluster/v1.3.1/tiup-clusterdeploymytidb-clusterv3.1.2./topo.yaml--userroot-pPleaseconfirmyourtopology:Clustertype:tidbClustername:mytidb-clusterClusterversion:v3.1.2TypeHostPortsOS/ArchDirectories-------------------------------pd192.168.59.1462379/2380linux/x86_64/tidb-deploy/pd-2379,/tidb-data/pd-2379tikv192.168.59.14620160/20180linux/x86_64/tidb-deploy/tikv-20160,/tidb-data/tikv-20160tidb192.168.59.1464000/10080linux/x86_64/tidb-deploy/tidb-4000tiflash192.168.59.1469000/8123/3930/20170/20292/8234linux/x86_64/tidb-deploy/tiflash-9000,/tidb-data/tiflash-9000prometheus192.168.59.1469090linux/x86_64/tidb-deploy/prometheus-9090,/tidb-data/prometheus-9090grafana192.168.59.1463000linux/x86_64/tidb-deploy/grafana-3000Attention:1.Ifthetopologyisnotwhatyouexpected,checkyouryamlfile.2.Pleaseconfirmthereisnoport/directoryconflictsinsamehost.Doyouwanttocontinue?[y/N]:yInputSSHpassword:+GenerateSSHkeys...Done+DownloadTiDBcomponents-Downloadpd:v3.1.2(linux/amd64)...Done-Downloadtikv:v3.1.2(linux/amd64)...Done-Downloadtidb:v3.1.2(linux/amd64)...Done-Downloadtiflash:v3.1.2(linux/amd64)...Done-Downloadprometheus:v3.1.2(linux/amd64)...Done-Downloadgrafana:v3.1.2(linux/amd64)...Done-Downloadnode_exporter:v0.17.0(linux/amd64)...Done-Downloadblackbox_exporter:v0.12.0(linux/amd64)...Done+Initializetargethostenvironments-Prepare192.168.59.146:22...Done+Copyfiles-Copypd->192.168.59.146...Done-Copytikv->192.168.59.146...Done-Copytidb->192.168.59.146...Done-Copytiflash->192.168.59.146...Done-Copyprometheus->192.168.59.146...Done-Copygrafana->192.168.59.146...Done-Copynode_exporter->192.168.59.146...Done-Copyblackbox_exporter->192.168.59.146...Done+CheckstatusEnablingcomponentpdEnablinginstancepd192.168.59.146:2379Enablepd192.168.59.146:2379successEnablingcomponentnode_exporterEnablingcomponentblackbox_exporterEnablingcomponenttikvEnablinginstancetikv192.168.59.146:20160Enabletikv192.168.59.146:20160successEnablingcomponenttidbEnablinginstancetidb192.168.59.146:4000Enabletidb192.168.59.146:4000successEnablingcomponenttiflashEnablinginstancetiflash192.168.59.146:9000Enabletiflash192.168.59.146:9000successEnablingcomponentprometheusEnablinginstanceprometheus192.168.59.146:9090Enableprometheus192.168.59.146:9090successEnablingcomponentgrafanaEnablinginstancegrafana192.168.59.146:3000Enablegrafana192.168.59.146:3000successCluster`mytidb-cluster`deployedsuccessfully,youcanstartitwithcommand:`tiupclusterstartmytidb-cluster`

7.启动集群

命令如下：

tiupclusterstartmytidb-cluster

启动成功日志如下：

[root@master~]#tiupclusterstartmytidb-clusterStartingcomponent`cluster`:/root/.tiup/components/cluster/v1.3.1/tiup-clusterstartmytidb-clusterStartingclustermytidb-cluster...+[Serial]-SSHKeySet:privateKey=/root/.tiup/storage/cluster/clusters/mytidb-cluster/ssh/id_rsa,publicKey=/root/.tiup/storage/cluster/clusters/mytidb-cluster/ssh/id_rsa.pub+[Parallel]-UserSSH:user=tidb,host=192.168.59.146+[Parallel]-UserSSH:user=tidb,host=192.168.59.146+[Parallel]-UserSSH:user=tidb,host=192.168.59.146+[Parallel]-UserSSH:user=tidb,host=192.168.59.146+[Parallel]-UserSSH:user=tidb,host=192.168.59.146+[Parallel]-UserSSH:user=tidb,host=192.168.59.146+[Serial]-StartClusterStartingcomponentpdStartinginstancepd192.168.59.146:2379Startpd192.168.59.146:2379successStartingcomponentnode_exporterStartinginstance192.168.59.146Start192.168.59.146successStartingcomponentblackbox_exporterStartinginstance192.168.59.146Start192.168.59.146successStartingcomponenttikvStartinginstancetikv192.168.59.146:20160Starttikv192.168.59.146:20160successStartingcomponenttidbStartinginstancetidb192.168.59.146:4000Starttidb192.168.59.146:4000successStartingcomponenttiflashStartinginstancetiflash192.168.59.146:9000Starttiflash192.168.59.146:9000successStartingcomponentprometheusStartinginstanceprometheus192.168.59.146:9090Startprometheus192.168.59.146:9090successStartingcomponentgrafanaStartinginstancegrafana192.168.59.146:3000Startgrafana192.168.59.146:3000success+[Serial]-UpdateTopology:cluster=mytidb-clusterStartedcluster`mytidb-cluster`successfully

8.访问数据库

因为TiDB支持mysql客户端访问，我们使用sqlyog登录TiDB，用户名root，密码空，地址192.168.59.149，端口4000，如下图：

登录成功如下图，左侧我们可以看到TiDB自带的一些表：

9.访问TiDB的Grafana监控

访问地址如下：

http://192.168.59.146:3000/login

初始用户名/密码:admin/admin，登录进去后修改密码，成功后页面如下：

10.dashboard

TiDB v3.x版本没有dashboard，v4.0开始加入，访问地址如下：

http://192.168.59.146:2379/dashboard

11.查看集群列表

命令：tiup cluster list，结果如下：

12.查看集群拓扑结构

命令如下：

tiupclusterlist

输入命令后，我本地集群的输出如下：

[root@master/]#tiupclusterlistStartingcomponent`cluster`:/root/.tiup/components/cluster/v1.3.1/tiup-clusterlistNameUserVersionPathPrivateKey-----------------------------mytidb-clustertidbv3.1.2/root/.tiup/storage/cluster/clusters/mytidb-cluster/root/.tiup/storage/cluster/clusters/mytidb-cluster/ssh/id_rsaYouhavenewmailin/var/spool/mail/root[root@master/]#tiupclusterdisplaymytidb-clusterStartingcomponent`cluster`:/root/.tiup/components/cluster/v1.3.1/tiup-clusterdisplaymytidb-clusterClustertype:tidbClustername:mytidb-clusterClusterversion:v3.1.2SSHtype:builtinIDRoleHostPortsOS/ArchStatusDataDirDeployDir----------------------------------------------192.168.59.146:3000grafana192.168.59.1463000linux/x86_64Up-/tidb-deploy/grafana-3000192.168.59.146:2379pd192.168.59.1462379/2380linux/x86_64Up|L/tidb-data/pd-2379/tidb-deploy/pd-2379192.168.59.146:9090prometheus192.168.59.1469090linux/x86_64Up/tidb-data/prometheus-9090/tidb-deploy/prometheus-9090192.168.59.146:4000tidb192.168.59.1464000/10080linux/x86_64Up-/tidb-deploy/tidb-4000192.168.59.146:9000tiflash192.168.59.1469000/8123/3930/20170/20292/8234linux/x86_64Up/tidb-data/tiflash-9000/tidb-deploy/tiflash-9000192.168.59.146:20160tikv192.168.59.14620160/20180linux/x86_64Up/tidb-data/tikv-20160/tidb-deploy/tikv-20160Totalnodes:6

遇到的问题

安装TiDB v4.0.9版本，可以部署成功，但是启动报错，如果topo.yaml中配置了3个节点，启动报错，tikv只能启动成功一个，日志如下：

[root@master~]#tiupclusterstartmytidb-clusterStartingcomponent`cluster`:/root/.tiup/components/cluster/v1.3.1/tiup-clusterstartmytidb-clusterStartingclustermytidb-cluster...+[Serial]-SSHKeySet:privateKey=/root/.tiup/storage/cluster/clusters/mytidb-cluster/ssh/id_rsa,publicKey=/root/.tiup/storage/cluster/clusters/mytidb-cluster/ssh/id_rsa.pub+[Parallel]-UserSSH:user=tidb,host=192.168.59.146+[Parallel]-UserSSH:user=tidb,host=192.168.59.146+[Parallel]-UserSSH:user=tidb,host=192.168.59.146+[Parallel]-UserSSH:user=tidb,host=192.168.59.146+[Parallel]-UserSSH:user=tidb,host=192.168.59.146+[Parallel]-UserSSH:user=tidb,host=192.168.59.146+[Parallel]-UserSSH:user=tidb,host=192.168.59.146+[Parallel]-UserSSH:user=tidb,host=192.168.59.146+[Serial]-StartClusterStartingcomponentpdStartinginstancepd192.168.59.146:2379Startpd192.168.59.146:2379successStartingcomponentnode_exporterStartinginstance192.168.59.146Start192.168.59.146successStartingcomponentblackbox_exporterStartinginstance192.168.59.146Start192.168.59.146successStartingcomponenttikvStartinginstancetikv192.168.59.146:20162Startinginstancetikv192.168.59.146:20160Startinginstancetikv192.168.59.146:20161Starttikv192.168.59.146:20162successError:failedtostarttikv:failedtostart:tikv192.168.59.146:20161,pleasechecktheinstance'slog(/tidb-deploy/tikv-20161/log)formoredetail.:timedoutwaitingforport20161tobestartedafter2m0sVerbosedebuglogshasbeenwrittento/root/.tiup/logs/tiup-cluster-debug-2021-01-05-19-58-46.log.Error:run`/root/.tiup/components/cluster/v1.3.1/tiup-cluster`(wd:/root/.tiup/data/SLGrLJI)failed:exitstatus1

查看日志文件/tidb-deploy/tikv-20161/log/tikv.log，提示下面2个目录下找不到文件：

[2021/01/0605:48:44.231-05:00][FATAL][lib.rs:482]["called`Result::unwrap()`onan`Err`value:Os{code:2,kind:NotFound,message:\"Nosuchfileordirectory\"}"][backtrace="stackbacktrace:\n0:tikv_util::set_panic_hook::{{closure}}\natcomponents/tikv_util/src/lib.rs:481\n1:std::panicking::rust_panic_with_hook\natsrc/libstd/panicking.rs:475\n2:rust_begin_unwind\natsrc/libstd/panicking.rs:375\n3:core::panicking::panic_fmt\natsrc/libcore/panicking.rs:84\n4:core::result::unwrap_failed\natsrc/libcore/result.rs:1188\n5:core::result::Result<T,E>::unwrap\nat/rustc/0de96d37fbcc54978458c18f5067cd9817669bc8/src/libcore/result.rs:956\ncmd::server::TiKVServer::init_fs\natcmd/src/server.rs:310\ncmd::server::run_tikv\natcmd/src/server.rs:95\n6:tikv_server::main\natcmd/src/bin/tikv-server.rs:166\n7:std::rt::lang_start::{{closure}}\nat/rustc/0de96d37fbcc54978458c18f5067cd9817669bc8/src/libstd/rt.rs:67\n8:main\n9:__libc_start_main\n10:<unknown>\n"][location=src/libcore/result.rs:1188][thread_name=main]

如果配置一个节点，启动还是失败，启动日志我们截取后半段：

StartingcomponentpdStartinginstancepd192.168.59.146:2379Startpd192.168.59.146:2379successStartingcomponentnode_exporterStartinginstance192.168.59.146Start192.168.59.146successStartingcomponentblackbox_exporterStartinginstance192.168.59.146Start192.168.59.146successStartingcomponenttikvStartinginstancetikv192.168.59.146:20160Starttikv192.168.59.146:20160successStartingcomponenttidbStartinginstancetidb192.168.59.146:4000Starttidb192.168.59.146:4000successStartingcomponenttiflashStartinginstancetiflash192.168.59.146:9000Error:failedtostarttiflash:failedtostart:tiflash192.168.59.146:9000,pleasechecktheinstance'slog(/tidb-deploy/tiflash-9000/log)formoredetail.:timedoutwaitingforport9000tobestartedafter2m0sVerbosedebuglogshasbeenwrittento/root/.tiup/logs/tiup-cluster-debug-2021-01-06-20-02-13.log.

在/tidb-deploy/tiflash-9000/log中文件如下：

[2021/01/0620:06:26.207-05:00][INFO][mod.rs:335]["startingworkingthread"][worker=region-collector-worker][2021/01/0620:06:27.130-05:00][FATAL][lib.rs:482]["called`Result::unwrap()`onan`Err`value:Os{code:2,kind:NotFound,message:\"Nosuchfileordirectory\"}"][backtrace="stackbacktrace:\n0:tikv_util::set_panic_hook::{{closure}}\n1:std::panicking::rust_panic_with_hook\natsrc/libstd/panicking.rs:475\n2:rust_begin_unwind\natsrc/libstd/panicking.rs:375\n3:core::panicking::panic_fmt\natsrc/libcore/panicking.rs:84\n4:core::result::unwrap_failed\natsrc/libcore/result.rs:1188\n5:cmd::server::run_tikv\n6:run_proxy\n7:operator()\nat/home/jenkins/agent/workspace/optimization-build-tidb-linux-amd/tics/dbms/src/Server/Server.cpp:415\n8:execute_native_thread_routine\nat../../../../../libstdc++-v3/src/c++11/thread.cc:83\n9:start_thread\n10:__clone\n"][location=src/libcore/result.rs:1188][thread_name=<unnamed>]

试了v4.0.1版本，也是一样的问题，都是报找不到文件的错误。

感谢各位的阅读，以上就是“怎么从0到1部署一套TiDB本地集群”的内容了，经过本文的学习后，相信大家对怎么从0到1部署一套TiDB本地集群这一问题有了更深刻的体会，具体使用情况还需要大家实践验证。这里是亿速云，小编将为大家推送更多相关知识点的文章，欢迎关注！