## 模块 2: 部署 TiDB 集群 ### 步骤 1: 查看 `solution-topology-ten-nodes.yaml` + `solution-topology-ten-nodes.yaml` 是用于部署 TiDB 集群的 yaml 文件。以下显示的 IP 地址会与您的有所不同。 ```bash cat solution-topology-ten-nodes.yaml ``` ```output global: user: "ec2-user" ssh_port: 22 deploy_dir: "/tidb-deploy" data_dir: "/tidb-data" arch: "amd64" server_configs: pd: replication.location-labels: ["region","zone","storage-performance"] replication.max-replicas: 3 pd_servers: - host: 10.90.2.84 client_port: 2379 peer_port: 2380 advertise_client_addr: "http://54.251.28.224:2379" config: server.labels: region: seattle zone: az1 - host: 10.90.2.134 client_port: 2379 peer_port: 2380 advertise_client_addr: "http://54.179.198.121:2379" config: server.labels: region: seattle zone: az2 - host: 10.90.2.214 client_port: 2379 peer_port: 2380 advertise_client_addr: "http://13.215.194.132:2379" config: server.labels: region: seattle zone: az3 tidb_servers: - host: 10.90.2.124 port: 4000 status_port: 10080 deploy_dir: /tidb-deploy/tidb-4000 log_dir: /tidb-deploy/tidb-4000/log - host: 10.90.2.219 port: 4000 status_port: 10080 deploy_dir: /tidb-deploy/tidb-4000 log_dir: /tidb-deploy/tidb-4000/log tikv_servers: - host: 10.90.3.239 port: 20160 status_port: 20180 config: server.labels: region: seattle zone: az1 storage-performance: high - host: 10.90.3.119 port: 20160 status_port: 20180 config: server.labels: region: seattle zone: az2 storage-performance: high - host: 10.90.3.181 port: 20160 status_port: 20180 config: server.labels: region: seattle zone: az3 storage-performance: high tiflash_servers: - host: 10.90.4.159 tcp_port: 9000 flash_service_port: 3930 deploy_dir: /tidb-deploy/tiflash-9000 data_dir: /tidb-data/tiflash-9000 log_dir: /tidb-deploy/tiflash-9000/log monitoring_servers: - host: 10.90.4.93 grafana_servers: - host: 10.90.4.93 alertmanager_servers: - host: 10.90.4.93 ``` ### 步骤 2. 检查节点上是否安装了 `numactl` + 所有节点上都应该已安装 `numactl`。我们只需登录到一个 `TiKV` 节点进行检查。 ```bash ssh -i ~/.ssh/pe-class-key-${LAB:REGION_NAME}.pem ec2-user@${HOST_KV1_PRIVATE_IP} ``` ```bash sudo yum -y install numactl ``` ```output Loaded plugins: extras_suggestions, langpacks, priorities, update-motd amzn2-core | 3.6 kB 00:00:00 60 packages excluded due to repository priority protections Package numactl-2.0.9-7.amzn2.x86_64 already installed and latest version Nothing to do ``` + 返回控制机: ```bash exit ``` ### 步骤 3. 检查集群部署的系统设置的先决条件 ```bash tiup cluster check ./solution-topology-ten-nodes.yaml \ --user ec2-user \ -i /home/ec2-user/.ssh/pe-class-key-${LAB:REGION_NAME}.pem ``` ```output + Detect CPU Arch Name - Detecting node 10.90.2.90 Arch info ... ⠙ Shell: host=10.90.2.90, sudo=false, command=`uname -m` - Detecting node 10.90.2.231 Arch info ... ⠙ Shell: host=10.90.2.231, sudo=false, command=`uname -m` - Detecting node 10.90.2.138 Arch info ... ⠙ Shell: host=10.90.2.138, sudo=false, command=`uname -m` - Detecting node 10.90.2.242 Arch info ... ⠙ Shell: host=10.90.2.242, sudo=false, command=`uname -m` ... 10.90.2.97 sysctl Fail fs.file-max = 376482, should be greater than 1000000 10.90.2.97 sysctl Fail net.core.somaxconn = 4096, should be greater than 32768 10.90.2.97 sysctl Fail net.ipv4.tcp_syncookies = 1, should be 0 10.90.2.97 sysctl Fail vm.swappiness = 60, should be 0 10.90.2.97 service Pass service firewalld not found, ignore 10.90.2.97 command Pass numactl: policy: default 10.90.2.97 timezone Pass time zone is the same as the first PD machine: UTC 10.90.2.97 os-version Pass OS is Amazon Linux 2 10.90.2.97 memory Pass memory size is 4096MB 10.90.2.97 limits Fail soft limit of 'nofile' for user 'ec2-user' is not set or too low 10.90.2.97 limits Fail hard limit of 'nofile' for user 'ec2-user' is not set or too low 10.90.2.97 limits Fail soft limit of 'stack' for user 'ec2-user' is not set or too low 10.90.2.97 selinux Pass SELinux is disabled 10.90.2.97 thp Fail THP is enabled, please disable it for best performance ``` ### 步骤 4. 对失败的检查启用自动修复 ```bash tiup cluster check ./solution-topology-ten-nodes.yaml \ --user ec2-user \ -i /home/ec2-user/.ssh/pe-class-key-${LAB:REGION_NAME}.pem \ --apply ``` ```output ... 10.90.3.79 cpu-governor Warn Unable to determine current CPU frequency governor policy, auto fixing not supported 10.90.3.79 sysctl Fail will try to set 'fs.file-max = 1000000' 10.90.3.79 sysctl Fail will try to set 'net.core.somaxconn = 32768' 10.90.3.79 sysctl Fail will try to set 'net.ipv4.tcp_syncookies = 0' 10.90.3.79 sysctl Fail will try to set 'vm.swappiness = 0' + Try to apply changes to fix failed checks - Applying changes on 10.90.2.231 ... Done - Applying changes on 10.90.2.138 ... Done - Applying changes on 10.90.2.242 ... Done - Applying changes on 10.90.2.158 ... Done - Applying changes on 10.90.1.56 ... Done - Applying changes on 10.90.4.138 ... Done - Applying changes on 10.90.2.90 ... Done - Applying changes on 10.90.1.169 ... Done - Applying changes on 10.90.4.126 ... Done - Applying changes on 10.90.2.95 ... Done ``` ### 步骤 5. 部署一个名为 `tidb-test` 的 TiDB 集群 ```bash tiup cluster deploy tidb-test 7.5.0 ./solution-topology-ten-nodes.yaml \ --user ec2-user \ -i /home/ec2-user/.ssh/pe-class-key-${LAB:REGION_NAME}.pem \ --yes ``` ```output + Detect CPU Arch Name - Detecting node 10.90.2.90 Arch info ... ⠙ Shell: host=10.90.2.90, sudo=false, command=`uname -m` - Detecting node 10.90.2.231 Arch info ... ⠙ Shell: host=10.90.2.231, sudo=false, command=`uname -m` - Detecting node 10.90.2.138 Arch info ... ⠙ Shell: host=10.90.2.138, sudo=false, command=`uname -m` - Detecting node 10.90.2.242 Arch info ... ⠙ Shell: host=10.90.2.242, sudo=false, command=`uname -m` ... Cluster `tidb-test` deployed successfully, you can start it with command: `tiup cluster start tidb-test --init` ```