1.2.2 TiDB-Operator 部署本地测试环境

1.2.2.1 背景介绍

本小结介绍如何在个人电脑（Linux 或 MacOS）上采用 kind 方式在 Kubernetes 上部署 TiDB Operator 和 TiDB 集群。部署包含三个关键环节：

  1. 通过 kind 部署 K8s 集群
  2. 在 K8s 集群上部署 TiDB Operator
  3. 在 K8s 集群中部署 TiDB 集群

在部署前，请确认资源满足以下要求：

  • 内存 4GB+、CPU 至少 2 cores
  • Docker 17.03+
  • Go 1.10+
  • net.ipv4.ip_forward 设置为 1

1.2.2.2 通过 kind 部署 K8s 集群

  1. 下载自动化部署程序
  1. # cd /root & git clone --depth=1 https://github.com/pingcap/tidb-operator && cd tidb-operator
  1. 通过程序创建集群
  1. # cd /root/tidb-operator && hack/kind-cluster-build.sh

执行成功后会有如下关键提示信息:

  1. ########## success create cluster:[kind] ##########
  2. To start using your cluster, run:  
  3.     kubectl config use-context kind-kind
  1. 将 K8s 集群相关命令路径加入 PATH 路径
  1.     # export PATH=$PATH:/root/tidb-operator/output/bin/
  1. 验证 K8s 环境是否符合要求
  1.     # kubectl cluster-info
  2.       Kubernetes master is running at https://127.0.0.1:32771
  3.       KubeDNS is running at https://127.0.0.1:32771/api/v1/namespaces/kube-system/services/kube-dns:dns/proxy

输出以上信息，则说明 K8s 服务符合要求。

  1.     # helm version
  2.       Client: &version.Version{SemVer:"v2.9.1", GitCommit:"20adb27c7c5868466912eebdf6664e7390ebe710", GitTreeState:"clean"}
  3.       Server: &version.Version{SemVer:"v2.9.1", GitCommit:"20adb27c7c5868466912eebdf6664e7390ebe710", GitTreeState:"clean"}

输出以上信息，则说明 Helm 客户端与服务端都符合要求。

1.2.2.3 在 K8s 集群上部署 TiDB Operator

  1. 通过 helm 安装 TiDB Operator

创建 TiDB CRD

  1.     # kubectl apply -f https://raw.githubusercontent.com/pingcap/tidb-operator/master/manifests/crd.yaml && kubectl get crd tidbclusters.pingcap.com
  2.     customresourcedefinition.apiextensions.k8s.io/tidbclusters.pingcap.com unchanged
  3.     customresourcedefinition.apiextensions.k8s.io/backups.pingcap.com unchanged
  4.     customresourcedefinition.apiextensions.k8s.io/restores.pingcap.com unchanged
  5.     customresourcedefinition.apiextensions.k8s.io/backupschedules.pingcap.com unchanged
  6.     customresourcedefinition.apiextensions.k8s.io/tidbmonitors.pingcap.com unchanged
  7.     customresourcedefinition.apiextensions.k8s.io/tidbinitializers.pingcap.com unchanged
  8.     customresourcedefinition.apiextensions.k8s.io/tidbclusterautoscalers.pingcap.com unchanged
  9.     NAME                       CREATED AT
  10.     tidbclusters.pingcap.com   2020-03-06T13:38:32Z

下载 TiDB Operator 的 Helm chart 文件：

  1.     # mkdir -p /root/chart/
  3.     从 https://github.com/pingcap/tidb-operator/releases 下载 tidb-operator-chart-v1.0.6.tgz 文件放到 /root/chart/ 路径下
  5.     # cd /root/chart/ && tar xvf tidb-operator-chart-v1.0.6.tgz

将 /root/tidb-operator/charts/tidb-operator/values.yaml 文件内的 scheduler.kubeSchedulerImageName 值修改为 registry.cn-hangzhou.aliyuncs.com/google_containers/kube-scheduler 以加快镜像拉取速度。

  1. 安装 TiDB Operator
  1.     # helm install --namespace=tidb-admin  --name=tidb-operator /root/tidb-operator/charts/tidb-operator -f /root/tidb-operator/charts/tidb-operator/values.yaml
  2.     NAME:   tidb-operator
  3.     LAST DEPLOYED: Fri Mar  6 14:24:09 2020
  4.     NAMESPACE: tidb-admin
  5.     STATUS: DEPLOYED
  6.     ...
  1. 验证 Operator 运行状态
  1.     # kubectl get pods -n tidb-admin
  2.     NAME                                       READY   STATUS    RESTARTS   AGE
  3.     tidb-controller-manager-85d8d498bf-2n8km   1/1     Running   0          19s
  4.     tidb-scheduler-7c67d6c77b-qd54r            2/2     Running   0          19s

以上信息显示 Operator 运行正常。

1.2.2.4 在 K8s 集群中部署 TiDB 集群

  1. 下载 TiDB Cluster 的 helm chart 文件
  1.     # mkdir -p /root/chart/
  2.     从 https://github.com/pingcap/tidb-operator/releases 下载 tidb-cluster-chart-v1.0.6.tgz 文件放到 /root/chart/ 路径下
  1. 安装 TiDB Cluster
  1.     # cd /root/chart/ && tar xvf tidb-cluster-chart-v1.0.6.tgz
  2.     # helm install --namespace dba-test --name=test /root/tidb-operator/charts/tidb-cluster -f /root/tidb-operator/charts/tidb-cluster/values.yaml
  3.     NAME:   test
  4.     LAST DEPLOYED: Fri Mar  6 14:50:25 2020
  5.     NAMESPACE: dba-test
  6.     STATUS: DEPLOYED

以上信息显示 TiDB Cluster 部署正常

  1. 观察 TiDB 的 POD 状态
  1.     # kubectl get pods -n dba-test
  2.     NAME                              READY   STATUS    RESTARTS   AGE
  3.     test-discovery-668b48577c-lqqbz   1/1     Running   0          7m37s
  4.     test-monitor-5b586d8cb-227qx      3/3     Running   0          7m37s
  5.     test-pd-0                         1/1     Running   0          7m37s
  6.     test-pd-1                         1/1     Running   0          7m37s
  7.     test-pd-2                         1/1     Running   1          7m37s
  8.     test-tidb-0                       2/2     Running   0          6m18s
  9.     test-tidb-1                       2/2     Running   0          6m18s
  10.     test-tikv-0                       1/1     Running   0          6m58s
  11.     test-tikv-1                       1/1     Running   0          6m58s
  12.     test-tikv-2                       1/1     Running   0          6m58s

以上信息显示 TiDB Cluster 所有 Pod 全部运行正常。

  1. 访问 TiDB 集群
  1.     # nohup kubectl port-forward svc/test-tidb 4000:4000 --namespace=dba-test &
  2.     # yum install -y mysql
  3.     # mysql -h 127.0.0.1 -uroot -P4000
  4.     mysql -h 127.0.0.1 -P 4000 -uroot
  5.     Welcome to the MariaDB monitor.  Commands end with ; or \g.
  6.     Your MySQL connection id is 1
  7.     Server version: 5.7.25-TiDB-v3.0.5 MySQL Community Server (Apache License 2.0)
  8.     Copyright (c) 2000, 2018, Oracle, MariaDB Corporation Ab and others.
  9.     Type 'help;' or '\h' for help. Type '\c' to clear the current input statement.
  10.     MySQL [(none)]>

显示以上输出显示 TiDB 集群部署成功。