Deploy TiDB on Alibaba Cloud Kubernetes

This document describes how to deploy a TiDB cluster on Alibaba Cloud Kubernetes with your laptop (Linux or macOS) for development or testing.

To deploy TiDB Operator and the TiDB cluster in a self-managed Kubernetes environment, refer to Deploy TiDB Operator and Deploy TiDB in General Kubernetes.

Prerequisites

• aliyun-cli >= 3.0.15 and configure aliyun-cli

  Note

  The access key must be granted permissions to control the corresponding resources.

• kubectl >= 1.12

• Helm 3

• jq >= 1.6

• terraform 0.12.*

You can use Cloud Shell of Alibaba Cloud to perform operations. All the tools have been pre-installed and configured in the Cloud Shell of Alibaba Cloud.

Required privileges

To deploy a TiDB cluster, make sure you have the following privileges:

• AliyunECSFullAccess
• AliyunESSFullAccess
• AliyunVPCFullAccess
• AliyunSLBFullAccess
• AliyunCSFullAccess
• AliyunEIPFullAccess
• AliyunECIFullAccess
• AliyunVPNGatewayFullAccess
• AliyunNATGatewayFullAccess

Overview of things to create

In the default configuration, you will create:

• A new VPC

• An ECS instance as the bastion machine

• A managed ACK (Alibaba Cloud Kubernetes) cluster with the following ECS instance worker nodes:

  • An auto-scaling group of 2 * instances (2 cores, 2 GB RAM). The default auto-scaling group of managed Kubernetes must have at least two instances to host the whole system service, like CoreDNS
  • An auto-scaling group of 3 * ecs.g5.large instances for deploying the PD cluster
  • An auto-scaling group of 3 * ecs.i2.2xlarge instances for deploying the TiKV cluster
  • An auto-scaling group of 2 * ecs.c5.4xlarge instances for deploying the TiDB cluster
  • An auto-scaling group of 1 * ecs.c5.xlarge instance for deploying monitoring components
  • A 100 GB cloud disk used to store monitoring data

All the instances except ACK mandatory workers are deployed across availability zones (AZs) to provide cross-AZ high availability. The auto-scaling group ensures the desired number of healthy instances, so the cluster can auto-recover from node failure or even AZ failure.

Deploy

Deploy ACK, TiDB Operator and the node pool for TiDB cluster

1. Configure the target region and Alibaba Cloud key (you can also set these variables in the terraform command prompt):

   export TF_VAR_ALICLOUD_REGION=${REGION} && \
   export TF_VAR_ALICLOUD_ACCESS_KEY=${ACCESS_KEY} && \
   export TF_VAR_ALICLOUD_SECRET_KEY=${SECRET_KEY}

   The variables.tf file contains default settings of variables used for deploying the cluster. You can change it or use the -var option to override a specific variable to fit your need.

2. Use Terraform to set up the cluster.

   git clone --depth=1 https://github.com/pingcap/tidb-operator && \
   cd tidb-operator/deploy/aliyun

   You can create or modify terraform.tfvars to set the values of the variables, and configure the cluster to fit your needs. You can view the configurable variables and their descriptions in variables.tf. The following is an example of how to configure the ACK cluster name, the TiDB cluster name, the TiDB Operator version, and the number of PD, TiKV, and TiDB nodes.

   cluster_name = "testack"
   tidb_cluster_name = "testdb"
   tikv_count = 3
   tidb_count = 2
   pd_count = 3
   operator_version = "v1.3.2"

   • To deploy TiFlash in the cluster, set create_tiflash_node_pool = true in terraform.tfvars. You can also configure the node count and instance type of the TiFlash node pool by modifying tiflash_count and tiflash_instance_type. By default, the value of tiflash_count is 2, and the value of tiflash_instance_type is ecs.i2.2xlarge.

   • To deploy TiCDC in the cluster, set create_cdc_node_pool = true in terraform.tfvars. You can also configure the node count and instance type of the TiCDC node pool by modifying cdc_count and cdc_instance_type. By default, the value of cdc_count is 3, and the value of cdc_instance_type is ecs.c5.2xlarge.

   Note

   Check the operator_version in the variables.tf file for the default TiDB Operator version of the current scripts. If the default version is not your desired one, configure operator_version in terraform.tfvars.

   After the configuration, execute the following commands to initialize and deploy the cluster:

   terraform init

   Input “yes” to confirm execution when you run the following apply command:

   terraform apply

   If you get an error while running terraform apply, fix the error (for example, lack of permission) according to the error description and run terraform apply again.

   It takes 5 to 10 minutes to create the whole stack using terraform apply. Once the installation is complete, the basic cluster information is printed:

   Apply complete! Resources: 3 added, 0 changed, 1 destroyed.
   Outputs:
   bastion_ip = 47.96.174.214
   cluster_id = c2d9b20854a194f158ef2bc8ea946f20e
   kubeconfig_file = /tidb-operator/deploy/aliyun/credentials/kubeconfig
   monitor_endpoint = not_created
   region = cn-hangzhou
   ssh_key_file = /tidb-operator/deploy/aliyun/credentials/my-cluster-keyZ.pem
   tidb_endpoint = not_created
   tidb_version = v3.0.0
   vpc_id = vpc-bp1v8i5rwsc7yh8dwyep5

   Note

   You can use the terraform output command to get the output again.

3. You can then interact with the ACK cluster using kubectl or helm:

   export KUBECONFIG=$PWD/credentials/kubeconfig

   kubectl version

   helm ls

Deploy the TiDB cluster and monitor

1. Prepare the TidbCluster and TidbMonitor CR files:

   cp manifests/db.yaml.example db.yaml && cp manifests/db-monitor.yaml.example db-monitor.yaml

   To complete the CR file configuration, refer to TiDB Operator API documentation and Configure a TiDB Cluster.

   • To deploy TiFlash, configure spec.tiflash in db.yaml as follows:

     spec
       ...
       tiflash:
         baseImage: pingcap/tiflash
         maxFailoverCount: 0
         nodeSelector:
           dedicated: TIDB_CLUSTER_NAME-tiflash
         replicas: 1
         storageClaims:
         - resources:
             requests:
               storage: 100Gi
           storageClassName: local-volume
         tolerations:
         - effect: NoSchedule
           key: dedicated
           operator: Equal
           value: TIDB_CLUSTER_NAME-tiflash

     To configure other parameters, refer to Configure a TiDB Cluster.

     Modify replicas, storageClaims[].resources.requests.storage, and storageClassName according to your needs.

     Warning

     Since TiDB Operator will mount PVs automatically in the order of the items in the storageClaims list, if you need to add more disks to TiFlash, make sure to append the new item only to the end of the original items, and DO NOT modify the order of the original items.

   • To deploy TiCDC, configure spec.ticdc in db.yaml as follows:

     spec
       ...
       ticdc:
         baseImage: pingcap/ticdc
         nodeSelector:
           dedicated: TIDB_CLUSTER_NAME-cdc
         replicas: 3
         tolerations:
         - effect: NoSchedule
           key: dedicated
           operator: Equal
           value: TIDB_CLUSTER_NAME-cdc

     Modify replicas according to your needs.

   To deploy Enterprise Edition of TiDB/PD/TiKV/TiFlash/TiCDC, edit the db.yaml file to set spec.<tidb/pd/tikv/tiflash/ticdc>.baseImage to the enterprise image (pingcap/<tidb/pd/tikv/tiflash/ticdc>-enterprise).

   For example:

   spec:
     ...
     pd:
       baseImage: pingcap/pd-enterprise
     ...
     tikv:
       baseImage: pingcap/tikv-enterprise

   Note

   • Replace all the TIDB_CLUSTER_NAME in the db.yaml and db-monitor.yaml files with tidb_cluster_name configured in the deployment of ACK.
   • Make sure the number of PD, TiKV, TiFlash, TiCDC, or TiDB nodes is >= the replicas value of the corresponding component in db.yaml.
   • Make sure spec.initializer.version in db-monitor.yaml is the same as spec.version in db.yaml. Otherwise, the monitor might not display correctly.

2. Create Namespace:

   kubectl --kubeconfig credentials/kubeconfig create namespace ${namespace}

   Note

   You can give the namespace a name that is easy to memorize, such as the same name as tidb_cluster_name.

3. Deploy the TiDB cluster:

   kubectl --kubeconfig credentials/kubeconfig create -f db.yaml -n ${namespace} &&
   kubectl --kubeconfig credentials/kubeconfig create -f db-monitor.yaml -n ${namespace}

Note

If you need to deploy a TiDB cluster on ARM64 machines, refer to Deploy a TiDB Cluster on ARM64 Machines.

Access the database

You can connect the TiDB cluster via the bastion instance. All necessary information is in the output printed after installation is finished (replace the ${} parts with values from the output):

ssh -i credentials/${cluster_name}-key.pem root@${bastion_ip}

mysql --comments -h ${tidb_lb_ip} -P 4000 -u root

tidb_lb_ip is the LoadBalancer IP of the TiDB service.

Note

• The default authentication plugin of MySQL 8.0 is updated from mysql_native_password to caching_sha2_password. Therefore, if you use MySQL client from MySQL 8.0 to access the TiDB service (TiDB version < v4.0.7), and if the user account has a password, you need to explicitly specify the --default-auth=mysql_native_password parameter.
• By default, TiDB (starting from v4.0.2) periodically shares usage details with PingCAP to help understand how to improve the product. For details about what is shared and how to disable the sharing, see Telemetry.

Monitor

Visit <monitor-lb>:3000 to view the Grafana dashboards. monitor-lb is the LoadBalancer IP of the Monitor service.

The initial login user account and password:

• User: admin
• Password: admin

Warning

If you already have a VPN connecting to your VPC or plan to set up one, it is strongly recommended that you go to the spec.grafana.service.annotations section in the db-monitor.yaml file and set service.beta.kubernetes.io/alicloud-loadbalancer-address-type to intranet for security.

Upgrade

To upgrade the TiDB cluster, modify the spec.version variable by executing kubectl --kubeconfig credentials/kubeconfig patch tc ${tidb_cluster_name} -n ${namespace} --type merge -p '{"spec":{"version":"${version}"}}.

This may take a while to complete. You can watch the process using the following command:

kubectl get pods --namespace ${namespace} -o wide --watch

Scale out the TiDB cluster

To scale out the TiDB cluster, modify tikv_count, tiflash_count, cdc_count, or tidb_count in the terraform.tfvars file, and then run terraform apply to scale out the number of nodes for the corresponding components.

After the nodes scale out, modify the replicas of the corresponding components by running kubectl --kubeconfig credentials/kubeconfig edit tc ${tidb_cluster_name} -n ${namespace}.

Note

• Because it is impossible to determine which node will be taken offline during the scale-in process, the scale-in of TiDB clusters is currently not supported.
• The scale-out process takes a few minutes. You can watch the status by running kubectl --kubeconfig credentials/kubeconfig get po -n ${namespace} --watch.

Configure

Configure TiDB Operator

You can set the variables in terraform.tfvars to configure TiDB Operator. Most configuration items can be modified after you understand the semantics based on the comments of the variable. Note that the operator_helm_values configuration item can provide a customized values.yaml configuration file for TiDB Operator. For example:

• Set operator_helm_values in terraform.tfvars:

  operator_helm_values = "./my-operator-values.yaml"

• Set operator_helm_values in main.tf:

  operator_helm_values = file("./my-operator-values.yaml")

In the default configuration, the Terraform script creates a new VPC. To use the existing VPC, set vpc_id in variable.tf. In this case, Kubernetes nodes are not deployed in AZs with vSwitch not configured.

Configure the TiDB cluster

See TiDB Operator API Documentation and Configure a TiDB Cluster.

Manage multiple TiDB clusters

To manage multiple TiDB clusters in a single Kubernetes cluster, you need to edit ./main.tf and add the tidb-cluster declaration based on your needs. For example:

module "tidb-cluster-dev" {
  source = "../modules/aliyun/tidb-cluster"
  providers = {
    helm = helm.default
  }
  cluster_name = "dev-cluster"
  ack          = module.tidb-operator
  pd_count                   = 1
  tikv_count                 = 1
  tidb_count                 = 1
}
module "tidb-cluster-staging" {
  source = "../modules/aliyun/tidb-cluster"
  providers = {
    helm = helm.default
  }
  cluster_name = "staging-cluster"
  ack          = module.tidb-operator
  pd_count                   = 3
  tikv_count                 = 3
  tidb_count                 = 2
}

Note

You need to set a unique cluster_name for each TiDB cluster.

All the configurable parameters in tidb-cluster are as follows:

Manage multiple Kubernetes clusters

It is recommended to use a separate Terraform module to manage a specific Kubernetes cluster. (A Terraform module is a directory that contains the .tf script.)

deploy/aliyun combines multiple reusable Terraform scripts in deploy/modules. To manage multiple clusters, perform the following operations in the root directory of the tidb-operator project:

1. Create a directory for each cluster. For example:

   mkdir -p deploy/aliyun-staging

2. Refer to main.tf in deploy/aliyun and write your own script. For example:

   provider "alicloud" {
       region     = ${REGION}
       access_key = ${ACCESS_KEY}
       secret_key = ${SECRET_KEY}
   }
   module "tidb-operator" {
       source     = "../modules/aliyun/tidb-operator"
       region          = ${REGION}
       access_key      = ${ACCESS_KEY}
       secret_key      = ${SECRET_KEY}
       cluster_name    = "example-cluster"
       key_file        = "ssh-key.pem"
       kubeconfig_file = "kubeconfig"
   }
   provider "helm" {
       alias    = "default"
       insecure = true
       install_tiller = false
       kubernetes {
           config_path = module.tidb-operator.kubeconfig_filename
       }
   }
   module "tidb-cluster" {
       source = "../modules/aliyun/tidb-cluster"
       providers = {
           helm = helm.default
       }
       cluster_name = "example-cluster"
       ack          = module.tidb-operator
   }
   module "bastion" {
       source = "../modules/aliyun/bastion"
       bastion_name             = "example-bastion"
       key_name                 = module.tidb-operator.key_name
       vpc_id                   = module.tidb-operator.vpc_id
       vswitch_id               = module.tidb-operator.vswitch_ids[0]
       enable_ssh_to_worker     = true
       worker_security_group_id = module.tidb-operator.security_group_id
   }

You can customize this script. For example, you can remove the module "bastion" declaration if you do not need the bastion machine.

Note

You can copy the deploy/aliyun directory. But you cannot copy a directory on which the terraform apply operation is currently performed. In this case, it is recommended to clone the repository again and then copy it.

Destroy

1. Refer to Destroy a TiDB cluster to delete the cluster.

2. Destroy the ACK cluster by running the following command:

   terraform destroy

If the Kubernetes cluster is not successfully created, the destroy operation might return an error and fail. In such cases, manually remove the Kubernetes resources from the local state:

terraform state list

terraform state rm module.ack.alicloud_cs_managed_kubernetes.k8s

It may take a long time to finish destroying the cluster.

Note

You have to manually delete the cloud disk used by the components in the Alibaba Cloud console.

Limitation

You cannot change pod cidr, service cidr, and worker instance types once the cluster is created.