概述

tidb为分布式的newSQLdb，我们结合operator和k8s 完成tidb的部署和扩容

硬盘使用local卷，如何管理local卷请翻阅 我的另一篇文章

https://www.jianshu.com/p/bfa204cef8c0

k8s机器信息 版本1.13

node1 10.16.16.119  master节点

node2 10.16.16.120 master节点

node3 10.16.16.68

node4 10.16.16.68

部署rcd资源

git clone https://github.com/pingcap/tidb-operator.git

cd /tidb-operator/

kubectl apply -f manifests/crd.yaml

安装TiDB Operator

查看可配置信息

helm inspect charts/tidb-operator

apiVersion: v1

description: tidb-operator Helm chart for Kubernetes

home: https://github.com/pingcap/tidb-operator

keywords:

- operator

- newsql

- htap

- database

- mysql

- raft

name: tidb-operator

sources:

- https://github.com/pingcap/tidb-operator

version: 0.1.0

---

# Default values for tidb-operator

# clusterScoped is whether tidb-operator should manage kubernetes cluster wide tidb clusters

# Also see rbac.create and controllerManager.serviceAccount

clusterScoped: true

# Also see clusterScoped and controllerManager.serviceAccount

rbac:

create: true

# operatorImage is TiDB Operator image

operatorImage: pingcap/tidb-operator:latest

imagePullPolicy: IfNotPresent

defaultStorageClassName: local-storage

controllerManager:

# With rbac.create=false, the user is responsible for creating this account

# With rbac.create=true, this service account will be created

# Also see rbac.create and clusterScoped

serviceAccount: tidb-controller-manager

logLevel: 2

replicas: 1

resources:

limits:

cpu: 250m

memory: 150Mi

requests:

cpu: 80m

memory: 50Mi

# autoFailover is whether tidb-operator should auto failover when failure occurs

autoFailover: false

# pd failover period default(5m)

pdFailoverPeriod: 5m

# tidb failover period default(5m)

tidbFailoverPeriod: 5m

scheduler:

# With rbac.create=false, the user is responsible for creating this account

# With rbac.create=true, this service account will be created

# Also see rbac.create and clusterScoped

serviceAccount: tidb-scheduler

logLevel: 2

replicas: 1

schedulerName: tidb-scheduler

resources:

limits:

cpu: 250m

memory: 150Mi

requests:

cpu: 80m

memory: 50Mi

kubeSchedulerImageName: gcr.io/google-containers/hyperkube

# This will default to matching your kubernetes version

# kubeSchedulerImageTag:

指定自身参数，改变指定的image，因为gcr。io国内无法访问

有两种类型的存储hhd和ssd

vi ./w11_tidb-operator.config

#这里使用的是local hhd

defaultStorageClassName: local-storage

#defaultStorageClassName: ssd-local-storage

scheduler:

kubeSchedulerImageName: googlecontainer/hyperkube

kubeSchedulerImageTag: v1.13.4

helm install charts/tidb-operator -f ./w11_tidb-operator.config  --name=tidb-operator --namespace=tidb-admin

kubectl get pods --namespace tidb-admin -l app.kubernetes.io/instance=tidb-operator

如果全部ready即为创建成功

创建TIDB集群

查看配置参数

helm inspect charts/tidb-cluster

过多这里不在罗列，主要分为

pd的设置

tidb的设置

kv的设置

monitor的设置

编写定义参数

clusterName: w11tidb

pd:

resources:

limits:

cpu: 4000m

memory: 4Gi

requests:

cpu: 1000m

memory: 1Gi

storage: 1Gi

tikv:

replicas: 3

resources:

limits:

cpu: 16000m

memory: 10Gi

storage: 100Gi

requests:

cpu: 1000m

memory: 2Gi

storage: 5Gi

tidb:

separateSlowLog: True

replicas: 3

resources:

limits:

cpu: 16000m

memory: 6Gi

requests:

cpu: 1000m

memory: 1Gi

monitor:

persistent: true

helm install charts/tidb-cluster -f ./w11.config --name=tidb-cluster --namespace=tidb

kubectl get pods --namespace tidb -l app.kubernetes.io/instance=tidb-cluster -o wide

看出是否全部ok

查看端口

kubectl get svc --namespace tidb

监控为w11tidb-grafana,tidb为w11tidb-tidb

连接tidb

监控

http://10.16.16.119:30976

密码账号全部 admin admin

在线扩容TIDB

编写定义参数

将tikv 和 tidb从3个节点扩用值6个 resource提高

clusterName: w11tidb

pd:

resources:

limits:

cpu: 4000m

memory: 4Gi

requests:

cpu: 1000m

memory: 1Gi

storage: 1Gi

tikv:

replicas: 6

resources:

limits:

cpu: 16000m

memory: 20Gi

storage: 100Gi

requests:

cpu: 1000m

memory: 2Gi

storage: 5Gi

tidb:

separateSlowLog: True

replicas: 6

resources:

limits:

cpu: 16000m

memory: 16Gi

requests:

cpu: 1000m

memory: 1Gi

monitor:

persistent: true

扩容

helm upgrade -f w11.config tidb-cluster ./charts/tidb-cluster/

观察节点情况

kubectl get pods --namespace tidb -l app.kubernetes.io/instance=tidb-cluster -o wide

两个节点扩用中，已有节点根据情况重启重新配置resource

slow收集的收集

开启单独记录slowlog

tidb:

separateSlowLog: True

更新集群

helm upgrade -f w11.config tidb-cluster ./charts/tidb-cluster/

kubectl logs -n${namespace}${tidbPodName}-c slowlog

或者用stern

stern -n${namespace}tidb -c slowlog

备份

创建root账号

kubectl --namespace tidb create secret generic backup-secret --from-literal=user=root --from-literal=password=

配置backup参数

helm inspect ./charts/tidb-backup/

vi ./w11_backup.config

clusterName: w11tidb

storage:

className: local-storage

size: 10Gi

部署备份

helm install --name w11backup ./charts/tidb-backup/ -f ./w11_backup.config --namespace tidb

查看job

kubectl get jobs --namespace default -l app.kubernetes.io/component=backup

查看备份信息

kubectl -n tidb get pv

kubectl -n tidb get pv local-pv-e85e3bd9 -o yaml

发现在node1上

创建crontabjob

创建root账号

kubectl --namespace tidb create secret generic backup-secret --from-literal=user=root --from-literal=password=

helm inspect ./charts/tidb-cluster/

根据自身配置一下参数

binlog:

pump:

create: false

replicas: 1

image: pingcap/tidb-binlog:v2.1.0

imagePullPolicy: IfNotPresent

logLevel: info

# storageClassName is a StorageClass provides a way for administrators to describe the "classes" of storage they offer.

# different classes might map to quality-of-service levels, or to backup policies,

# or to arbitrary policies determined by the cluster administrators.

# refer to https://kubernetes.io/docs/concepts/storage/storage-classes

storageClassName: local-storage

storage: 10Gi

# a integer value to control expiry date of the binlog data, indicates for how long (in days) the binlog data would be stored.

# must bigger than 0

gc: 7

# number of seconds between heartbeat ticks (in 2 seconds)

heartbeatInterval: 2

drainer:

create: false

image: pingcap/tidb-binlog:v2.1.0

imagePullPolicy: IfNotPresent

logLevel: info

# storageClassName is a StorageClass provides a way for administrators to describe the "classes" of storage they offer.

# different classes might map to quality-of-service levels, or to backup policies,

# or to arbitrary policies determined by the cluster administrators.

# refer to https://kubernetes.io/docs/concepts/storage/storage-classes

storageClassName: local-storage

storage: 10Gi

# parallel worker count (default 1)

workerCount: 1

# the interval time (in seconds) of detect pumps' status (default 10)

detectInterval: 10

# disbale detect causality

disableDetect: false

# disable dispatching sqls that in one same binlog; if set true, work-count and txn-batch would be useless

disableDispatch: false

# # disable sync these schema

ignoreSchemas: "INFORMATION_SCHEMA,PERFORMANCE_SCHEMA,mysql,test"

# if drainer donesn't have checkpoint, use initial commitTS to initial checkpoint

initialCommitTs: 0

# enable safe mode to make syncer reentrant

safeMode: false

# number of binlog events in a transaction batch (default 1)

txnBatch: 1

# downstream storage, equal to --dest-db-type

# valid values are "mysql", "pb", "kafka"

destDBType: pb

mysql: {}

# host: "127.0.0.1"

# user: "root"

# password: ""

# port: 3306

# # Time and size limits for flash batch write

# timeLimit: "30s"

# sizeLimit: "100000"

kafka: {}

# only need config one of zookeeper-addrs and kafka-addrs, will get kafka address if zookeeper-addrs is configed.

# zookeeperAddrs: "127.0.0.1:2181"

# kafkaAddrs: "127.0.0.1:9092"

# kafkaVersion: "0.8.2.0"

scheduledBackup:

create: false

binlogImage: pingcap/tidb-binlog:v2.1.0

binlogImagePullPolicy: IfNotPresent

# https://github.com/tennix/tidb-cloud-backup

mydumperImage: pingcap/tidb-cloud-backup:latest

mydumperImagePullPolicy: IfNotPresent

# storageClassName is a StorageClass provides a way for administrators to describe the "classes" of storage they offer.

# different classes might map to quality-of-service levels, or to backup policies,

# or to arbitrary policies determined by the cluster administrators.

# refer to https://kubernetes.io/docs/concepts/storage/storage-classes

storageClassName: local-storage

storage: 100Gi

# https://kubernetes.io/docs/tasks/job/automated-tasks-with-cron-jobs/#schedule

schedule: "0 0 * * *"

# https://kubernetes.io/docs/tasks/job/automated-tasks-with-cron-jobs/#suspend

suspend: false

# https://kubernetes.io/docs/tasks/job/automated-tasks-with-cron-jobs/#jobs-history-limits

successfulJobsHistoryLimit: 3

failedJobsHistoryLimit: 1

# https://kubernetes.io/docs/tasks/job/automated-tasks-with-cron-jobs/#starting-deadline

startingDeadlineSeconds: 3600

# https://github.com/maxbube/mydumper/blob/master/docs/mydumper_usage.rst#options

options: "--chunk-filesize=100"

# secretName is the name of the secret which stores user and password used for backup

# Note: you must give the user enough privilege to do the backup

# you can create the secret by:

# kubectl create secret generic backup-secret --from-literal=user=root --from-literal=password=

secretName: backup-secret

# backup to gcp

gcp: {}

# bucket: ""

# secretName is the name of the secret which stores the gcp service account credentials json file

# The service account must have read/write permission to the above bucket.

# Read the following document to create the service account and download the credentials file as credentials.json:

# https://cloud.google.com/docs/authentication/production#obtaining_and_providing_service_account_credentials_manually

# And then create the secret by: kubectl create secret generic gcp-backup-secret --from-file=./credentials.json

# secretName: gcp-backup-secret

# backup to ceph object storage

ceph: {}

# endpoint: ""

# bucket: ""

# secretName is the name of the secret which stores ceph object store access key and secret key

# You can create the secret by:

# kubectl create secret generic ceph-backup-secret --from-literal=access_key= --from-literal=secret_key=

# secretName: ceph-backup-secret

我这里配置如下：

在cluster的配置文件后面追加我们这里是w11.condig

vi ./w11.config

###

#tidb的参数忽略

###

binlog:

pump:

create: true

replicas: 1

image: pingcap/tidb-binlog:v2.1.0

imagePullPolicy: IfNotPresent

logLevel: info

storageClassName: local-storage

storage: 10Gi

gc: 7

heartbeatInterval: 2

drainer:

create: true

image: pingcap/tidb-binlog:v2.1.0

imagePullPolicy: IfNotPresent

logLevel: info

storageClassName: local-storage

storage: 10Gi

# parallel worker count (default 1)

workerCount: 10

# the interval time (in seconds) of detect pumps' status (default 10)

detectInterval: 10

# disbale detect causality

disableDetect: false

# disable dispatching sqls that in one same binlog; if set true, work-count and txn-batch would be useless

disableDispatch: false

# # disable sync these schema

ignoreSchemas: "INFORMATION_SCHEMA,PERFORMANCE_SCHEMA,mysql,test"

# if drainer donesn't have checkpoint, use initial commitTS to initial checkpoint

initialCommitTs: 0

# enable safe mode to make syncer reentrant

safeMode: false

# number of binlog events in a transaction batch (default 1)

txnBatch: 10

# downstream storage, equal to --dest-db-type

# valid values are "mysql", "pb", "kafka"

destDBType: pb

mysql: {}

kafka: {}

scheduledBackup:

create: true

binlogImage: pingcap/tidb-binlog:v2.1.0

binlogImagePullPolicy: IfNotPresent

mydumperImage: pingcap/tidb-cloud-backup:latest

mydumperImagePullPolicy: IfNotPresent

storageClassName: local-storage

storage: 20Gi

schedule: "0 0 * * *"

创建备份任务

helm upgrade tidb-cluster ./charts/tidb-cluster/ -f ./w11.config

kubectl get pods --namespace tidb -l app.kubernetes.io/instance=tidb-cluster -o wide

查看增量备份pump日志

kubectl --namespace tidb logs w11tidb-pump-0

查看增量备份drain的日志

kubectl --namespace tidb logs w11tidb-drainer-0

查看全量crontab

kubectl get cronjob --namespace tidb -l app.kubernetes.io/instance=tidb-cluster