018. TiDB upgrade for database management

Use TiUP for patch upgrade (HotFix)

-R : all -N : specified nodes

Upgrade all TiDB instances on the cluster:

tiup cluster path <cluster-name> /tmp/tidb-hotfix.tar.gz -R tidb

Replace one of the TiDB instances:

tiup cluster path <cluster-name> /tmp/tidb-hotfix.tar.gz _N ${
    
    Node_IP}:${
    
    Node_Port}

Version upgrade process

升级TiUP=》修改TiUP Cluster拓扑配置文件=》检查当前集群监控状况=》将集群升级到指定版本=》验证

Upgrade Preparation - Updating TiUP

tiup update --self # Upgrade TiUP version
tiup update cluster # Upgrade tiup cluster version

Upgrade Preparation - Edit TiUP Cluster

tiup cluster edit-config <cluster-name>

Upgrade Preparation - Cluster Monitoring Status Check

Check the current region health status of the cluster
tiup cluster check --cluster

Upgrade TiDB cluster

Non-stop upgrade
tiup cluster upgrade

downtime upgrade

停止集群
tiup cluster stop <cluster-name>
离线升级
tiup cluster upgrdate <cluster-name> <version> --offline
启动集群
tiup cluster start <cluster-name>

Verify the TiDB cluster upgrade result

Check the cluster version

tiup cluster display <cluster-name>

Upgrade FAQ

Question 1: The terminal reports an error during the upgrade. How to proceed with the upgrade after processing the error?
1. Check the operation record and find the ID of the failed upgrade operation record

tiup cluster audit

2. Retry the last upgrade operation record

tiup cluster replay <audit-id>

Question 2: The upgrade process takes a long time, if you skip this step to upgrade quickly

tiup cluster upgrade <cluster-name> <version> --force

Question 3: Upgrade and update peripheral tools such as pd-ctl

tiup install ctl:v5.0.0

おすすめ

転載: blog.csdn.net/wangzhicheng987/article/details/131045505