Note:
This topic has been translated from a Chinese forum by GPT and might contain errors.
Original topic: tidb数据备份问题
[TiDB Usage Environment] Production Environment / Testing / Poc
[TiDB Version]
[Reproduction Path] What operations were performed that caused the issue
[Encountered Issue: Issue Phenomenon and Impact]
[Resource Configuration] Go to TiDB Dashboard - Cluster Info - Hosts and take a screenshot of this page
[Attachments: Screenshots/Logs/Monitoring]
Is there a contradiction here? Does it support partial database/table backup and restore or not?
- You can restore a full backup, or partial databases/tables from a full backup, to restore the target cluster to the data state corresponding to that full backup.
- Based on backup data (full and log), you can specify any point in time to restore the target cluster to the data state of the backup cluster corresponding to that point in time (Point-in-time recovery, PITR).
What does the second sentence mean?
Documentation link: TiDB 备份与恢复概述 | PingCAP 文档中心
Full backup can specify tables and databases, but incremental backup cannot specify databases or tables. That should be the meaning.
What are the prerequisites for using snapshot data backup to restore data? If the original file has been deleted, can it still be restored?
Full backup recovery can specify databases/tables, but log backups cannot.
Full backup is possible, but PITR (Point-in-Time Recovery) is not. Someone has raised this requirement:
Regarding PITR not being able to restore data by table or database - TiDB Q&A Community (asktug.com)
This mentions two dimensions:
One is time-based recovery.
The other is recovery of full or partial databases/tables, without considering time.
It is indeed a bit confusing.
Question, please confirm with the official submission.
Use BR to restore a specified table or database.
Full backup can specify tables and databases, but incremental backup cannot specify databases or tables. That should be the meaning.
Full backup is possible, but PITR (Point-in-Time Recovery) is not. This is a basic fact in distributed databases, and none have implemented it (except OB).
br supports specifying tables, pitr does not.
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.