Is the built-in restore also ten times faster if the 8.1br restore speed is ten times faster?

Note:
This topic has been translated from a Chinese forum by GPT and might contain errors.

Original topic: 8.1br还原速度快十倍那么内置的restore也快十倍吗

| username: tidb狂热爱好者

[TiDB Usage Environment] Production Environment / Testing / PoC
[TiDB Version]
[Reproduction Path] What operations were performed when the issue occurred
[Encountered Issue: Issue Symptoms and Impact]
[Resource Configuration] Go to TiDB Dashboard - Cluster Info - Hosts and take a screenshot of this page
[Attachments: Screenshots/Logs/Monitoring]
Can someone help confirm this?

| username: zhaokede | Original post link

There are no plans to upgrade to 8.1 yet. Those who have upgraded to 8.1 can test it out.

| username: zhanggame1 | Original post link

It needs to be tested in practice. I always take the documentation descriptions as advertising slogans; they are likely measured under specific conditions.

| username: lemonade010 | Original post link

At that time, an official performance test report will be released, and the results will be evident.

| username: Inkjade | Original post link

Just about to start testing with 7.5. Haven’t tried 8.1 yet. Can someone test the results?

| username: Jack-li | Original post link

Waiting for an official response.

| username: Kongdom | Original post link

:yum: Not specific, it definitely covers most situations.

| username: yytest | Original post link

In TiDB version 7.6.0, a new algorithm was introduced to enhance the BR snapshot recovery speed, experimentally increasing the recovery speed by up to 10 times. This indicates that in certain scenarios, significant speed improvements can indeed be achieved through optimized backup and restore techniques. Version 8.1 has just been released, and there are no related tests yet; we need to wait for official or production environment tests.

| username: FutureDB | Original post link

Official tests are generally conducted under relatively ideal hardware configurations and scenarios. For example, in the case of TiDB Lightning and Dumpling for import and export, the results obtained in most real-world scenarios will be less favorable compared to the official test data.

| username: zhanggame1 | Original post link

Waiting for the expert to test.

| username: 像风一样的男子 | Original post link

The performance of BR theoretically improves by 10 times, but I think the bottleneck will be stuck in disk I/O and server bandwidth.

| username: 呢莫不爱吃鱼 | Original post link

Theoretical performance is likely to differ from actual application. The hardware performance of the server itself also needs to be considered.