The Impact of the TiDB Parameter split-table on Performance

Translated
1

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

Original topic: TIDB 参数split-table对性能的影响

| username: residentevil

[TiDB Usage Environment] Production Environment
[TiDB Version] V 7.1.0
[Encountered Problem: Problem Phenomenon and Impact] If the number of tables in the TiDB cluster is relatively large [reaching 50,000+], I see there is a split-table parameter in the official documentation. I would like to ask if this parameter will have a performance impact in a scenario with more than 50,000 tables.

2
| username: h5n1 | Original post link

This way, at least one table will have one region, and excessive regions can also burden the cluster, such as region heartbeats, etc. Another feature of cross-table merging is to reduce the number of regions by placing small tables into one region.

3
| username: residentevil | Original post link

For 50,000 tables, is it recommended to set this parameter to FALSE?

4
| username: tidb菜鸟一只 | Original post link

It is recommended to set it to false.

5
| username: h5n1 | Original post link

You can keep this parameter at its default setting. In previous versions, the parameter schedule.enable-cross-table-merge needed to be set to true to control cross-table merging. In the current version, neither of these parameters needs to be adjusted.

6
| username: Fly-bird | Original post link

50,000 has no impact.