Reducing the number of replicas (max-replicas) from 3 to 2 results in extremely slow removal speed

Translated
1

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

Original topic: 设置副本数(max-replicas)从3减少到2，移除速度特别慢

| username: magic

[TiDB Usage Environment] Testing
[TiDB Version] v6.5.0
[Reproduction Path] pd config set max-replica 2
[Encountered Issue: Symptoms and Impact]
After changing the number of replicas from 3 to 2, the PD CPU has been consistently at 200%, and the store region of the node has been continuously decreasing, but the process is particularly slow.
[Resource Configuration]
[Attachments: Screenshots/Logs/Monitoring]

PD CPU

image
image1380×538 59.3 KB

Region Related

image
image1380×552 60.5 KB

Slow Decrease

image
image1380×430 37.8 KB

Parameters
replica-schedule-limit 128
region-schedule-limit 8192

How to speed up the process?

2
| username: magic | Original post link

Currently, it seems that whether the cluster has data being written or not does not affect this speed.

3
| username: Kongdom | Original post link

Try referring to the official documentation:

4
| username: magic | Original post link

Tried it, no significant improvement.

5
| username: db_user | Original post link

For a test environment, use this to adjust the store limit; the effect should be quite noticeable.

6
| username: magic | Original post link

It took effect. Thank you, thank you.

7
| username: magic | Original post link

The image you provided is not visible. Please provide the text you need translated.

8
| username: 考试没答案 | Original post link

Learned a new trick.

9
| username: magic | Original post link

Use with caution in production environments.

10
| username: system | Original post link

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.