DM Incremental Data Validation Checkpoint Not Advancing

Translated
1

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

Original topic: dm 增量数据校验validation 位点不前进

| username: Vincent_Wang

[TiDB Usage Environment] Production Environment
[TiDB Version] 6.5.5
[Reproduction Path]
MySQL synchronizes data to TiDB, enabling incremental data validation.

[Encountered Problem: Phenomenon and Impact]

DM incremental data validation position does not advance, staying at the same position for several days, resulting in incorrect incremental validation results.
Syncer has already synchronized to binlog.002722, 487335044.
Validation has been stuck at binlog.002350, 73958530 and hasn’t moved.

image
image1186×1000 106 KB

[Resource Configuration] Go to TiDB Dashboard - Cluster Info - Hosts and take a screenshot of this page
[Attachments: Screenshots/Logs/Monitoring]

2
| username: dba远航 | Original post link

Is there no error log?

3
| username: DBAER | Original post link

Check the DM logs.

4
| username: Vincent_Wang | Original post link

There is this warning “context deadline exceeded,” and I don’t know how to handle it. Restarting the validation didn’t work.

5
| username: WinterLiu | Original post link

I don’t understand, waiting for an expert to help.

6
| username: TiDBer_JpdctRSx | Original post link

CPU, memory, or disk I/O resource constraints may cause the verification speed to slow down. Lagging does not necessarily mean the system has crashed.

7
| username: Vincent_Wang | Original post link

Resources are sufficient.

8
| username: WalterWj | Original post link

Try restarting the task. :thinking:

9
| username: Vincent_Wang | Original post link

Will changing this parameter in the configuration file take effect after a restart?

10
| username: WalterWj | Original post link

I see there is a direct command for task resumption. :thinking:

11
| username: Vincent_Wang | Original post link

This is the sync process.

12
| username: WalterWj | Original post link

Then pause first and start again?

13
| username: xmlianfeng | Original post link

First, check the logs to see if there is any abnormal information.

14
| username: xmlianfeng | Original post link

Pause the task and then resume it to see if it works. If it still doesn’t work, you can try rebuilding the task and record the binlog position information.