Abnormal secondsBehindMaster Metric in DM Tool

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

Original topic: DM工具的secondsBehindMaster指标异常

| username: 特雷西-迈克-格雷迪

[TiDB Usage Environment] Production Environment / Testing / PoC
[TiDB Version] v6.1.2
[Reproduction Path]
For an alter table add column xx, add column xxx. DM reported that the index already exists, and after skipping the error, problems occurred.

[Encountered Problem: Phenomenon and Impact]
For an alter table add column xx, add column xxx. DM reported that the index already exists. After I skipped the error binlog, synchronization was normal but secondsBehindMaster kept increasing.

Checked the dm-worker.log, after handling the DM error, there were indeed no “meet heartbeat event and then flush jobs” log records (time point A); before (time point A), there were “meet heartbeat event…” logs.
[Resource Configuration]
[Attachment: Screenshot/Log/Monitoring]

| username: Jjjjayson_zeng | Original post link

I haven’t encountered this situation before.
How about trying to put the binlog locally?

| username: tidb狂热爱好者 | Original post link

This issue won’t occur. How could binlog record the creation of two identical indexes?

| username: 考试没答案 | Original post link

Please upload a screenshot of your error so we can take a look.

| username: 考试没答案 | Original post link

Also, please share the status of this synchronization task.

| username: 特雷西-迈克-格雷迪 | Original post link

This was the previous error. Now DM is synchronizing normally, but the secondsBehindMaster keeps increasing.

| username: 特雷西-迈克-格雷迪 | Original post link

So it could be a bug.

| username: 考试没答案 | Original post link

Is the secondsBehindMaster still increasing?

| username: 特雷西-迈克-格雷迪 | Original post link

Still increasing