The MEM value in information_schema.processlist is very large

Translated
1

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

Original topic: information_schema.processlist MEM值很大

| username: 天锁斩月

【TiDB Usage Environment】Production
【TiDB Version】v6.1.1
【Encountered Problem】
Using “select * from information_schema.processlist order by MEM desc;” query, it was found that the MEM values are very large.

image
image1380×162 77.9 KB

2
| username: jansu-dev | Original post link

It should be an issue with the corresponding memTracker, possibly this issue → Should clean the memory usage of MemTracker when a hashJoin query ends · Issue #34571 · pingcap/tidb · GitHub
Fix PR → executor: fix processlist's Mem is not zero when HashJoin finished by wshwsh12 · Pull Request #34569 · pingcap/tidb · GitHub

3
| username: 顾刚-数据 | Original post link

Hello, have you found out the reason for this issue? We are also encountering it, especially with update and commit operations. However, even when there are only 2 rows of data in the table I am updating, this situation still occurs.

4
| username: 江湖故人 | Original post link

Bugs existing in version 4.0
mem field in information_schema.processlist display abnormal · Issue #18588 · pingcap/tidb · GitHub