TiDB internally panics with concurrent DML and DDL

[Lily2025]

Bug Report

Please answer these questions before submitting your issue. Thanks!

1. Minimal reproduce step (Required)

oltp_fun_002
1、run sysbench
2、add column for all tables
sysbench --mysql-user=root --db-driver=mysql --time=0 --report-interval=10 --mysql-db=test --mysql-host=tc-tidb.endless-oltp--tps-512791-1-934 --mysql-port=4000 --threads=512 --tables=32 --table-size=10000000 --mysql-ignore-errors=2013,1213,1105,8028,9007 --rand-type=pareto oltp_read_write run

2. What did you expect to see? (Required)

qps can not drop to zero

3. What did you see instead (Required)

QPS drop to zero

4. What is your TiDB version? (Required)

./tidb-server -V

Release Version: v5.0.6
Edition: Community
Git Commit Hash: 181b4c2
Git Branch: heads/refs/tags/v5.0.6
UTC Build Time: 2021-12-21 09:10:51
GoVersion: go1.13
Race Enabled: false
TiKV Min Version: v3.0.0-60965b006877ca7234adaced7890d7b029ed1306
Check Table Before Drop: false

[Lily2025]

/assign wjhuang2016

[Lily2025]

/severity Critical

[you06]

IMO, the problem is TiDB internally panics. This issue is already fixed in master by #26759. In my test, sysbench quits when meeting internal panic with error state = 'HY000': MySQL server has gone away, after restarting sysbench, it works well.

@Lily2025 may we change the issue title to "TiDB internally panics with concurrent DML and DDL"?

[github-actions]

Please check whether the issue should be labeled with 'affects-x.y' or 'fixes-x.y.z', and then remove 'needs-more-info' label.