MySQL Bugs: #109552: TRUNCATE TABLE causes "Stopping / Starting purge" log entries

Bug #109552	TRUNCATE TABLE causes "Stopping / Starting purge" log entries
Submitted:	9 Jan 2023 12:52	Modified:	9 Jan 2023 13:23
Reporter:	Sami Ahlroos	Email Updates:
Status:	Verified	Impact on me:	None
Category:	MySQL Server: Logging	Severity:	S3 (Non-critical)
Version:	5.7	OS:	Any
Assigned to:		CPU Architecture:	Any

Description:
After upgrading to 5.7.40, every time TRUNCATE TABLE is executed, this gets logged in the error log:
2023-01-09T12:42:08.822717Z 2 [Note] InnoDB: Stopping purge
2023-01-09T12:42:08.824934Z 2 [Note] InnoDB: Resuming purge

How to repeat:
With log_error_verbosity=3, execute TRUNCATE TABLE and check error log:
----
mysql> create table t(id int primary key);
Query OK, 0 rows affected (0.00 sec)

mysql> truncate table t;
Query OK, 0 rows affected (0.01 sec)

mysql> truncate table t;
Query OK, 0 rows affected (0.00 sec)

mysql> truncate table t;
Query OK, 0 rows affected (0.00 sec)
----
Followed by:
----
# tail /var/log/mysqld.log
2023-01-09T12:41:17.788062Z 0 [Note] Failed to start slave threads for channel ''
2023-01-09T12:41:17.792762Z 0 [Note] Event Scheduler: Loaded 0 events
2023-01-09T12:41:17.792998Z 0 [Note] /usr/sbin/mysqld: ready for connections.
Version: '5.7.40-log'  socket: '/var/lib/mysql/mysql.sock'  port: 3306  MySQL Community Server (GPL)
2023-01-09T12:42:08.822717Z 2 [Note] InnoDB: Stopping purge
2023-01-09T12:42:08.824934Z 2 [Note] InnoDB: Resuming purge
2023-01-09T12:42:09.564364Z 2 [Note] InnoDB: Stopping purge
2023-01-09T12:42:09.566206Z 2 [Note] InnoDB: Resuming purge
2023-01-09T12:42:09.944561Z 2 [Note] InnoDB: Stopping purge
2023-01-09T12:42:09.946289Z 2 [Note] InnoDB: Resuming purge
#
----
On a system doing lots of truncating the error log is flooded with the above. This did not happen on 5.7.39.

I suspect this change is what causes this: https://github.com/mysql/mysql-server/commit/c7d0df6b872b60b7f0a67f69ecdc85c188792f7a#diff...
It adds a call to trx_purge_stop() in row0trunc.cc which seems to be the cause of these notes in the log.

Hi Mr. Ahiroos,

Thank you for your bug report.

We have repeated your test case with 5.7.40.

It is possible that this report will finish as a Documentation bug, but that is not our call to make.

Verified as reported.

The same behaviour was NOT observed with 8.0.31.

Hence, this is 5.7-only bug.