MySQL Bugs: #93034: replication logging - please move to performance

Bug #93034	replication logging - please move to performance_schema
Submitted:	31 Oct 2018 18:24	Modified:	5 Oct 2022 7:18
Reporter:	Simon Mudd (OCA)	Email Updates:
Status:	Verified	Impact on me:	None
Category:	MySQL Server: Replication	Severity:	S4 (Feature request)
Version:	8.0.13	OS:	Any
Assigned to:		CPU Architecture:	Any
Tags:	logging, performance_schema, replication

Description:
I see these logs in my system every 2 minutes:

2018-10-31T17:52:42.509623Z 5 [Note] [MY-010559] [Repl] Multi-threaded slave statistics for channel '': seconds elapsed = 121; events assigned = 12525569; worker queues filled over overrun level = 0; waited due a Worker queue full = 0; waited due the total size = 0; waited at clock conflicts = 2730510184400 waited (count) when Workers occupied = 0 waited when Workers occupied = 0
2018-10-31T17:54:43.617469Z 5 [Note] [MY-010559] [Repl] Multi-threaded slave statistics for channel '': seconds elapsed = 121; events assigned = 12592129; worker queues filled over overrun level = 0; waited due a Worker queue full = 0; waited due the total size = 0; waited at clock conflicts = 2744546840600 waited (count) when Workers occupied = 0 waited when Workers occupied = 0
2018-10-31T17:56:43.001382Z 5 [Note] [MY-010559] [Repl] Multi-threaded slave statistics for channel '': seconds elapsed = 120; events assigned = 12658689; worker queues filled over overrun level = 0; waited due a Worker queue full = 0; waited due the total size = 0; waited at clock conflicts = 2759298694200 waited (count) when Workers occupied = 0 waited when Workers occupied = 0
2018-10-31T17:58:43.785779Z 5 [Note] [MY-010559] [Repl] Multi-threaded slave statistics for channel '': seconds elapsed = 120; events assigned = 12759041; worker queues filled over overrun level = 0; waited due a Worker queue full = 0; waited due the total size = 0; waited at clock conflicts = 2780638594600 waited (count) when Workers occupied = 0 waited when Workers occupied = 0
2018-10-31T18:00:44.268489Z 5 [Note] [MY-010559] [Repl] Multi-threaded slave statistics for channel '': seconds elapsed = 121; events assigned = 12834817; worker queues filled over overrun level = 0; waited due a Worker queue full = 0; waited due the total size = 0; waited at clock conflicts = 2797422939800 waited (count) when Workers occupied = 0 waited when Workers occupied = 0

This is noisy and not easy to read. Docs on the MySQL web site don't help much. I got pointed at https://www.percona.com/blog/2017/07/19/multi-threaded-slave-statistics/ which helps a lot more.  Thanks JF.

$ sudo grep -c MY-010559 /var/log/mysqld.log
200

not been running long but it’s noisy.

How to repeat:
setup replication, at least with write set and watch the noise.
Then wonder what it means.

Suggested fix:
1. At least i can turn off the noise with dragnet logging
2. However the information is useful. Please expose this via a performance_schema table.
3. If you don't stop the logging make the display interval configurable, including setting it to 0 meaning don't show anything.
4. Improve the documentation to mention this output or the P_S table if you create one so it can be better understood as https://dev.mysql.com/doc/refman/8.0/en/error-messages-server.html does not help.

Hello Simon,

Thank you for the feature request!

regards,
Umesh

I'm also facing this issue in 5.7.17 with MTS(Multi Thread slave)

FWIW in 8.0 you can make this go away. 

Load the dragnet filter and run:

SET GLOBAL log_error_services='log_filter_dragnet; log_sink_internal';
SET GLOBAL dragnet.log_error_filter_rules='IF err_code == 10559 then drop.';

To suppress this spam:

[mysqld]
log_error_suppression_list='MY-010559'

Indeed, but log_error_suppression_list was not available when I opened this bug report.

However, the new setting does help filter out such entries.

Actually I'm wrong.

https://dev.mysql.com/doc/refman/8.0/en/server-system-variables.html indicates it was released in 8.0.13. I believe I'd seen this previously and had not seen this new setting when I reported the issue.