MySQL Bugs: #117657: [ERROR] [MY-013935] [Server] IO-layer timeout before wait

Bug #117657	[ERROR] [MY-013935] [Server] IO-layer timeout before wait_timeout was reached
Submitted:	10 Mar 16:47	Modified:	12 Mar 14:31
Reporter:	Serge Medelice	Email Updates:
Status:	Not a Bug	Impact on me:	None
Category:	MySQL Server	Severity:	S3 (Non-critical)
Version:	8.4.4	OS:	Other (Rocky Linux 8)
Assigned to:	MySQL Verification Team	CPU Architecture:	x86 (x86_64 / CPU 2)

Description:
After MySQL upgrade from LTS release 8.0.41 to 8.4.4, the following error message is seeing in the mysql server log:
[ERROR] [MY-013935] [Server] IO-layer timeout before wait_timeout was reached
This message repeats itself everyday since the upgrade at no regular time with no related application logs

How to repeat:
Unable to reproduce it

Hi I upgraded few test setups and I do not see this error in the log. Without a way to reproduce the problem there is not much we can do.

Thank you for your response.

I have also upgraded a Rocky Linux 8 test server without seeing this occurrence. I have seen in internet someone getting the same error message though (https://stackoverflow.com/questions/79486567/mysql-my-013935-io-layer-timeout-before-wait-...)

this is just an informational message that the connection dropped before it was expected to timeout from wait_timeout.

Probably a firewall or vpn, router terminated the connection early.

Consider setting a lower tcp_keepalive_time and see if it helps.
https://tldp.org/HOWTO/TCP-Keepalive-HOWTO/usingkeepalive.html

Hello. For your reference, I also encountered the same issue. 

In my case, I was using GCP(GCE), and the VPC idle timeout was defined as 10 minutes. Connections exceeding this timeout will be forcibly dropped. And i was using the C# MySQLConnector as the client, and its KeepAlive setting was set to the system default of 2 hours. I resolved it by changing this to 1 minute.

Hi MySQL Team, 

Considering the following mention from the previous reply:
"this is just an informational message that the connection dropped before it was expected to timeout from wait_timeout."

Is there any reason for having it as [ERROR]? Why not classify it as [Note] for "Aborted connection" error log messages? I believe that having it as [ERROR] brings more attention, hence more concern for the "informational message".