Bug #117115 The server_id of the record written to the error log by semisync is incorrect

• Submitted: 7 Jan 7:12
• Modified: 7 Jan 8:27
• Reporter: karry zhang (OCA)
• Status: Verified
• Category: MySQL Server: Replication
• Severity: S3 (Non-critical)
• Version: 5.7.44, 8.0.40
• OS: Any
• CPU Architecture: Any
• Tags: Contribution

[7 Jan 7:12] karry zhang

Description:
When we set replica's server_id bigger than 2147483647, and enable semisync. We will find the follwing record in master-error.log:

2025-01-06T21:09:14.266515+08:00 8775 [Note] Start binlog_dump to master_thread_id(8775) slave_server(3306128045), pos(, 4)
2025-01-06T21:09:14.266531+08:00 8775 [Note] Start asynchronous binlog_dump to slave (server_id: -988839251), pos(, 4)
2025-01-06T21:09:28.933976+08:00 8775 [Note] Stop asynchronous binlog_dump to slave (server_id: -988839251)

The server-id output by the semisyn plugin is a negative number -988839251. In fact, the server-id is 3306128045.

How to repeat:
Enable semisync and set the replica server-id to be greater than 2147483647. You can observe the above bug in the source's master-error.log.

By the way, I gave the error log of MySQL 5.7. MySQL 8.0 has the same problem because the relevant logic is the same.

Suggested fix:
The reason for this problem is that the server-id is of type ulong, but when semisyn outputs it to the error log, it uses %d to format the output.

Change %d to %u.

[7 Jan 8:27] MySQL Verification Team

Hello karry zhang,

Thank you for the report and feedback.

regards,
Umesh

[7 Jan 8:40] karry zhang

Fix patch

(*) I confirm the code being submitted is offered under the terms of the OCA, and that I am authorized to contribute it.

Contribution: diff.txt (text/plain), 1.65 KiB.