Bug #113814 The Rpl_semi_sync_master_yes_tx and Rpl_semi_sync_master_no_tx are inaccurate

• Submitted: 31 Jan 2:42
• Modified: 31 Jan 13:15
• Reporter: karry zhang (OCA)
• Status: Verified
• Category: MySQL Server: Replication
• Severity: S3 (Non-critical)
• Version: MySQL8.0,5.7
• OS: Any
• CPU Architecture: Any

[31 Jan 2:42] karry zhang

Description:
We can see that Rpl_semi_sync_master_yes_tx is described in the document as follows:
 
Rpl_semi_sync_master_yes_tx

The number of commits that were acknowledged successfully by a replica.

But in the bug https://bugs.mysql.com/bug.php?id=113813 I reported, we can see Rpl_semi_sync_master_yes_tx increased to 7, but the corresponding commits have not been acknowledged by replica.

How to repeat:
You can see test case and result in https://bugs.mysql.com/bug.php?id=113813.

You can see the corresponding description in https://dev.mysql.com/doc/refman/8.0/en/server-status-variables.html#statvar_Rpl_semi_sync....

Suggested fix:
Rpl_semi_sync_master_yes_tx

The number of commits on master.

Or I can submit a patch to change the collection logic of Rpl_semi_sync_master_yes_tx so that Rpl_semi_sync_master_yes_tx conforms to the semantics.

[31 Jan 5:53] karry zhang

In fact, due to https://bugs.mysql.com/bug.php?id=113813, the Rpl_semi_sync_master_yes_tx and Rpl_semi_sync_master_no_tx are inaccurate. I think it's not a documentation problem, we need to fix https://bugs.mysql.com/bug.php?id=113813 or add some information in documentation to prompt the user for possible errors.

[31 Jan 13:15] MySQL Verification Team

Hello karry zhang,

Thank you for the report and feedback.

regards,
Umesh