Comments on: Is MySQL Statement-Based / Mixed Replication Really Safe? https://www.percona.com/blog/is-mysql-statement-based-mixed-replication-really-safe/ Tue, 13 Feb 2024 17:02:45 +0000 hourly 1 https://wordpress.org/?v=6.5.2 By: Sodjamts https://www.percona.com/blog/is-mysql-statement-based-mixed-replication-really-safe/#comment-10973820 Wed, 16 Nov 2022 23:30:23 +0000 https://www.percona.com/blog/?p=81099#comment-10973820 Any replication performance difference between mixed and row-based replication?
Did you guys do any stress test and compare them?

]]> By: Nicolás Rotunno https://www.percona.com/blog/is-mysql-statement-based-mixed-replication-really-safe/#comment-10973638 Sat, 04 Jun 2022 15:18:22 +0000 https://www.percona.com/blog/?p=81099#comment-10973638 Hi there, nice work with these findings, we had similar issues with MIXED format in my organization.

The issue itself is not the updates but the INSERT that failed, since all statements seem ACID complain, updates are a byproduct.
Check the datetime on master vs replica “10:13:37” “09:13:37”

Unless server date/time are the same this is bound to happen, even in different time zones, we solved it by setting them all in the same timezone as master. This had the pluss of also helping when checking server events outside mysql.

I can’t recall the exact config, but there was a setting (current time =NOW) where mysql master sent the value of current time inside the binlog to the replicas and the now() used that value.

]]>