Hi, Beekshan.
How are you?
These errors typically indicate data inconsistency or logical/physical corruption, especially common in large environments with InnoDB, ROW replication, and partitioned tables. The cause is rarely network or GTID.
The safe approach is to identify drift with checksum, analyze the exact binlog event, validate table integrity, and, if necessary, resynchronize tables or the entire replica, avoiding skips.
In a controlled window, I recommend checking the tables in both the primary and secondary libraries:
--> CHECK TABLE table_name EXTENDED;
I recommend checking out these links:
https://dev.mysql.com/doc/refman/8.0/en/replication-features-partitioning.html
https://dev.mysql.com/doc/refman/8.0/en/replication-problems.html
“Replication is supported between partitioned tables as long as they use the same partitioning scheme…
Statements such as ALTER TABLE … DROP PARTITION might produce different results on the source and the replica.”
Sincerely,
Henrique Garcia.