When are using 2 level replication:
MASTER1 (mysql 5.6.16-log) <-----> MASTER2 (mysql 5.6.16-log) -----> SLAVE1 (Percona 5.6.39-83.1-log)
The SLAVE1 has been setup recently ( a month ago) and after initial replication setup we run checksum between MASTER2 and SLAVE1 and was good.
But later when we compared the the total row count of some core tables between MASTER2 and SLAVE1 there is a mismatch and checksum also show such difference
We identified a huge difference in the total row count (using “select count(*) from table”) and some rows are not even synced to SLAVE1. For getting more information I am attaching the my.cnf of SLAVE1.
The purpose of the new slave is to migrate the database to a new server running on Percona and due to this unusual behaviorI am stucked and under time pressure.
Your analysis and feedback on this is very helpful.
my.cnf_slave1.txt (4.28 KB)