Last Updated: 2013-11-01
Condition or Error
The latency of updates on the slaves is very high
First the reason and location of the delay should be identified. It is possible for replication data to have been replicated quickly, but applying the data changes is taking a long time. Using row-based replication may increase the latency due to the increased quantity of data that must be transferred.
Check the replication format:
shell> grep binlog_format /etc/my.cnf binlog_format=ROW
Slow slaves can be the cause, but it may require some configuration changes.