"packet too big" errors in replication are often a sign of corrupt
binary logs. If it's on a slave reading the relay log
often times flushing it will temporarily solve the problem. The easiest
way to flush the relay logs is to do a slave stop; change master to back
to the current file name and p
Hello.
In my opinion (it is based on looking through the source code, but I
can be wrong) - the max_allowed_packet should be at least more than
any event in the master's binary log. The tool for researching
binary logs is mysqlbinlog utility. See:
http://dev.mysql.com/doc/mysql/en/mysqlbi
