two parameters, one is binlog_ignore_db. the other is replicate-ignore-db difference:
binlog_ignore_db The parameters are set on the main library, for example: Binlog-ignore_db=test, Then for Test all operations under the library (incremental deletions) are recorded so that slave in accepting the main library's Binlog File , the amount of files will be reduced.
Pros: Reduce network I/O , reducing slave End I/O Thread's I/O volume, Optimizing Replication Performance
replicate-ignore-db replicate-ignore-db=test, performance No binlog-ignore-db good (whether or not you want, copy the binlog will be I/O thread read to slave i/o slave relaylog
Advantages: Guarantee Master and the slave consistency of data.
Summary: Personal view of the size of the data volume, if you want to filter the database is large in the main library settings. If the database to be filtered is small, it is set from the library. Generally use the replicate-ignore-db= Library name
This article from the "Clear Sky" blog, declined reprint!
Mysql.master and slave a master multi-slave, Çeku transmission