On 2013-03-06 16:39, Ken Linder wrote:
> Hi all,
> I have a question about replication. We are not sure if this is a bug, but it
> certainly feels like it. Here is the simplified situation...
> We have a larger database (~50GB) that is replicated to a second MySQL machine as a
> sort-of live backup. It has over 200 million daily transactions and we need to keep a
> live backup for disaster planning.
> We have some other non-replicated databases on the Master MySQL machine for
> reporting, research work, etc... Here is the oddity...
> When the currently active database is not replicated, and an update/delete/insert is
> issued against a table in a replicated database, the change is not replicated.
> MachineA.ReplicatedDB (Master DB)
> MachineZ.ReplicatedDB (Slave DB)
> MachineA.NotReplicatedDB (Not Replicated DB)
> Doing this on MachineA...
> Use NotReplicatedDB;
> Insert into ReplicatedDB.SomeTable (SomeString) values('This is a test');
> Will insert a record in the MachineA.ReplicatedDB.SomeTable HOWEVER, that record
> will not be replicated into MachineZ.ReplicatedDB.SomeTable
> Bug? Feature? Side effect?
This is described here:
and this is how replication-filtering works.
This works different with row-based replication.
> Server version: 5.5.25a-log MySQL Community Server (GPL) by Remi
> Linux: 2.6.32-220.13.1.el6.x86_64 CentOS 6.2
> Ken Linder
> ANI Networks
Senior System Consultant
Nydalen Allé 37, N-0484 Oslo
+47 917 90 643 | +47 4000 4100
truls@stripped | www.basefarm.no
techincal exellence - caring for your business