[Users] MySQL 2-way replication with SER

Marcello Lupo lupo at itspecialist.it
Thu Nov 17 10:03:30 CET 2005


Hi,
what you mean for 2-way replication in mysql??
The replica as A-->B-->A or the clustering of mysql 5??
I see limitation in both situation:

In A-->B-->A you have trouble in case of failure of one server  to get again a 
working environment being sure that all is syncrhonized well. If you can 
suggest a good way to accomplish it will be welcome. I'm trying to do the 
same thing.

As i know (may be i'm wrong and i will be happy to be) In clustering you do 
not have support for fereign keys and there are other kind of limitation.

Let me know if you can suggest a good way to do it.

I need the transactional functions, foreign keys and to be sure that data are 
flushed to disk almost immediately after the query executed (this is 
something that the cluster do not assure to you because it keep in memory lot 
of data and flush it to disk in big chunks)  for the application that run on 
top of SER/OpenSer.

The goal of my application is to have 2 OpenSer (balanced using SRV records or 
something like this) servers working on two separate mysql but dealing with 
the same data. For now the best solution for reliability that i found is to 
have 2 OpenSer servers to operate on the 1st mysql and the 2nd  in sync as 
slave with one way replication and in case of failure to move the IP address 
of mysql on the slave and restart the slave as master. All this with 
Heartbeat package.

Bye,
Marcello


>Hi Daryl,
>
>the 2-way replication in Mysql works and it might be used along with 
>OpenSER - depends of your configuration.
>
>regards,
>bogdan
>
>Daryl Sanders wrote:





More information about the sr-users mailing list