Re: Setting up Circular replication in existing replication infrastructure

[Phil Marsay]

The nature of what I am doing means that there are many more read operations than writes. Some of the read operations are quite intensive - reporting on large amounts of data.

In this instance, surely it would be better to split the load over 2 servers?

I am not intending on putting all three in a circle.

Rather A and B in circular replication and C just replicating one way off one of those two (C is in a separate data center as an off site backup only)


The example I gave was, perhaps a bad one - I would not have writes going on to the same database across 2 different servers.

It would be split by DB - so users of DB A,B,C always use server A, users of DB D,E,F always use server B, so the issue of stale information on reads would not come up.