We are developing a web application (Spring mvc based app) that uses a MySQL DB. (let's call it the "original app")
The DB architecture uses a MASTER-SLAVE configuration with replication.
The web app uses the MASTER for the functionality.
We are now developing a Back-office application for our "original app". This back office is in itself a web app (spring mvc based).
In order to reduce the performance impact on the "original app" we would like to use the SLAVE replica as the DB for the back-office.
This replica DB will include unique tables (that will exists only on the SLAVE, to serve the back-office app).
Non of us ever worked in such architecture, and we are pretty intimidated by this, therefore we have some questions:
- Is this architecture even viable?
- In case of a replication-error , How hard will it be to recover? remember, that the SLAVE will consist of tables that are not in the MASTER
- Any other tips \ information we can use?
- We cannot reconstruct the slave-only tables from the replicated tables. No That is one of my biggest concerns. Is that a show stopper?