Replication Question

#1

Hello!  I have a few questions!

I have followed the instructions on replication and I can see that is helps with:

Prompt disaster recovery - because if master goes down I can switch the slave into master, or copy the slave data to another master and resume.

But how does it help with:

high availability and fault tolerance? Unless the application is read only and no write, as long as the master goes down, the application will go down with it. Please do correct me if I'm wring so I can understand.

Also I am assuming the cluster is not designed for spreading out the load, because according to the doc, the following code will always connect to the master unless the master goes down.  So basically both read and write will always be going to the master and the load is never distributed to the slaves(unless master goes down), correct?

 

EntityManagerFactory emf = Persistence.createEntityManagerFactory(
    "objectdb://localhost:6000/my.odb;user=a;password=a|" +
    "objectdb://localhost:6001//localhost:6000/my.odb;user=b;password=b"
);

Main point I want to ask:

Can I conclude that ODB Clustering is:

-  Only for the purpose of data redundancy?(e.g. my data is constantly backed up to another location so I can recover easily) 

- Not for performance(spreading the load, not even read load)

- Not for Read+Write failover?

 

Thank you.

#2

Not exactly.

  1. You should be able to keep your systems partly working, with read-only operations supported.
  2. You can manually make the switch between the master and the slave to enable write operations as well.
ObjectDB Support

Reply