I would like to set up a backup/disaster site for jboss, and I am wondering how to do this.
Basically we have two physical locations in our company. Primary Site A and Backup Site B
I would like to set up a primary jboss running on site A, and a backup JBOSS running on site B
each would have their own database. What I would like to do is set up a hot backup system where the replication would be happening on the JBoss level (not on the db level) and saved to the database. There would be no requests coming from the clients to the backup jboss server while the primary is up. There is also no need for a reverse replication when the failover occurs, this can be done manually at some maintanence window time.
This is really not a cluster setup, because, as far as I understand, a clustered JBOSS needs a single db instance (or db cluster to connect to). I have two databases only, which cannot be clustered (because of network latency issues). Can I use JbossCache for replicating entity beans and statefull session beans to the backup jboss? How do i set this up?
Interesting that you mention this. I'm currently working on the CacheLoader architecture, which will allow for exactly this scenarion (see jboss-head/cache/docs/design/cacheloader.txt).
It is not yet part of JBossCache, I will release version 1.1 of JBossCache with this feature before JavaOne this year (end of June).