I am not sure if the question is still relevant. I will quickly skim through the solution that we have implemented, and, if there is interest, will provide more details.
Ensemble servers: Two data centers plus a cloud instance for quorum.
Active nodes: Two data centers, either active active or through "master:" component when needed.
All fabric machines are configured with GlusterFS, enabling the deployed components to by synced between nodes automatically.
DB: Oracle Active/Passive, with autofailover in the JDBC connection config.
Hope that helps.
Do you have any architecture diagram of this?
can you please provide architecture diagram for this.