Edge On Premise multi datacenter design

Hello

I am new to Apigee, and we just started using Apigee Edge on-premise AIO (single server) for our DEV/QA environments.

We are planning to go into production and just needed some help on the design.

Our infrastructure is in Azure and we are planning to deploy a 2 data center for active/active Apigee Edge solution.

We will be using Azure traffic manager for geo-location (US East and US West region)

As far as the data center, do we need to go to a 12 nodes cluster for active/active solution (is it 12 nodes on each data center? or 6 nodes on each data center) or can we run on a smaller footprint?

From the doc:
Two data centers, each consisting of the following:

Nodes 1 & 7: Cassandra, Edge UI, Management Server OpenLDAP, Zookeeper
Nodes 2 & 3, 8 & 9: Cassandra, Message Processor, Router, Zookeeper
Nodes 4 & 5, 10 & 11: Qpid Server
Node 6: Postgres Server (master in Data Center 1)
Node 12: Postgres Server (slave in Data Center 2)

Operation question, since it's an active/active, how do we make changes/create API proxies? Does it automatically get replicated between the datacenter? (via Postgres replication?)

0 1 388
1 REPLY 1

Postgres is just for analytics database, proxies kvm and all rum time info are stored in cassandra and zookeeper nodes.