WSO2 ESB/GREG High Availability

Question

I’m attempting to configure WSO2 ESB/GREG in a High Availability configuration, as follows:

Two GREG-ESB pairs, installed/configured on two separate Solaris servers.

Each server has an instance of GREG 4.5.3 (port offset 0) and ESB 4.7.0 (port offset 1) installed in separate installation directories.

GREG installations are configured to use ORACLE with jdbc DataSource, both connecting to the same database/schema, so adding something to one GREG is visible in the other.

ESB installations are configured with remote GREGs from above (each on the same server) and pointing to the same ORACLE database/schema for configuration/governance registry artifacts.

Tribes synchronization is enabled on all 4 installations.

We plan to use our own Load Balancer to round-robin traffic to either one or the other ESB with the idea that if one of the Solaris servers is down, we still have the full functionality on the other.

I couldn’t find an example of such HA configuration in WSO2 documentation.

The questions are:

. did anyone attempt such configuration (did it work)?

. is it even possible?

Answer 1

You can refer this documentation for information on deploying a cluster of Governance Registry instances to achieve load balancing and high availability.

To avoid the risk of system downtime due to failure of the ELB, you can implement a fail-proof ELB deployment. This documentation explains how to implement failover using two identical ELB setups running in active and passive modes respectively.