Netcool/OMNIbus is part of many, if not all, AIOps deployments due to the nature of the event sources. If you are deploying Netcool Probes as part of your AIOps solution, you will most likely also include ObjectServers as part of the design to implement automation, such as housekeeping and event flood control.
Most deployments include a 2-site georedundancy capability however on rare occasions, 3-site georedundancy is desired. This blog outlines a design for a 3-site "trinity" deployment of OMNIbus to provide such a solution.
Consider the following diagram:
NOTES:
- The "standard multitier architecture" is extended with a third ObjectServer: AGG_T
- AGG_T is configured as a backup Aggregation ObjectServer
- Two additional bidirectional Gateways are added, one being a cold standby
- The Netcool interfaces file supports multiple "Backup" targets for the virtual AGG_V
- AGG_P is the "Primary", AGG_B is the first "Backup", and AGG_T is a second "Backup"
- Failover and failback is automatic and seamless in the usual Netcool manner
- A set of carefully constructed configuration manages the various failover/failback scenarios
- Probe peering will still only work in pairs however, even though they can have 3 potential targets
A working prototype of the above design has been built and the configuration files to build it are available from here:
IMPORTANT: This design and set of configurations are provided as-is and is not intended for production use without thorough testing in a non-production environment with a representative data load. IBM accepts no liability for its use in any context and users download and use it at their own risk.