Hi,
I have two sets of redundant Ignition servers at the regional level, each collecting data from the same field devices (as part of disaster recovery measures). I’d like to configure MQTT on all four of these servers to publish data to a third set of Ignition servers at the country level.
The goal is for the country-level Ignition servers to subscribe to only one of the regional server pairs at a time. If the currently active regional server becomes unavailable, the system should automatically switch to the next available regional server.
Can this setup be achieved using the Cirrus Link MQTT Engine module?
Thanks
Afthab
This can be done, but it isn’t handled in exactly the way you describe. If your redundant Edge systems (at the regional level) are using Ignition redundancy and your ‘country level’ server is also using Ignition redundancy, this can all be handled automatically via the Transmission and Engine modules. I’d recommend taking a look at this tutorial:
In my setup, both regional redundant sets are running the same backup, and all tag paths are identical. Will this configuration work? Also, we have two national sets as well.
Are the backups live/running at the same time as the primary systems? Or are you using Ignition redundancy? If the latter, it will work. If the former, it would violate a number of Sparkplug rules and would not work.
Yes, it is the former. I understand the issue, and we are currently exploring alternative options, any help appreciated.
Thank you.
Is there a reason the architecture in the tutorial above would not work for you?
In our setup, we have 2 redundant ignition server pairs at the regional level, each consisting of 4 machines. Both servers in each redundant pair run the same SCADA system—including identical screens, tags, and even the underlying PLCs. These regional redundant pairs are required to send data to another set of redundant servers at the national level. Our goal is to achieve multi-node redundancy. I’ve observed that Siemens OA and Honeywell implement similar approaches—interestingly, Honeywell uses MQTT for this purpose.
That architecture could work with one of the two options off the top of my head:
- Use two different MQTT Servers - one for the primary system, another for the backup system.
- Use different Sparkplug IDs (i.e. different Group or Edge Node IDs) for the primary and backup systems.
