If I'm to target zero downtime for historizing data, is the Mission Critical architecture the only way to go if we're keeping things within the Ignition platform (the historian is a separate box, assume it's alive & well) ? Can I have 3 nodes vs the 2 shown in the documentation? reference: Redundancy Architecture - Ignition User Manual 8.1 - Ignition Documentation
thanks in advance,
Cannot be done entirely within Ignition for sample intervals faster than 10-20 seconds. Ignition redundancy has both a non-zero detection time, typically ~10 seconds, but also a transition delay on the standby gateway that adds more time. On top of that, when an active gateway dies, there's no clear boundary on the alternate where to pick up its recordings.
Those numbers can be tweaked with both configuration settings and over-specified hardware, but I doubt you could get the dead time to less than a second. Even getting ≤5 seconds of dead time would be difficult.
In general, zero lost data can only be achieved with PLC buffers and a handshake protocol where Ignition signals to the PLC when critical data has been committed in the database (not possible with store and forward). This is true for every SCADA system with which I'm familiar.
Thank you for the quick & helpful response! Much appreciated!!