Currently I am having an issue when an edge node re-connects it will cause a value of 0 to be written to the historian database before writing the actual value a few milliseconds later. It seems to be more prominent on tags that do not change while the node is disconnected but it still does occur.
Looking at my stored data it seems to have been happening with every version of the Transmission and Engine module I have used (4.0.6-4.0.10). Really not sure if this is an issue with the MQTT modules or the Ignition Tag Historian so I was hoping someone here had seen and corrected this issue.
There are a number of caveats with regard to MQTT store and forward configuration. Have you been through this doc? MQTT History - MQTT Modules for Ignition 8.x - Confluence
Yes, that is actually what I used as the basis for setting up these tags, they are configured to write directly to the History provider. Also this was occurring at about the same frequency before and after changing the ‘Enable History Storage by Default’ setting and using the StoreAndForward custom property.
This is something that used to be fairly intermittent but has started to become more common separately from any module changes. It also doesn’t seem to leave any evidence in the log and doesn’t match up to any of the duplicate entry quarantines that occur from time to time.
@mgrishaber, we have the exact same issue. It messes with the historian accuracy quite a bit. I will let you know when we find a solution, but exact same story, also tried 4.0.8-4.0.10, but it keeps this way.
What versions of Ignition are you running? On my system the gateway with the Engine module is running 8.1.14 and the edge nodes with the Transmission module are running a few different versions, 8.1.5 to 8.1.12.
It’s not Ignition edge, but we have 2 plant leveI gnition servers, 1 central one. All are on 8.1.15