I'm testing with OPC Expert - a software package to try and enable redundant opc connections for our clients. I can connect successfully with the redundant server served by the OPC Expert software, and browse successfully - including with UA Expert (without any errors), however any tag on Ignition that is not actively changing, turns bad after 5 or so seconds. Any tag that has a value changing, remains good, screen capture below for reference.
I edited/saved the ignition opc connection at the start of the capture. All of the tags were healthy at around 39 seconds, and started to go bad at around 90 seconds.
There's a bunch of PublishResponses with data change notifications indicating a StatusCode of BadLicenseExpired.
Not sure if it's OpcExpert applying that StatusCode or it's from one of the underlying servers.
There's also at least 50 tags for which the server is not recognizing the NodeId at all when CreateMonitoredItems is called (BadNodeIdUnknown). Based on the NodeIds it looks like it could be UDT params not evaluating or being incorrect (e.g. {OpcPrefix}GPMPPC_RTAC.eProgramState)
I've applied a trial license from the software provider now so those errors shouldn't be an issue, however I'm not sure why the bad node id errors are appearing on most tags.
I've got a few more threads to pull on, but would be interested to hear if you had any more thoughts. UA Expert can subscribe to tags and they remain good indefinitely, unsure if there's any peculiarities in this instance with the combination of OPC Expert and the Ignition OPCUA driver.
I think there’s something wrong with your UDT setup.
You can get another capture so we can be sure without the other errors in the way, but I don’t think your bad tags were ever good, they’re just failing to have monitored items created because the OPC Item Paths are incorrect.
OK I'll take another look at that to be sure, thanks. They were working before I placed this server in the middle, and they do work for the first scan of all tags - it's just after 10 seconds or so they all appear to time out and report node id not found.
I've captured again, from the point of pressing save on the OPC connection. At about 30s the data comes GOOD, at about 58s the static data begins to show BAD. The changing data continues being GOOD. Total record time of ~100s.
Same IP addresses as before.
I'm not too concerned about the tags that are showing incorrect opc item paths, this is an offline dev system and there are a few moving parts. At the moment just chasing reliable OPC redundancy.
A rep from OPC Expert has offered to jump on a call with IA so I will open a ticket as well.
I'm still not really seeing anything damning from either side. A couple BadWaitingForInitialData in the PublishReponses, that's it. There's a couple other IP addresses communicating with that server that are returning a lot of BadNodeIdUnknown though.
There's a logger in Ignition called opcua.client.OpcUaConnection.DataValueConsumer that you can set to TRACE and it will show you every value as it arrives from the OPC server. Maybe that will show us something that makes sense.
New revelation, updating all of my opc tags to be polled, means that the tag values forever remain good.
It seems like when we use subscribed tags, after ~10-20 seconds without a value change, the tag goes bad, with the bad node id unknown error. However when I trigger a manual poll on OPC Expert, which updates the timestamp, the tag on ignition comes good, going bad again after ~10 seconds.
Food for thought anyway. Kevin are there settings that might support this on Ignition?
