Not much info on the web out there about this.
I did see something on Niagara Central, which suggested turning off "Switch to sync io", so I did that.
So far so good, but I'm not sure how long it took to fail before.
I've been having this problem for a while now, and it seems to be increasing in frequency.
I have a bunch of points from some PLCs on our Kepware OPC server. In our Niagara AX G3 ( Version 3.6.34 ) system, I bring these points in via Drivers -> OPCNetwork.
Basically, some of the points eventually fail to update. On the bad ones, the "Opc Quality" still shows as "Good", but the "To String" will show "{fault}", with a "fault cause" of "Read fault: Read fail: null".
It's not always the same points that fail. To get the points to come back, I select the point, "Edit", change "Enabled" from "true" to "false" and click okay. After that, I repeat the process changing it back to "true" and the point is fine (at least for a while).
Any ideas? Do I need to increase the update rate up from 1000ms? There are probably about 100 of these OPC points total that I'm dealing with here.
I don't really understand why the OPC Quality would be good, but it would have a fault.
The attached graphic shows what I'm talking about.
Not much info on the web out there about this.
I did see something on Niagara Central, which suggested turning off "Switch to sync io", so I did that.
So far so good, but I'm not sure how long it took to fail before.
Dang, I am still losing points, but the failure mode is now different since I turned off "Switch to sync io". Now it doesn't show the read fault errors, but the points just stop updating even though the status still shows {ok}. Using a direct OPC client shows that the values on the OPC server are different than the stale values that are on the G3 server when this happens.
I have some of these points set up to log to histories, and when I go look at the histories, the data will have stopped being collected, since the point stopped updating. If I go to the point, disable and then re-enable it, it comes back to life.
There may be a bigger problem with our G3 server, as I also seem to be losing data points which are brought over via OBIX from our R2 server. In a similar manner, I'll note that the history points have stopped, and a disable/re-enable on the data points seem to bring them back to life.
Anyone have any troubleshooting ideas so I can fix these problems?
Just a thought on this problem.
Have you tried reducing the batch limit to something like 50. As you only have 100 or so points then doing half at a time shouldnt cause too much problem with points going stale.
No dice.
The points lasted about a day as normal before stopping to update, and not showing any error when they failed. That's quite annoying when you think that the points are okay because no errors are showing, but then you check them against the values shown on a real OPC client and they don't match....
I suppose I better spend some more time over on Niagara Central doing research.
As an update to anyone who happened to be following this situation, it appears that we now have the points working.
As discussed elsewhere on here, the 3.6.47 release of AX had some improvements to the OPC functionality. Unfortunately, we're running a hybrid Vykon Security / AX workstation machine, and the only 3.6.47 update available was for straight AX, which wipes out all of the Vykon Security stuff if you install it.
Anyways, I was able to extract the 3.6.47 modules JAR files and put them in my modules directory, and then on somebody's suggestion, I also copied over the new OPC.dll file to the correct location. With those changes, the points have stayed alive for over a week now, which is a major accomplishment.
Thanks for the update!