9 Replies Latest reply on Jul 26, 2018 8:32 AM by rahulbhavsar

    Interface node disconnection almost everyday

    rahulbhavsar

      Hi ALL,

       

      We are facing an issue recently, We have a PI OPC interface node which is getting the data from OPC server (at platform, exploration site) through Matrikon tunneller. Almost every other day we are facing disconnection of PI system with the platform data at around same time in the morning. The disconnection is for couple of minutes but when the link is back some of the tag shows jumped value say earlier it was 1.5 but now it is showing 13.1 or so. At the interface node we have matrikon explorer where we can see the opc source value which shows correct value even after the link is back i.e. say 1.5 and not 13.1. Please guide, let me know what details are required.

      Also for the same interface instance first scan class assigned is as per below picture: 99:59:59, 00:00:00. why so high ?

       

      Scanfrequency.png

        • Re: Interface node disconnection almost everyday
          rahulbhavsar

          Hi Guys,

           

          Please help in solving the issue, if anybody came across to similar issue kindly state your views.

           

          Thank you

          • Re: Interface node disconnection almost everyday
            John Messinger

            There are a number of additional questions to ask here.

             

            1. What is the reasoning behind such an unusual scan class configuration for scan class 1? You have configured a scan class for every 4 and a bit days here. Why?

            2. What does the local PI message log show around the time of your disconnect each day? This should be the first place to look.

            3. Is there anything unusual recorded in the Windows Event log around the time of the disconnect each day? This is the second place to look.

            4. What do the OPC Tunneler logs show at the time of the disconnect?

            5. Is the tunneler disconnecting from the OPC server, or is the interface dosconnecting from the PI Data Archive?

            6. Are your interface tags configured for Advise or Polling data collection? If they are Advise, how many are assigned to scan class 1?

             

            What other initial troubleshooting or investigation have you performed? It would be good to state this information up front to avoid going over areas you have already covered.

              • Re: Interface node disconnection almost everyday
                rahulbhavsar

                Hi,

                Sorry for late reply. The answers are:

                1. I am not sure why it is configured. Please let me know in which cases it is established like this, I can see the same scan rate in some other interfaces as well but no tag is mapped to this scan class.

                2. Image attached from the log, everytime it is same with same tag name which are hidden in the image.

                PIsdkerrorlog.png

                3. No information in the windows event log.

                4. OPC tunneller logs are not possible to look.

                5. Interface disconnecting from PI data archive

                6. There are both advise and polled. Most of them polled (issue lies in some of the polled tags only but not all). There is no tag mapped to scan class1

                 

                When we restart the PI services at interface, the data comes back to normal. Also currently i changed all the tags to advise to see if it is going to help.

                 

                Thank you