2 Replies Latest reply on Nov 23, 2017 4:33 PM by CambronP

    Filling Data gaps with PI Cloud

    CambronP

      Hi,

       

      Is it possible to create with PI Cloud Connect two suscriptions/publications that will write on the same tags. I have some holes in the data on the suscriber server. However this data is available on the publisher server. But I don't want to stop the current suscription. So can I create a second one that could fill the gaps?

        • Re: Filling Data gaps with PI Cloud
          pgonzalez

          Hello Philippe,

           

          It is not possible to have multiple subscriptions writing to the same tag. When s subscription is created PI Cloud Connect attempts to create the associated PI Points for any PI Point Data References included in the subscription. If another PI Point with the same name is not found then the Subscription ID is entered into the Extended Descriptor (exdesc) field for all of the PI Points created for that specific subscription; this is the mechanism to mark ownership of the PI Tag.However, if the service finds that there is already a PI Point on the destination PI System with the same name and the same Subscription ID (and this subscription has not been deleted in the customer's portal) into the Extended Descriptor, PI Cloud Connect wont write any values to the tag and generates and alarm.

           

          Why do you have data gaps on your system? When did they start? When was the publication created?

            • Re: Filling Data gaps with PI Cloud
              CambronP

              This publication was created in September. Last week, the suscriber stopped receiving data. This is not the first time I experienced a failure in the suscription. My solution is to re-install PI Could Connect on the concerned server. This is not a pretty solution, but it works so far. So I have restarted the publication. For some reasons I don't understand, for some tags, there a few gaps in the data.

               

              However, I don't want to restart the publications. This would be to long because I need the live actual data.

               

              Is there a way to start a new publication specifying the how the data should be transfered? Like with a priority on actual data?