1 Reply Latest reply on Aug 17, 2018 11:39 AM by gregor

    Access specific member of a collective


      I have a question regarding reading data from a collective using ProcessBook, Coresight, PI Vision.


      Given the case:

      I have a collective <Collective123>, containing two servers, <PrimaryServer> and <SecondaryServer>

      I would like to look at data from tag <Tag-ABC-123>

      Obviously I will be able to see what ever is provided from the collective itself. Kind of referenced like <Collective123>.<Tag-ABC-123>

      The question is, will I be able to also look at the tag <Tag-ABC-123> from the two servers separately? Kind of referenced like <PrimaryServer>.<Tag-ABC-123> and <SecondaryServer>.<Tag-ABC-123> ?


      Case is to monitor if, for any reason, the data from PIBuffer is not properly read into both collective members.

        • Re: Access specific member of a collective

          Hello Øystein,


          The short answer is "Yes and No".

          PI ProcessBook allows you to switch the Collective member but I wouldn't know PI Vision, which is the successor of PI Coresight, allows to do this. This leaves us to PI ProcessBook and I assume you would like to look at both (all) Collective members simultaneously. I believe the Server object doesn't allow connecting to all members at the same time. The only option which I could think of is running multiple instances of PI ProcessBook, each connecting against a dedicated Collective member.


          There are however other ways to verify if PI Buffer Subsystem is fanning to all members of a PI Data Archive Collective. Buffering Manager (BufferingManager.exe) is designed for the purpose and informs you about issues with buffering. Please see KB00808 - Common Buffering problems for details and additional resources on the subject.


          PI Archive Subsystem exposes a Performance Counter 'Archived Events/sec' and PI Snapshot Subsystem offers 'Snapshots/sec'. I suggest collecting those to PI Points for all members of a PI Data Archive Collective and to use e.g. Asset Analytics to build totals and to write them back to PI Points which you can than use for monitoring with any client you like. This is to give you the big picture but in case the total amount of archived / snapshot events drift apart, you would like to know where the issue might be. You get the required numbers through the Performance Counters exposed by each PI Buffer Subsystem instance. If those indicate a problem with the Buffer Subsystem on a specific interface node, you will need to log on there anyhow and check what's going on e.g. using Buffering Manager.

          1 of 1 people found this helpful