AnsweredAssumed Answered

Does the PI-AF collective feature requires special consideration in code?

Question asked by Noga on Feb 18, 2019
Latest reply on Feb 19, 2019 by rfox

I am new to the HA PI-AF collective feature, and I have two main questions I would be glad to get answers for-


I am writing an application that uses the AF SDK to connect to a single PI-AF server, create, update and delete AFObjects on that single server.

Now I wish to use the PI-AF collective feature and support that feature in my implementation.


My question is-

From an application point of view- Is there any need to consider the fact that the application connects to a PI-AF collective instead of a single PI-AF server?

Are there any adjustment needed to be made in code (Such as: different method for connecting to PI-AF server in a collective, updating AFObjects on the collective etc..)?

Or is the collective feature is "transparent" to the implementation?


I wrote in the past an application for transferring tags between two PI Archive servers (not in collectives), and later configured each server to be part of a collective,

My implementation there was not changed in order to support the collective feature- except for the connection to the collective where the name of the collective was passed to the connect() method instead of the single standing PI Archive server name.

So, Is this the same for PI-AF collective?


My second question is-

In the PI Archive collective feature there is a limitation as follows:

PI Points can be created or edited only on the primary member of a PI Collective. Points on secondary members are created and edited by replicating the primary's PIPOINT table to all the secondary member nodes.

Is there a similar limitation on the PI-AF collective (such as: AFbjects can be edited only on the primary member)?

Or any other limitation that arise when using the PI-AF collective?