2 Replies Latest reply on Jun 11, 2014 2:12 PM by steveglass
      • 1. Re: How to properly handle SRC_EVENT_UME_MESSAGE_NOT_STABLE LBMSourceEvent state
        New Member

        Vlad,

         

        The question that you are asking can only be answered by you.  Let me explain why I say that.

         

        That event means that a message that you have sent from a UMP source was reported as unrecoverable loss by the stores, or at least, enough of stores that the stability criteria was not meet.

         

        How that is handled depends on your needs.  The appropriate response can vary from "do nothing" to resending that particular message again and handling all potential duplicate data issues.  Another possibility would be to inform the upstream application that this particular message was not delivered.  Other possibilities exist as well,

         

        Given that multiple retry mechanisms had to fail to reach this point, recovering will be a challenge.

         

        Kind Regards,

        Sherwin Clutter

        • 2. Re: How to properly handle SRC_EVENT_UME_MESSAGE_NOT_STABLE LBMSourceEvent state
          New Member

          The source event SRC_EVENT_UME_MESSAGE_NOT_STABLE would not be in the 5.3.6 documentation because it did not exist in any of the 5.x versions of UMP. SRC_EVENT_UME_MESSAGE_NOT_STABLE was first introduced in UMP 6.0, alongside with the UMP feature Pro-active Retransmissions (Pro-RX).

           

          Is the question how to work with this callback? If so, I would recommend reading this KB article: https://mysupport.informatica.com/whitepapers/4/Pages/1/148936.aspx and also checking out its usage in the sample application umesrc.c (or .java or .cs for the Java and .NET API's)

           

          If the question is how should the application handle this particular event, then as Sherwin mentioned above, it is entirely up to your application requirements.

           

          Thanks,

          Steve Glass