>Thanks for the explanation Larry but IMO when you release a session object the data session associated with that object should go away as well, but it doesn't if a object reference was created with a scatter name. Have you run the sample code that I provided? If so, do you get the same behavior?
>
>I'm I the only one that thinks this behavior isn't correct?

I can't answer that question, but you definately aren't looking at the behavior from the right direction. The session does not have a "containership" relationship to the object.

The relationship goes from the object to the data session it was created in. As long as the object exists so must the data session. Your sessionbase object is sort of a red herring in the discussion, it creates the DS but it's only one of two objects with attachments to the DS. The sRecordNNN object is what's forcing the DS to remain open, it contains a reference to the data session. Once sRecordnnn is released your data session can release.

This behaviour is not a bug or indication of a memory leak.