Quote - Thanks - yeah, I guess I'd expect to see this sort of thing if either the server was reassigning the session ID extremely early (which seems unlikely) or if there was some sort of truncation of IDs occurring... as in the session id, which is quite possibly a long integer, somehow being in a range outwith the scope of a code statement somewhere that is truncating the long integer to a shorter data type (probably a string type)...

Some of the other issues I've seen would also make me suspect some kind of truncation of long data type numerical ids (e.g.) when they're parsed into a string data type...

...and I guess that this kind of problem would be consistent with an issue occuring post data migration, that somehow wasn't captured in test; e.g. if the initialization of an index key somewhere started out at a much higher value, post migration, than it ever got to in the test phase.

But that's all pure speculation on my part of course... sorry! Getting carried away there... LOL.

Hopefully we'll see the horse brought back into the paddock shortly ;-)

Well, I thought it was a recursive gribble in the squonkulator that was the problem.  But on reflection yours may be a bit more testicle.  I mean technical. 

Coppula eam se non posit acceptera jocularum.

My Store

My Gallery