> From: Dave Shield [mailto:D.T.Shield@...]
> Sent: Friday, August 24, 2001 01:42 AM
> One possibility is that the "next object" is somewhat expensive to
> retrieve, and the client is timing out and asking for it again, (and
> again and again) while the agent is still working on the
> first request.
> But even when it finally answers the first request, it then has to
> answer the retries as well - each with the same delay. It's not
> unknown for this to completely tie up the agent for a few minutes.
Hmmm? Shouldn't requests which match all fields (srcHost, srcPort,
msgId) of a request in process be dropped?