michael@??? said:
> did anybody ever try multiple exim processes that accept() connections
> from the same socket? It would reduce the latency on heavily loaded
> systems, but I have no clue if that would make any difference for mail
> systems.
I don't think this optimisation would work well for an MTA which will
tend to have longer session than an httpd daemon (which is where that
optimisation comes from).
Additionally you would then have the problem of losing the current
capability for session limiting and prevention of a single host
monopolising the SMTP slots unless you invent a new scheme for
inter-process communications and stats building... which could be more
of a bottleneck than the current single accepter.
ph10@??? said:
> AFAIK, this is not possible. How would the operating system's TCP/IP
> stack know which process to hand an incoming call to?
Its done by contention - first process to wake and take it gets the
session - in general on most Unices all the processes get woken, but
only one gets something useful returned from accept.
Nigel.
--
[ - Opinions expressed are personal and may not be shared by VData - ]
[ Nigel Metheringham Nigel.Metheringham@??? ]
[ Phone: +44 1423 850000 Fax +44 1423 858866 ]
