[Spread-users] INET unable to bind to port - force kill?

Marc ZYNGIER mzyngier at freesurf.fr
Wed Aug 1 11:27:55 EDT 2001


>>>>> "JS" == Jonathan Stanton <jonathan at cnds.jhu.edu> writes:

JS> Ok. I think this is right, but I don't have a copy of Stevens or
JS> other documentation handy. The reason I thought it was not well
JS> defined is that I vaugely remember that some implementations (from
JS> different OS's) did actually do slightly different things here,
JS> even thought what you describe makes the most sense and might be
JS> the 'standard' behaivor (as well as the socket interface is
JS> standardized).

At least Solaris, AIX and Linux are behaving this way... I could also
check on NetBSD if you want. Any *sane* OS should have a similar
behaviour (and Windows should be sane on this subject, if my memory
serves me well).

JS> Yes. Spread could examine the interfaces when it starts and bind to all of
JS> them which would be similar to INADDR_ANY, I think the main difference
JS> would be with interfaces which were not up when Spread starts, but come up
JS> afterwards. With INADDR_ANY Spread will also receive traffic on these new
JS> interfaces ( I think ) with specific binds it will receive only the traffic
JS> for the original interfaces.

Have a look at what BIND (named) does... It binds on each available
interface, and can detect interfaces addition/deletion.

JS> Sure, that makes sense and should work fine on the client-server side. Do
JS> you also bind the daemon-to-daemon ports to specific interfaces ( I assume
JS> yes ) and noone (or just one) binds the localhost address?

That's exactly my setup. Each instance of the daemon have all his ports
bound to one interface, and noone uses localhost.

Regards,

        M.
-- 
Places change, faces change. Life is so very strange.





More information about the Spread-users mailing list