Sockets stuck in FIN_WAIT_1

[Robert Blayzor]

On May 30, 2008, at 4:47 AM, David Malone wrote:
> There has been some talk about this sort of problem on the IETF TCP
> Maintainers list. I don't think any good conclusion was reached -
> whatever the solution was certainly needs to be tunable per-socket
> because this behaviour is perfectly valid in some situations but a
> bit of a pain in others.



A timeout value would be fine.  Obviously if the client keeps sending  
back packets with a 0 size, there should be some option or work around  
to tell the stack to drop the connection.  There than to have the  
server lock up resources on a "dead connection".  Unfortunately we're  
talking about the internet here, we can't insure that every one of the  
clients connecting to our servers behaves correctly! ;-)

On a side note, I could easily fix this problem by frontending the  
server with a Cisco PIX or ASA.  I believe they have "half closed"  
timers just for this purpose... Perhaps a kernel tunable knob would be  
a nice option/fix/hack also.

-- 
Robert Blayzor, BOFH
INOC, LLC
rblayzor at inoc.net
http://www.inoc.net/~rblayzor/