Advice needed on running idiotic test for client

[Mike Silbersack]

On Tue, 14 Jun 2005, PSI, Mike Smith wrote:

> Idiot test requested by client - Same as above but see how fast 10,000 
> can be sent (will never come close to happening in the real world).
>
> Problem: Because among other things I am using let's say minimal and 
> ancient computer power, I hit a DOS state after about 3700 connects. I 
> know (think?) I am running out of mbufs because of the 2MSL delay on the 
> close.
>
> Is there any way to close a TCP connection and have all resources 
> immediately freed? I KNOW this is really violating all the data 
> integrity features of TCP, but this is an idiotic test remember. And in 
> this case I am not concerned about data integrety. I have noticed that 
> LINUX has a TCP_LINGER2 option that can at least minimize the closing 
> delay (freeing resources) but I haven't found anything comparable in 
> FreeBSD.

Are the TIME_WAIT sockets building up on the client or the server?

5.x has some features so that it does not allow too many TIME_WAIT sockets 
to build up beyond a certain threshold, but if you're using 4.x we can 
still tweak some sysctl values to achieve the effect you want.

Mike "Silby" Silbersack