httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dean Gaudet <>
Subject Re: roy's l_c perf patch and spareservers
Date Sat, 15 Feb 1997 19:47:22 GMT
When I did those SO_LINGER experiments last week I sent a message to Steve
Alexander <>, and he said that w.r.t.  SO_LINGER,
IRIX is probably "bug-for-bug compatible with 4.3BSD as originally
shipped."  He also said that SO_LINGER didn't have a timer in all releases
prior to 4.3-Reno.  I'm not sure of the history of net1 or net2, but this
probably gives you a better idea.


On Sat, 15 Feb 1997, Jim Jagielski wrote:

> Roy T. Fielding wrote:
> > 
> > >I put Roy's l_c perf patch on HotWired's servers for a few hours... but
> > >the FIN_WAIT_2s flew through the roof, so I had to go back to
> > >-DNO_LINGCLOSE.  (IRIX 5.3).  There was some speculation that this might
> > >help the FIN_WAIT_2 situation, but it doesn't look like it.  As far as
> > >performance goes I didn't really get a chance to compare, I'll try again.
> > 
> > Thanks, that narrows it a bit.  I think what may be occurring is that
> > the shutdown(sd, 1) is changing some flag on IRIX's (and apparently other)
> > TCP stacks such that the later close() is not working, or we just block
> > on shutdown(sd, 1) because the OS is too stupid to realize that it is
> > supposed to be a non-blocking call.  Can you modify that test program
> > you posted earlier and truss it to see where it gets lost?  I have run
> > out of ideas here, since it works fine on our Solaris machines.  Hmmm,
> > I might be able to test it on SunOS4 tomorrow.
> Hmmmm It would be interesting to know if IRIX 5.3 is based on
> Net1 or Net2 BSD code... A/UX also has severe problems with l_c()
> and it's sort of a combo of both (Net1.5 :) ).
> > 
> > Is there any portable way to abort a connection?
> > 
> Well, using SO_LINGER with a linger-time of 0 should send an
> immediate RST...
> -- 
> ====================================================================
>       Jim Jagielski            |       jaguNET Access Services
>           |
>                   "Not the Craw... the CRAW!"

View raw message