Date: 29 Sep 2001 03:12:09 +0000 From: "Mark" <markd@BushWire.Net> To: "Nguyen-Tuong Long Le" <le@cs.unc.edu> Cc: freebsd-net@freebsd.org Subject: Re: FD_SETSIZE Message-ID: <20010929031209.18095.qmail@prefix.bushwire.net> In-Reply-To: <Pine.LNX.4.32.0109282129490.1971-100000@le-cs.cs.unc.edu>; from le@cs.unc.edu on Fri, Sep 28, 2001 at 09:30:33PM -0400 References: <Pine.LNX.4.32.0109282129490.1971-100000@le-cs.cs.unc.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Sep 28, 2001 at 09:30:33PM -0400, Nguyen-Tuong Long Le allegedly wrote: > Hi, > > I am wondering what is the side effects of increasing FD_SETSIZE > beyond 1024? Our group have a propiertary web server software that > handles a large number of sockets. While increasing the kern.maxfiles > and kern.maxfilesperproc gives our web server more connections, > select() seems to fail if the descriptor is beyond 1024. > > Can I just increase FD_SETSIZE and recompile the kernel? I saw > some magic numbers in kern/sys_generic.c and am not sure whether > there are some side effects. I know this doesn't answer your question directly, but do you have to use select()? If not, you might want to consider poll() or better yet, kqueue(). Neither suffer from a predefined limit such as FD_SETSIZE and both should scale and perform as well as, or better than, select(). Regards. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20010929031209.18095.qmail>