From owner-freebsd-hackers Wed Apr 7 18:19: 9 1999 Delivered-To: freebsd-hackers@freebsd.org Received: from moebius2.Space.Net (moebius2.Space.Net [195.30.1.100]) by hub.freebsd.org (Postfix) with SMTP id 2F1F214CE3 for ; Wed, 7 Apr 1999 18:19:00 -0700 (PDT) (envelope-from maex@Space.Net) Received: (qmail 2420 invoked by uid 1013); 8 Apr 1999 01:17:19 -0000 Message-ID: <19990408031719.M28680@space.net> Date: Thu, 8 Apr 1999 03:17:19 +0200 From: Markus Stumpf To: hackers@FreeBSD.org Subject: squid and FreeBSD (was: Re: Nice FreeBSD mention) References: <19990407151931.A36308@ikhala.tcimet.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.93.2i In-Reply-To: <19990407151931.A36308@ikhala.tcimet.net>; from Natty Rebel on Wed, Apr 07, 1999 at 03:19:31PM -0400 Organization: SpaceNet GmbH, Muenchen, Germany Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On Wed, Apr 07, 1999 at 03:19:31PM -0400, Natty Rebel wrote: > > In a different section, they mention some tuning that they had to do to > > get FreeBSD to hop a long a bit faster... I don't have the exact URL > > handy, but it's in the software section... Perhaps the tips they suggest > > could be incorporated into a FAQ or note somewhere... > here's the url > http://polygraph.ircache.net/caveats.html The IMHO more speed limiting factor I experience is with swapping. I have asked this list some time ago, but none of the answers (thanks again anyway ;-) did help. I have a FreeBSD 2.2.5 box with 256 MB RAM. It's dedicated to squid (1.NOVM.22), no other services running. "top" says PID USERNAME PRI NICE SIZE RES STATE TIME WCPU CPU COMMAND 8542 root 2 -15 204M 177M select 60:27 0.00% 0.00% squid From that point I would imagine that 256 MB RAM is enough to fit in squid and - with about 50 megs left - some other processes. However, due to the precautionary swapping algorithm in FreeBSD squid gets swapped out and after running some time. squid then reports a ratio of page faults to http requests > 1 and the system is constantly swapping. This can also be monitored with "systat -vmstat". My current solution is to restart the squid every 2-3 weeks or so. This puts the above ratio down to 0.5 and from that it increases within this period to 1.0 again. I also thought of completely removing swap space on the machine. Do you guys think it would help? Or would it crash squid/the system? (I am rather careful with try and error in this case as not to interupt service for our customers, who rely on the cache a lot). Any comments/suggestions/solutions welcome! Other than that we are more than happy with FreeBSD: $ uptime 3:05AM up 251 days, 3:03, 3 users, load averages: 0.00, 0.03, 0.09 :-)))) \Maex -- SpaceNet GmbH | http://www.Space.Net/ | In a world without Research & Development | mailto:maex-sig@Space.Net | walls and fences, Joseph-Dollinger-Bogen 14 | Tel: +49 (89) 32356-0 | who needs D-80807 Muenchen | Fax: +49 (89) 32356-299 | Windows and Gates? To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message