Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 6 Dec 2000 00:42:29 -0800
From:      Boris <koester@x-itec.de>
To:        Brian McGroarty <snowfox@yipyap.net>
Cc:        Gabriel Ambuehl <gabriel_ambuehl@buz.ch>, Alfred Perlstein <bright@wintelcom.net>, <questions@FreeBSD.ORG>
Subject:   Re[2]: Load-Balancing - any solutions?
Message-ID:  <1833232197.20001206004229@x-itec.de>
In-Reply-To: <20001205110841.A14207@yipyap.net>
References:  <105102226954.20001205163641@x-itec.de> <20001205074217.Y8051@fw.wintelcom.net> <5676150678.20001205165002@buz.ch> <20001205110841.A14207@yipyap.net>

next in thread | previous in thread | raw e-mail | index | archive | help
Hello Brian,

Tuesday, December 05, 2000, 9:08:41 AM, you wrote:

BM> With just two systems, running squid on one machine to access the
BM> second is a decent solution.  This has the advantage of being
BM> completely self-updating and trivial to maintain.

VERY interesting. Do i have to setup squid to listen on port 80
instead of 3128 for this? Or are they any other configuration tips
somewhere?  I think i have to change the port to 80, and the request
to a specific domain is resolved (internally) by an internal dns server combined
with round robin? At this point we have load sharing with cached http
pages (very userful with static pages), but even if this works we still not have
real load balancing.

Lets assume i put a squid server in front of the real webserver.
Dynamic webpages will still be processed over the real server btw the
cache must be always refreshed. As seen at many providers, this will
provide by high load servers some timing problems, the result is that
you request a page and it takes 2-5 secounds before you can see
anything. The overloaded squid server requests the overloaded master
server, the cache needs to be rebuid and send to the client.

By this way we have two machines working and the cpu´s have much to
do. I want one machine per request of a client, not two even if
running squid is a VERY interesting idea.

I need this solution (only for interesting purposes).

machine A (apache server, "master-server")
machine B (apache server, "replicated directories")

I would use dns round robin for load sharing if there is really no
good load balancing solution (free) available (this is ok), but there is still the
problem to replicate the content from A to B if anything on A changes
somewhere.

Running a cronjob every minute kills the harddrive someadays.

I  think about the coda-fs, this would be interesting, but i have
never get it up and running -(


-- 
Best regards,
 Boris                            mailto:koester@x-itec.de




To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1833232197.20001206004229>