From owner-freebsd-isp Fri Nov 6 15:32:52 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id PAA21536 for freebsd-isp-outgoing; Fri, 6 Nov 1998 15:32:52 -0800 (PST) (envelope-from owner-freebsd-isp@FreeBSD.ORG) Received: from unix.kawartha.com (unix.kawartha.com [204.101.15.2]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id PAA21531 for ; Fri, 6 Nov 1998 15:32:51 -0800 (PST) (envelope-from pstewart@kawartha.com) Received: from shell.kawartha.com (shell.kawartha.com [204.101.15.43]) by unix.kawartha.com (8.8.8/8.8.7) with SMTP id SAA22918; Fri, 6 Nov 1998 18:38:00 -0500 (EST) Date: Fri, 6 Nov 1998 18:56:08 -0500 (EST) From: Paul Stewart To: Deepwell Internet cc: freebsd-isp@FreeBSD.ORG Subject: Re: Consolidate or isolate? In-Reply-To: <4.1.19981106095213.00ab5100@mail1.dcomm.net> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-isp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Just my two cents worth... we used to run ALL services off of two servers when we first started out in the ISP industry....now we have 26 CORE servers running critical services... some of the decision depends on your size.... The thing we really like about running so many servers is that if one server happens to go down that doesn't have fault tolerance, then big deal.... our dial-up users can still surf, our websites are still online or whatever the case may be.... If you move to a few BIG servers with RAID5, dual fault tolerant power supplies etc.... that's great.. but what happens if you take a box down my mistake (software glitch whatever)... it's happened the best sys admin's out there.... oops! Time to reboot... while you're rebooting your whole operation if offline or a good majority of it... Personally, I just did this the other day... I was downing an IP address (virtual web account) as the customer now wants Front Page extensions which is on a seperate server... instead of downing the IP itself, I took down the entire ethernet for that server.... dumb mistake but it happens... however instead of taking down ALL our web customers at that moment, I only took down about 200.... :) If we had all web customers on one server, then more people would have been yelling.... thankfully either way, the server was only down for about 30-45 seconds.. but even so... Just my two cents worth... Paul On Fri, 6 Nov 1998, Deepwell Internet wrote: > We run FreeBSD almost exclusively here and we have divided into two camps > of thought. We're running all the standard services that every other ISP > runs; DNS, SMTP/pop3, webserver, SSL webserver, shell server, radius and > probably a few others that aren't popping into my head. > > Basically, we are divided here. Some people think that we should maintain > a machine for DNS, a separate Radius server, a separate mail server...yadda > yadda yadda. The other group says we should consolidate and throw > DNS/Radius on one machine and work towards consolidating everything towards > a few core machines. Obviously we'd have to watch the loads and not throw > our large mail server and a large webserver together. And allowing shell > accounts on the primary webserver or radius server seems a little risky. > > Can you all give me your input. > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-isp" in the body of the message > > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-isp" in the body of the message