Here's an old post I forgot to cc the list to. Thanks to James for
forwarding it.
----------Forwarded message ----------
From: [EMAIL PROTECTED]
To: 5@ @@ay4it.com
Date: 2 Aug 2000 14:44:14 -0000
Subject: failure notice
>
> Hi. This is the qmail-send program at muncher.math.uic.edu.
> I'm afraid I wasn't able to deliver your message to the following addresses.
> This is a permanent error; I've given up. Sorry it didn't work out.
>
> <[EMAIL PROTECTED]>:
> I do not accept blind carbon copies. Please use To or Cc.
>
> --- Below this line is a copy of the message.
>
> Return-Path: <[EMAIL PROTECTED]>
> Received: (qmail 9238 invoked from network); 2 Aug 2000 14:44:13 -0000
> Received: from maybe.domainregistry.ie (193.1.142.4)
> by muncher.math.uic.edu with SMTP; 2 Aug 2000 14:44:13 -0000
> Received: (qmail 8695 invoked by uid 1000); 2 Aug 2000 14:43:49 -0000
> Resent-Message-ID: <[EMAIL PROTECTED]>
> Delivered-To: [EMAIL PROTECTED]
> Received: (qmail 8382 invoked from network); 2 Aug 2000 14:13:26 -0000
> Received: from drno-int.domainregistry.ie (HELO smtp.domainregistry.ie)
>(?hIO52HHLEyWUw4SKuCTAep5ydEYRa8/[email protected])
> by maybe.domainregistry.ie with SMTP; 2 Aug 2000 14:13:26 -0000
> Received: (qmail 31881 invoked by uid 501); 2 Aug 2000 14:13:25 -0000
> Delivered-To: [EMAIL PROTECTED]
> Received: (qmail 31878 invoked from network); 2 Aug 2000 14:13:25 -0000
> Received: from unknown (HELO fwder.com) (216.42.24.88)
> by smtp0.domainregistry.ie with SMTP; 2 Aug 2000 14:13:25 -0000
> Received: (qmail 9113 invoked by uid 99); 2 Aug 2000 14:05:57 -0000
> Message-ID: <[EMAIL PROTECTED]>
> References: <[EMAIL PROTECTED]>
> <[EMAIL PROTECTED]>
> <[EMAIL PROTECTED]>
> In-Reply-To: <[EMAIL PROTECTED]>
> From: "JuanE" <[EMAIL PROTECTED]>
> To: James Raftery <[EMAIL PROTECTED]>
> Subject: Re: updated load balancing qmail-qmqpc.c mods
> Date: Wed, 02 Aug 2000 14:05:57 GMT
> Mime-Version: 1.0
> Content-Type: text/plain; charset="us-ascii"
> Content-Transfer-Encoding: 7bit
> Resent-From: [EMAIL PROTECTED]
> Resent-Date: Wed, 2 Aug 2000 15:43:49 +0100
> Resent-To: [EMAIL PROTECTED]
>
>
> Not quite. The possibility does exist, but the likelihood os not the same
> as choosing an even spread. If only one server is down, you will pick it
> twice with probability 1/N^2, 3 times with p = 1/N^3, etc...
>
> The probability that you deliver the message in 2 or less attempts is
> (N-1)/N + (N-1)/N^2. So the odds of distributing the load evenly is much
> greater than acuatlly polling the same down server twice.
>
> But, you bring up a valid point. If you want to absolutely avoid the
> possibility of polling the same down server twice, you need to keep track
> of the servers you poll. When you sample a random number, just check if it
> already came up and skip it. This can easily be accomplished with flags. If
> you have less than 32 servers (or is it 16, i don't remember), you can just
> use an integer (or long?) and use every bit as a flag servers you polled.
> If the bit for the server you are going to poll is on, it means you already
> polled it, so you should skip it and get a random sample again.
>
> JES
>
> James Raftery writes:
>
> > On Wed, Aug 02, 2000 at 11:13:22AM +0000, JuanE wrote:
> > > This way, you will redistribute the load randomly to all
> > > servers and not just the next one on the list.
> >
> > Not quite. If choose truly randomly the liklehood of you choosing the
> > same server /every/ time is the same as the liklihood of choosing a
> > nice even spread across the servers. This isn't what you want.
> >
> > Regards,
> >
> > james
> > --
> > James Raftery (JBR54) - Programmer Hostmaster - IE TLD Hostmaster
> > IE Domain Registry - www.domainregistry.ie - (+353 1) 706 2375
> > "Managing 4000 customer domains with BIND has been a lot like
> > herding cats." - Mike Batchelor, on [EMAIL PROTECTED]
> >
>
>
>
