From: Jim Choate <ravage@ssz.com>
Date: Fri, 24 Oct 1997 11:46:21 +0800
To: cypherpunks@ssz.com (Cypherpunks Distributed Remailer)
Subject: Re: ADMIN: lost messages (fwd)
Message-ID: <199710240359.WAA21656@einstein.ssz.com>
MIME-Version: 1.0
Content-Type: text



Forwarded message:

> From: Jeff Barber <jeffb@issl.atl.hp.com>
> Subject: Re: ADMIN: lost messages (fwd)
> Date: Thu, 23 Oct 1997 13:21:27 -0400 (EDT)

> I understand this rationale.

It isn't a rationale, its reality. This view is not moderated by political
or ideological beliefs but the harsh reality that things don't work the
right way all the time.

> But that doesn't explain what's 
> happening to the missing messages.

Your implied base assumption is that the reason is always the same. I
seriously doubt it.

>  Is there a flaw in the 
> distributor/reflector duplicate-filtering algorithms? 

The process is so simple it ain't funny. I believe, but won't guarantee it,
that we are all using a procmail script to remove duplicate messages, and
that is all the 'distributor/reflector duplicate-filter algorithms' there
is. Unless one of the other repeaters is doing something more complicated
they haven't said. In short, you are under the misconception that this is
something special, it ain't.

SSZ receives traffic from the other remailers at cpunks@ssz.com, just a
normal user account, which has the procmail script filtering duplicates. It
then forwards to cypherpunks@ssz.com, which is the majordomo managed list.
cpunks@ssz.com as well as the rest of the subscribers get their traffic from
here. If the procmail breaks then I get a mail loop at SSZ and my box fills
up and lets me know pretty quickly something is broke because majordomo
recognizes this through the X-loop: and kills them in that case as well as
sending postmaster (me) all kinds of nifty hints. The other sites in this
case still filter duplicates so things still don't run away.

>From my end to add another node is to put it in the procmail script for
outbound and have them add cpunks to their outbound list.

> Or is this
> a problem of certain messages being introduced into a node whose
> product does not get propagated to all subscribers of all the other
> nodes?

Sounds good to me...<shrug>

> Or are you saying that if ssz (say) goes down for a few
> minutes it forgets about any mail it has queued up for its
> counterparts at cyberpass and algebra (et al.)?

No, SSZ has full inbound and outbound traffic queing for up to 4 days
or the hard drive fills up here and at my ISP.


    ____________________________________________________________________
   |                                                                    |
   |    The financial policy of the welfare state requires that there   |
   |    be no way for the owners of wealth to protect themselves.       |
   |                                                                    |
   |                                       -Alan Greenspan-             |
   |                                                                    | 
   |            _____                             The Armadillo Group   |
   |         ,::////;::-.                           Austin, Tx. USA     |
   |        /:'///// ``::>/|/                     http://www.ssz.com/   |
   |      .',  ||||    `/( e\                                           |
   |  -====~~mm-'`-```-mm --'-                         Jim Choate       |
   |                                                 ravage@ssz.com     |
   |                                                  512-451-7087      |
   |____________________________________________________________________|


ps The ISDN was apparently moved to a new pair. It seems rock solid at
   this point...time will tell.