This Message was undeliverable due to the following reason:
A message was not delivered because a loop was found in the Mail
eXchanger record database (the MX loop condition). The destination
host has an MX record that points to this host, but there is no
account for the designated recipient.
Here are a few possible solutions for clearing up this problem:
- add the domain name to your list of Local Mail Domains
if this mail server has complete authority over all
mail accounts within that domain
- return the mail since the address was invalid
- fix the MX records in the DNS for the host in question
and then resubmit the message
- add an account or alias for the recipient on this machine
and then resubmit the message
If this machine should be handling all mail for the destination
host, it should be added to the list of Local-Mail-Domains in
the "system" configuration.
Destination host: geekpowered.com
The following recipients did not receive this message:
<david@geekpowered.com>
Please reply to Postmaster@mail.mcsweb1.com
if you feel this message to be in error.
--===========================_ _= 2066266(2109)
Content-Type: message/rfc822
Received: from walnut.iea-software.com ([207.53.165.8]) by mail.mcsweb1.com
(Post.Office MTA v3.1.2 release (PO203-101c)
ID# 0-44404U2500L250S0) with ESMTP id AAA9467
for <david@geekpowered.com> Wed, 14 Oct 1998 22:30:37 -0700
Received: from walnut.iea-software.com (walnut.iea-software.com [207.53.165.8]) by walnut.iea-software.com (NTMail 4.00.0020/NT6651.00.c89adb95) with ESMTP id za002321 for <radiusNT@iea-software.com> Wed, 14 Oct 1998 22:43:09 -0700
Received: from [207.53.173.19] by walnut.iea-software.com (NTMail 4.00.0020/NT6651.00.c89adb95) with ESMTP id hvkbaaaa for <radiusNT@iea-software.com> Wed, 14 Oct 1998 22:43:05 -0700
Message-ID: <36258E7D.C24F1E0A@iea-software.com>
Date: Wed, 14 Oct 1998 22:56:13 -0700
From: "Dale E. Reed Jr." <daler@iea-software.com>
Organization: IEA Software, Inc.
X-Mailer: Mozilla 4.06 [en] (WinNT; I)
MIME-Version: 1.0
To: sberlin@csonline.net
CC: radiusNT@iea-software.com
Subject: Re: RadiusNT slow to Autenticate with Sequel
References: <2.2.32.19981013220413.01367228@mail.csonline.net>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Script: \\FilterScript.mml
X-DomainScript: csonline.net\\script.mml
X-UserScript: iea-software.com\radiusNT\script.mml
X-Info: -Reply-To: radiusnt@iea-software.com
sberlin@csonline.net wrote:
>
> RadiusNT 2.2 Running as a service
> Two Radius Servers one Sequel Server
>
> Radius servers both Pentium 166, 128mb ram, SCSI HD
> Sequel server Dual Pentium 200, 196mb ram, 4 gig SCSI HD, Adaptec 2940
> NAS Ascend Max 4000's (14) Have tried 5.x and 6.x code
> Sequel server 6.5 300mb for database, 30mb temp db
>
> Running only authentication (no accounting) Radius will run and verify ok
> but radius response times show to range from 1600ms to 4000ms. No other
> debug errors.
Realize two things. First, running in foreground debug mode will slow
RadiusNT down by a factor of 10 soley because of screen refresh issues.
If you minimize the window, make a request, and restore it, you'll see
what I am talking about (versus leaving it open). Secondly, the
performance
of RadiusNT is directly related to the configuration of your SQL Server
and the design of the database. For example, indexes, proper
maintenance,
and proper configuration can make a database perform countless times
faster
than a system that isn't properly setup. It only gets worse as your DB
grows if this isn't setup right.
The SQL Server scripts included with RadiusNT are not fully optimized.
They are designed as a base to start from to create your own database
from. For example, the login field is indexed, but I don't see one
for the shell and Email fields, which are equally as important for
authentication (the email is an addition in 2.5, whereas that script
was originally written for 2.2).
> Then at some random interval (a day or a week) the Sequel server processor
> usage will skyrocket and authentications will fail. Debug shows no errors
> but it seems as if the NAS is sending many repeated requests that can't be
> processed. User names will be repeated from one or more Max's. Opening the
> Max profile, leaving the authentication servers set, will result in the
> Sequel server dropping back to normal after a period (20-30min). The Max's
> can be the reset to use the radius profile again and may run for another few
> days or weeks.
This could be some issues with the calls table. RadiusNT 2.2 used the
same
DSN for auth and accounting. By delting records from the calls table (a
part
of the maintenance I mentioned above) the calls table will be locked.
In
2.2, that stopped everything, including authentication (it would timeout
after 30 seconds, but would basically be dead). RadiusNT 2.5 solved
this
problem by using two DSNs and two threads (1 each for Auth and
Accounting).
This easilt doubles the performance of RadiusNT alone. I would highly
recommend looking at RadiusNT 2.5.
> There seems to be no ryme or reason why this happens it will happen during
> high and low auth periods. This only thing that seems to show is the
> 1800-4000ms times for radius which I thought would be a lot quicker.
Typical RadiusNT auth times should be ~300ms max. Anything more needs
to
be investigated.
> Any Ideas why this might happen, where to look or what can be tweaked with
> Sequel for a preformace boost. We have been trying to work this out for a
> couple of months.
Two things I would do is run a checkdb to make sure the DB is in order
and also update your statistics/indexes. The laster is vital when you
change a lot of data in your database, since the index may be out of
date and therefore causing SQL Server to do a full table scan (which
means slow).
-- Dale E. Reed Jr. (daler@iea-software.com)_________________________________________________________________ IEA Software, Inc. | RadiusNT, Emerald, and NT FAQs Internet Solutions for Today | http://www.iea-software.com--===========================_ _= 2066266(2109)--