qmail Digest 30 May 1999 10:00:01 -0000 Issue 656
Topics (messages 26109 through 26121):
Fwd: Warning: message 10nNFt-0004ji-00 delayed 24 hours
26109 by: "Fred Lindberg" <[EMAIL PROTECTED]>
26114 by: "Sam" <[EMAIL PROTECTED]>
26116 by: "Fred Lindberg" <[EMAIL PROTECTED]>
26119 by: Jason Haar <[EMAIL PROTECTED]>
26120 by: "Sam" <[EMAIL PROTECTED]>
Mailing lists on dial-up box
26110 by: Eric Dahnke <[EMAIL PROTECTED]>
26111 by: "Adam D. McKenna" <[EMAIL PROTECTED]>
26115 by: RaTao von J <[EMAIL PROTECTED]>
Q: getting a virtual domain
26112 by: [EMAIL PROTECTED]
~/Mailbox and "you have mail"
26113 by: Brian Butler <[EMAIL PROTECTED]>
receiving bounces...
26117 by: Cris Daniluk <[EMAIL PROTECTED]>
26118 by: Dustin Marquess <[EMAIL PROTECTED]>
tcp-wrappers, qmail-1.03 and Solaris 2.7
26121 by: Mark Recio <[EMAIL PROTECTED]>
Administrivia:
To subscribe to the digest, e-mail:
[EMAIL PROTECTED]
To unsubscribe from the digest, e-mail:
[EMAIL PROTECTED]
To bug my human owner, e-mail:
[EMAIL PROTECTED]
To post to the list, e-mail:
[EMAIL PROTECTED]
----------------------------------------------------------------------
Here's another reason against DSN. Got this for posting to the qmail
list. The idea of generating more than one message per message without
extreme care strikes me as a recipe for disaster. VERP doesn't.
==================BEGIN FORWARDED MESSAGE==================
>Return-Path: <>
>Delivered-To: [EMAIL PROTECTED]
>Received: (qmail 273 invoked from network); 29 May 1999 14:11:41 -0000
>Received: from serv1.is1.u-net.net (195.102.240.129)
> by id.wustl.edu with SMTP; 29 May 1999 14:11:41 -0000
>Received: from exim by serv1.is1.u-net.net with local (Exim 2.00 #2)
> for [EMAIL PROTECTED]
> id 10njlS-0005OA-00; Sat, 29 May 1999 14:07:06 +0000
>From: Mail Delivery System <[EMAIL PROTECTED]>
>To: [EMAIL PROTECTED]
>Subject: Warning: message 10nNFt-0004ji-00 delayed 24 hours
>Message-Id: <[EMAIL PROTECTED]>
>Date: Sat, 29 May 1999 14:07:06 +0000
>
This message was created automatically by mail delivery software.
A message that you sent has not yet been delivered to all its
recipients
after more than 24 hours on the queue on serv1.is1.u-net.net.
The message identifier is: 10nNFt-0004ji-00
The date of the message is: Fri, 21 May 1999 08:57:51 -0500
The subject of the message is: Re: Virtual domains and ezmlm-idx
The address to which the message has not yet been delivered is:
[address of a qmail list subscriber]
No action is required on your part. Delivery attempts will continue for
some time, and this warning may be repeated at intervals if the message
remains undelivered. Eventually the mail delivery software will give
up,
and when that happens, the message will be returned to you.
===================END FORWARDED MESSAGE===================
-Sincerely, Fred
(Frederik Lindberg, Infectious Diseases, WashU, St. Louis, MO, USA)
On Sat, 29 May 1999, Fred Lindberg wrote:
> Here's another reason against DSN. Got this for posting to the qmail
> list. The idea of generating more than one message per message without
> extreme care strikes me as a recipe for disaster. VERP doesn't.
[ snip ]
This is not an RFC-compliant DSN, unless you did not include all the
headers and attachments. Looks like a home-brewed, broken, mailer-daemon
message. Even if it were a standard DSN, it is a site configuration
issue, since delayed DSNs are not required.
In fact, I can guarantee you that this is not an RFC-compliant DSN,
because RFC-compliant DSNs go to the envelope sender address, which would
be ezmlm in this case, and not the header address. You will never see
RFC-compliant DSNs from mailing list traffic.
That does bring up an interesting point. An RFC-compliant DSN will go to
the envelope sender address, and I'm not sure if ezmlm is smart enough to
identify delayed DSNs. If it doesn't, it'll look like a bounce to it, and
eventually it will kick you off the list. I'm sure that many will
disagree, but in this case I believe that this would be an ezmlm bug.
On Sat, 29 May 1999 13:37:29 -0400 (EDT), Sam wrote:
The point I was trying to make is that when MTA are misconfigured, the
more traffic generation points (bounce, delay notification, etc) there
are, then more garbage will be sent and the more difficult it becomes
to use E-mail for communication.
>This is not an RFC-compliant DSN, unless you did not include all the
No it's not. Format wrong, and sent to From: address, not envelope
sender. If one does DSN, of course one should do it right.
>That does bring up an interesting point. An RFC-compliant DSN will go to
>the envelope sender address, and I'm not sure if ezmlm is smart enough to
>identify delayed DSNs. If it doesn't, it'll look like a bounce to it, and
>eventually it will kick you off the list. I'm sure that many will
>disagree, but in this case I believe that this would be an ezmlm bug.
ezmlm-weed filters out DSN. Of course, who can keep up with all the
home brew by major software vendors ;-)
Speaking of ezmlm: I used to assume that on a ezmlm list, one would see
less that 5% bounces. I have now looked at a few large lists, and the
number of bouncing addresses at any time on a busy list is about 10%.
ezmlm removes the legitimate problem addresses. remaining are bounces
for "over quota", "we do not relay" (we _are_ an MX for the host, but
our sysadmin installed the new sendmail without knowing what /she was
doing), "sender domain must resolve" (people do a lot of DNS lookups
for spam checks, but aren't well connected to DNS - some even make this
a permanent error), looping message (>15 received headers), etc, and my
all time favorite "syntax error in Reply-to: header" (yes, they check
each header and seem to do one DNS lookup per address in the message).
Thus, the majority of "bounces" are intermittent failures due to
configuration errors and spam control, and of course DSN adds to this.
Conclusion: In an attempt to make it better and safer it has been made
too complicated for at least 10% of sysadmins. Any new "feature", such
as DSN should be viewed against this.
-Sincerely, Fred
(Frederik Lindberg, Infectious Diseases, WashU, St. Louis, MO, USA)
On Sat, May 29, 1999 at 01:37:29PM -0400, Sam wrote:
> That does bring up an interesting point. An RFC-compliant DSN will go to
> the envelope sender address, and I'm not sure if ezmlm is smart enough to
> identify delayed DSNs. If it doesn't, it'll look like a bounce to it, and
> eventually it will kick you off the list. I'm sure that many will
> disagree, but in this case I believe that this would be an ezmlm bug.
Spot on. Some months ago my workstation was down for 8 hours - and ezmlm
managing this list logged the generated "warning: cannot deliver for 4
hours" messages as bounces - it bumped me into djb's "possibly badaddress
list" and some weeks later I got a monitoring message from ezmlm checking to
see that I still existed (an Email-ping :-). I was surprised that ezmlm
considered a 4-hour warning to be a bounce... I mentioned it to Dan and he
didn't think it was a problem. As it doesn't actually drop you from the list
after a warning-bounce, it's probably not a bug - but a feature ;-)
--
Cheers
Jason Haar
Unix/Network Specialist, Trimble NZ
Phone: +64 3 3391 377 Fax: +64 3 3391 417
On Sun, 30 May 1999, Jason Haar wrote:
> Spot on. Some months ago my workstation was down for 8 hours - and ezmlm
> managing this list logged the generated "warning: cannot deliver for 4
> hours" messages as bounces - it bumped me into djb's "possibly badaddress
> list" and some weeks later I got a monitoring message from ezmlm checking to
> see that I still existed (an Email-ping :-). I was surprised that ezmlm
> considered a 4-hour warning to be a bounce...
ezmlm probably thinks that anything that's sent to the return address is a
bounce. Since DJB doesn't like DSNs, I doubt that ezmlm's code includes
logic to parse bounces, looking for delayed DSNs.
I don't think any of you are understanding his problem. The problem is
that singular e-mail is not sent out singularly, but that it is
separated in to 500 separate messages, creating 500 times the traffic
load.
I have the same problem with a few customers.
I know that is how qmail is designed.
Anyway, I believe that is the problem he is referring to.
- eric
RaTao von J escribi�:
>
> so, if you have a ISP machine (with a T1) that will relay mail for you (with a 2
> 8.8k) I think that you should use it always! it's better to connect to the machi
> ne that is 2 hop's away than machines 200 hops away :)
>
> you should have your domain in locals, so the only step left is adding:
>
> :relay.some.isp.net
>
> to your control/smtproutes file
>
> that will create a "default" smtproute to the relay.
>
> regards,
> ratao
>
> On 29-May-99 Doug Lumpkin wrote:
> > Ok... Entire situation. One linux box with internet access (28.8 modem), on
> > a
> > network. They do not want employees to have internet access, so none of the
> > machines can reach anything other than what is on the local network. Qmail
> > is
> > set-up as the SMTP server and processes both interoffice and internet mail.
> > They
> > have small internet mail load and a large interoffice mail load, except when
> > once
> > a week a large mailing list is distributed. The interoffice mail is
> > distributed
> > locally and never has to traverse the internet. Importantly, This is what
> > they
> > want!
> >
> > What I would like to do is have qmail notice that the message it is
> > processing is
> > to more than 30 bcc addresses and then decide to pass that to a different
> > SMTP
> > server to be processed at the ISP. This way their dial-up line is not
> > cruching
> > messages for hours non-stop.
> >
> > I would appreciate any suggestions you might have...
> > --
> > Doug Lumpkin
> > PacInfo Internet
> > [EMAIL PROTECTED]
> >
> >
> >
> >
> >
> > John Gonzalez/netMDC admin wrote:
> >
> >> I dont see why this is necessary. Have you ever heard of virtual hosts?
> >> Mail exchangers? POP boxes? Virtual Domains? etc, etc?
> >>
> >> It might help us to better help you, if you explain the entire situation?
> >>
> >> On Fri, 28 May 1999, Doug Lumpkin wrote:
> >>
> >> >I realize there might be better ways to do this, but none of their machines
> >> >are connected to the internet, only the gateway machine is. So it has to
> >> >be
> >> >running SMTP to accept their messages and then direct them out onto the
> >> >net...
> >> >--
> >> >Doug Lumpkin
> >> >[EMAIL PROTECTED]
> >> >
> >> >
> >> >
> >> >John Gonzalez/netMDC admin wrote:
> >> >
> >> >> On Fri, 28 May 1999, Frederik Lindberg wrote:
> >> >>
> >> >> >qmail isn't made for dialups. Use the serialmail package for remote mail
> >> >> >instead. Local delivery with qmail and all remote mail goes to a Maildir
> >> >> >from where it is sent to the smarthost via serialmail.
> >> >>
> >> >> SMTP itself really isnt optimized for dialup, it's not just qmail.
> >> >>
> >> >> There are tons of ways to run a more efficient mailer from a dialup box
> >> >> without using SMTP or even serialmail.
> >> >>
> >> >> qmtp is an option
> >> >>
> >> >> Bruce Guenter has a nullmailer package that might be of some use.
> >> >>
> >> >> _ __ _____ __ _________
> >> >> ______________ /_______ ___ ____ /______ John Gonzalez/Net.Tech
> >> >> __ __ \ __ \ __/_ __ `__ \/ __ /_ ___/ MDC Computers/netMDC!
> >> >> _ / / / `__/ /_ / / / / / / /_/ / / /__ (505)437-7600/fax-437-3052
> >> >> /_/ /_/\___/\__/ /_/ /_/ /_/\__,_/ \___/ http://www.netmdc.com
> >> >> [---------------------------------------------[system info]-----------]
> >> >> 5:40pm up 113 days, 43 min, 3 users, load average: 0.13, 0.17, 0.18
> >> >
> >> >--
> >> >Doug Lumpkin
> >> >PacInfo Internet
> >> >[EMAIL PROTECTED]
> >> >
> >> >
> >> >
> >>
> >> _ __ _____ __ _________
> >> ______________ /_______ ___ ____ /______ John Gonzalez/Net.Tech
> >> __ __ \ __ \ __/_ __ `__ \/ __ /_ ___/ MDC Computers/netMDC!
> >> _ / / / `__/ /_ / / / / / / /_/ / / /__ (505)437-7600/fax-437-3052
> >> /_/ /_/\___/\__/ /_/ /_/ /_/\__,_/ \___/ http://www.netmdc.com
> >> [---------------------------------------------[system info]-----------]
> >> 6:00pm up 113 days, 1:03, 3 users, load average: 0.02, 0.09, 0.12
>
> ----------------------------------
> E-Mail: RaTao von J <[EMAIL PROTECTED]>
> Date: 29-May-99 Time: 01:56:40
> ----------------------------------
From: Eric Dahnke <[EMAIL PROTECTED]>
: I don't think any of you are understanding his problem. The problem is
: that singular e-mail is not sent out singularly, but that it is
: separated in to 500 separate messages, creating 500 times the traffic
: load.
:
: I have the same problem with a few customers.
:
: I know that is how qmail is designed.
No, that is how SMTP is designed. The only difference with qmail is that if
more than one of those messages are destined for the same host, qmail will not
combine them into a single SMTP session. With qmail you will get 500 separate
messages. With sendmail you will get somewhere less than 500, but certainly
pretty close to 500, unless all of the users on your mailing list are on only
a few different hosts.
--Adam
hi,
since the problem is only one e-mail with lots of BCC's I would suggest that you
don't use qmail for this message. instead setup a port forwarding from your mac
hine port 25000 to the ISP's relay port 25.
That user who sends the mlist send's _only_ the mlist to port 25000 (right to th
e relay...).
eheheh I know that this has nothing to do with qmail, but it's a quick hack that
does the job :)
any more ideas?
regards,
ratao
On 29-May-99 Adam D. McKenna wrote:
> From: Eric Dahnke <[EMAIL PROTECTED]>
>
>: I don't think any of you are understanding his problem. The problem is
>: that singular e-mail is not sent out singularly, but that it is
>: separated in to 500 separate messages, creating 500 times the traffic
>: load.
>:
>: I have the same problem with a few customers.
>:
>: I know that is how qmail is designed.
>
> No, that is how SMTP is designed. The only difference with qmail is that if
> more than one of those messages are destined for the same host, qmail will
> not
> combine them into a single SMTP session. With qmail you will get 500
> separate
> messages. With sendmail you will get somewhere less than 500, but certainly
> pretty close to 500, unless all of the users on your mailing list are on only
> a few different hosts.
>
> --Adam
----------------------------------
E-Mail: RaTao von J <[EMAIL PROTECTED]>
Date: 29-May-99 Time: 18:40:29
----------------------------------
Please pardon this off-topic question!
I'm about to lose my static IP--then it's back to dialup from
home. The problem is that I'm hosting an ezmlm mailing list, and want
to keep doing so.
Where should I look for a service which will rent me an MX record, and
spool my incoming mail for delivery when I am online? Autoturn would
be preferred, but any arrangement which preserves envelope information
is okay. I just don't want to grok headers, for all the obvious
reasons.
I've checked several ISPs, the Qmail archives, and DejaNews to no
avail.
Thanks in advance,
Len.
--
19. Let your Countenance be pleasant but in Serious Matters Somewhat
grave.
-- George Washington, "Rules of Civility & Decent Behaviour"
Hello.
We have a small, single-real-IP, masquaderaded Linux network. We run qmail
on the so-called metabox, which uses NFS home directories to put received
mail in users' ~/Mailbox. Works quite well, generally.
On the hosts containing the NFS exported home directories, there is also the
lingering /var/spool/mail tree. In some cases we've used this in
conjunction with fetchpop or mutt's function to get at other mailboxes, and,
although a bit awkward, it's a nice facility.
Problem is that everything "detecting" mail checks for the
/var/spool/mail/user file, not ~/Mailbox. I can't find documentation about
how to have the "you have mail"/"you have new mail" etc. be based on the
"real" mailbox in the home directory.
Another issue is this: I intend to get procmail (or is there a better way?)
to sort out mailing lists from BOTH the mailboxes written to by qmail, and
the fetchpop'd ones. Has anyone been through this? I'm reading the man
pages and they look good, but I'm looking for pointers in this particular
unique situation.
TIA,
Brian Butler
[EMAIL PROTECTED]
This is strange. We have a customer who sends out a monthly newsletter to about 20,000
people. He's done this every month for about 5 months, but last month, for some reason,
all the bounces came to ME. This is definitely quite mysterious because it is actually
placing my email address, with the full system host name as the Return-path. The mail
is
being sent via qmail-inject and even when a proper Return-path is specified, it still
replaces it with my personal email address. Like I said, no aliases are attached to it.
The only possible explanation I can come up with for qmail knowing my email address is
that I was su'd to root from my account when I compiled it but I haven't recompiled it
for at least 4 months.
Any one have any ideas?
Thanks
--
Cris Daniluk [EMAIL PROTECTED]
-------------------------------------------------------------
Digital Services Network, Inc. http://www.dsnet.net
1129 Niles-Cortland Road, Warren, Ohio 44484 [EMAIL PROTECTED]
(330) 609-8624 ext. 20 Fax (330) 609-9990
The Web Hosting Specialists
-------------------------------------------------------------
Check to see if the QMAILINJECT environment variable is set..
-Dustin
On Sat, 29 May 1999, Cris Daniluk wrote:
> This is strange. We have a customer who sends out a monthly newsletter to about
>20,000
> people. He's done this every month for about 5 months, but last month, for some
>reason,
> all the bounces came to ME. This is definitely quite mysterious because it is
>actually
> placing my email address, with the full system host name as the Return-path. The
>mail is
> being sent via qmail-inject and even when a proper Return-path is specified, it still
> replaces it with my personal email address. Like I said, no aliases are attached to
>it.
> The only possible explanation I can come up with for qmail knowing my email address
>is
> that I was su'd to root from my account when I compiled it but I haven't recompiled
>it
> for at least 4 months.
>
> Any one have any ideas?
>
> Thanks
>
>
> --
> Cris Daniluk [EMAIL PROTECTED]
> -------------------------------------------------------------
> Digital Services Network, Inc. http://www.dsnet.net
> 1129 Niles-Cortland Road, Warren, Ohio 44484 [EMAIL PROTECTED]
> (330) 609-8624 ext. 20 Fax (330) 609-9990
> The Web Hosting Specialists
> -------------------------------------------------------------
>
>
>
>
Hello!
I had the tcp-wrappers/relayclient happily running along on one of our
Solaris 2.5.1 boxes with out any problems.
However, we have just upgraded to Solaris 2.7, recompiled tcp-wrappers
for sunos5, and now when a user attempts to send a message, they get a
time out message for port 25 of the given machine. Without washing the
connection through tcpd, everything works fine.
Anyone have any suggestions?
Thanks,
Mark
--
Mark I. Recio
mailto:[EMAIL PROTECTED]
