qmail Digest 26 Mar 1999 11:00:01 -0000 Issue 591
Topics (messages 23438 through 23500):
failure notice (fwd)
23438 by: Jos Backus <[EMAIL PROTECTED]>
Doc Wars (was: keyserver)
23439 by: [EMAIL PROTECTED]
23444 by: Mark Bitting <[EMAIL PROTECTED]>
keyserver
23440 by: [EMAIL PROTECTED]
23450 by: Mark Delany <[EMAIL PROTECTED]>
23452 by: Kai MacTane <[EMAIL PROTECTED]>
23471 by: "D. J. Bernstein" <[EMAIL PROTECTED]>
23472 by: [EMAIL PROTECTED]
23474 by: "Adam D. McKenna" <[EMAIL PROTECTED]>
23476 by: Paul Farber <[EMAIL PROTECTED]>
23485 by: Russell Nelson <[EMAIL PROTECTED]>
23487 by: Russ Allbery <[EMAIL PROTECTED]>
23488 by: Russell Nelson <[EMAIL PROTECTED]>
23492 by: [EMAIL PROTECTED]
O.T.(?): .fetchmailrc
23441 by: "Sam" <[EMAIL PROTECTED]>
What goes in the official patch?
23442 by: Russell Nelson <[EMAIL PROTECTED]>
23443 by: Dave Sill <[EMAIL PROTECTED]>
23445 by: Van Liedekerke Franky <[EMAIL PROTECTED]>
23449 by: Kai MacTane <[EMAIL PROTECTED]>
23451 by: Chris Johnson <[EMAIL PROTECTED]>
Specific bounces
23446 by: Jean Caron <[EMAIL PROTECTED]>
23486 by: Anand Buddhdev <[EMAIL PROTECTED]>
Rewriting headers
23447 by: Dirk Alboth <[EMAIL PROTECTED]>
23453 by: Scott Schwartz <[EMAIL PROTECTED]>
23454 by: Mark Delany <[EMAIL PROTECTED]>
23466 by: Stefan Paletta <[EMAIL PROTECTED]>
23499 by: Dirk Alboth <[EMAIL PROTECTED]>
23500 by: Dirk Alboth <[EMAIL PROTECTED]>
fixcr?
23448 by: Bill Luckett <[EMAIL PROTECTED]>
virtualdomains with fastforward
23455 by: Niklas Alberth <[EMAIL PROTECTED]>
qmail-inject multiple recipients?
23456 by: Jeff Hill <[EMAIL PROTECTED]>
23473 by: Mark Delany <[EMAIL PROTECTED]>
23475 by: "Adam D. McKenna" <[EMAIL PROTECTED]>
How to monitor qmail-send and friends?
23457 by: "D. J. Bernstein" <[EMAIL PROTECTED]>
dot-qmail security
23458 by: "D. J. Bernstein" <[EMAIL PROTECTED]>
Handling of dead remote hosts
23459 by: "D. J. Bernstein" <[EMAIL PROTECTED]>
23469 by: Steven Levis <[EMAIL PROTECTED]>
qmail version 2?
23460 by: David Villeger <[EMAIL PROTECTED]>
23461 by: Stefan Paletta <[EMAIL PROTECTED]>
ezmlm and "delay notifies" (was: Re: mini-bounce)
23462 by: "D. J. Bernstein" <[EMAIL PROTECTED]>
daemontools and sshd
23463 by: <[EMAIL PROTECTED]>
MX Problems through Firewall
23464 by: "D. J. Bernstein" <[EMAIL PROTECTED]>
23482 by: [EMAIL PROTECTED]
ETRN, qmail-1.03 and etrn patch v0.1f
23465 by: "D. J. Bernstein" <[EMAIL PROTECTED]>
mail delayed, why ?
23467 by: Siegfried Kerkow <[EMAIL PROTECTED]>
23468 by: Stefan Paletta <[EMAIL PROTECTED]>
23493 by: Siegfried Kerkow <[EMAIL PROTECTED]>
23494 by: Anand Buddhdev <[EMAIL PROTECTED]>
23495 by: Les Klein <[EMAIL PROTECTED]>
23496 by: Siegfried Kerkow <[EMAIL PROTECTED]>
23497 by: Siegfried Kerkow <[EMAIL PROTECTED]>
Aliases - silly newbie question but I can't figure it out
23470 by: "D. J. Bernstein" <[EMAIL PROTECTED]>
Sendmail for NT
23477 by: Kai MacTane <[EMAIL PROTECTED]>
23478 by: [EMAIL PROTECTED]
23479 by: [EMAIL PROTECTED]
23480 by: "Scott D. Yelich" <[EMAIL PROTECTED]>
23481 by: "Adam D. McKenna" <[EMAIL PROTECTED]>
23483 by: Joel Shellman <[EMAIL PROTECTED]>
23489 by: Uwe Wuerdinger <[EMAIL PROTECTED]>
qmail vs lsmtp
23484 by: David Villeger <[EMAIL PROTECTED]>
23490 by: Dustin Marquess <[EMAIL PROTECTED]>
cyclog problem on Solaris
23491 by: Dongping Deng <[EMAIL PROTECTED]>
qmail, perl, and header rewrites
23498 by: "Efg�" <[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]
----------------------------------------------------------------------
Hi Greg,
On Tue, Mar 23, 1999 at 06:02:16PM -0500, xs wrote:
> hey all, since it seems that jos' email address is no longer valid, i'd
> figure i'd throw the question out here for an answer.
<[EMAIL PROTECTED]> works (I changed jobs twice quite a while ago).
> hi there jos, i was taking a look at your toolarge program and i had a few
> questions:
>
> in:
>
> echo '|condredirect $USER-toolarge $HOME/toolarge test'>.qmail-default
> echo '|$HOME/toolarge bounce"'>.qmail-toolarge
>
> i currently have "./Maildir/" in my .qmail, now would i enter this code
> before or after "./Maildir/",
Before. You would use
$ cat $HOME/.qmail
|condredirect $USER-toolarge $HOME/toolarge test
./Maildir/
and
$ cat $HOME/.qmail-toolarge
|$HOME/toolarge bounce
> or does the fact that it's in .qmail-default make qmail look at it first
> then .qmail?
Nope, you can stick this in whatever .qmail file you wish, as long as you take
care to point ``$USER-toolarge'' to ~user/.qmail-toolarge.
> thanks for the help.
Hope this does.
Cheers,
--
Jos Backus _/ _/_/_/ "Reliability means never
_/ _/ _/ having to say you're sorry."
_/ _/_/_/ -- D. J. Bernstein
_/ _/ _/ _/
[EMAIL PROTECTED] _/_/ _/_/_/ use Std::Disclaimer;
Forgive me plunging into the fray, but...
Mr. Yelich, I doubt a sould would disagree with you if you said,
"Golly, wouldn't a nice comprehensive O'Reilly book be great?
Something like the great classics--the expect book and the Perl camel
book." BTW, such a book is in progress.
You set up a straw man, however, when you slam "the
documentation". True, there is not yet a rich literature, including a
dummies book and a nutshell book, but "the documentation" is quite
complete.
"Scott D. Yelich" <[EMAIL PROTECTED]> writes:
> On Wed, 24 Mar 1999, Adam D. McKenna wrote:
> > ...I went to www.qmail.org and I read man pages for 3 hours. So
> > when someone tells you to RTFM, it might not be because they're
> > trying to be hostile, but that RTFM'ing is really what you need to
> > do in order to start grasping the concepts that go along with this
> > piece of radically different software.
>
> We've addressed this. What do you do when the FM isn't enough?
Straw man. You wrote this mailing list; apparently you do comprehend
that this is a qmail resource. Are you unaware that it is archived?
Having asked questions so inane that they were completely ignored by
the nice folks of this list, I learned through personal shame that TFM
includes the archive of this mailing list, and to a lesser extent
bugtraq and dejanews.
Had you employed those resources, you would have had your
answer--without needing to post a rather obnoxious rant, and starting
a flame-war.
> when is djb coming out with his own unix?
Are you asking, or ranting? Can't quite tell. You should know that
beside qmail, DJB is a regular contributor to bugtraq, faq-keeper of
comp.security.unix and sci.crypt, and writes crypto software. Unlike
Stallman, he doesn't want to replace everything with DJB-ware; just
the stuff that is insecure or faulty.
> Also, what is part of qmail and what isn't?
You downloaded qmail, right? Look very carefully inside the
tarball. There's your answer. Is that RTFM? Or merely the moral
equivalent?
> Does anyone have a list of what is "official" and what is not? does
> anyone care?
The "patch culture" evolved because DJB will not risk qmail's
security, reliability and speed by folding in everyone's pet projects.
Everyone thinks their pet, unlike everyone else's pet, is beautiful
and necessary, and belongs in qmail. DJB resists.
Hence "official patch" is rather an oxymoron. Use what you like, at
your own risk. Personal responsibility is rather bracing--try it out.
Len.
--
26. In Pulling off your Hat to Persons of Distinction, as Noblemen,
Justices, Churchmen &c make a Reverence, bowing more or less according
to the Custom of the Better Bred, and Quality of the Person. Amongst
your equals expect not always that they Should begin with you first,
but to Pull off the Hat when there is no need is Affectation, in the
Manner of Saluting and resaluting in words keep to the most usual Custom.
-- George Washington, "Rules of Civility & Decent Behaviour"
As the most inept and clueless *nix sysadmin on the web, I'd like to say
a few words about the qmail docs... They're damned good. I fought with
sendmail for a long time, the quarterly upgrades became less and less
functional for me. At 8.9.3, all I could get it to do was bounce
everything that came its way. Sure, there's the 25-pound Bat Book, but
it's written in computerese, not English. Figuring there had to be
something easier to set up, I searched and found qmail, and decided to
give it a try out of desperation. I RTFM, followed the instructions,
and it all worked. I dug around a little more, found tcpserver and the
instructions for that, and got pop boxes working on multiple domains and
selective relaying for the machines inside the firewall. Some things
were a bit vague, it took maybe a half hour of experimentation to find
out what worked. Sure, the docs on the conceptual points of qmail could
be improved, but they're still better than anything I ever found on
sendmail. Hell, I'm dumber than a brick and I got three qmail servers
running the way I want them to and doing the tricks I could never get
sendmail to do. My experience with this list has been that there is a
wealth of knowledge and help available to those who have made an honest
effort and got stuck, but little sympathy for those who want all the
answers without any expending any energy of their own. RTFM - it works.
</rant>
Mark Bitting
"Scott D. Yelich" <[EMAIL PROTECTED]> writes:
>
> How many [rock solid, secure, fast, small, multi-use] programs does
> qmail have? I've lost count... I have qmail, ezmlm, dotforward,
> accustamp, cyclog, tcp-env, tcpserver, setuser...
As a side note, I use accustamp, setuser, cyclog and tcpserver all the
time. Any boot-script I write uses them in preference to more
"standard" tools. And you did miss a few...
You missed supervise, which comes in the daemontools package with
accustamp. Thanks to supervise, writing flexible reliable daemons is a
snap. Thanks to accustamp, splogger and cyclog, interacting with the
system logger is cake. Thanks to tcpserver and tcp-env, building
internet services is a breeze.
Err, did you miss the point that modularity means flexibility? Those
many tools exist because of the many different tasks that need doing?
> Sure, it works... eventually... maybe. It might be efficient (as
> long as you're not on a serial line?).
That's FAQ 2.4. RTFM.
> And, to top all of this off, the docs suck! I mean, it's not that
> qmail is bad, really, it isn't! It's the docs.
Terse != sucky.
Try this, when you have questions in the future:
1. man qmail (Follow cross-references, including:)
a. man qmail-start (addresses resource limits)
b. man qmail-control (Indexes ALL control files. Follow references.)
2. less /usr/doc/qmail/FAQ (better keep the FAQ somewhere...)
3. lynx|netscape http://www-archive.ornl.gov:8000/ (Search the archive)
4. lynx|netscape http://www.qmail.org/
> You can find a doc on sendwhale to do just about anything you need.
It's there for qmail; it's just not said twelve ways for every kind of
wizard and dummy, true...
> Besides, I can grok sendmail.cf -- I kinda like it. I have often
> thought of writing an AI/Expert-System that uses many of its techniques.
Good idea. Make a mission-critical, 100% reliable, secure system using
those techniques. Good luck to you!
Len.
--
44. When a man does all he can though it Succeeds not well blame not
him that did it.
-- George Washington, "Rules of Civility & Decent Behaviour"
> Ya know.... I just got back from a new consulting job. This place paid
> someone several hundred dollars to install qmail and the person never
Sounds cheap for a real qmail installation.
> did get it working -- after several WEEKs of work. I'm sure you'll be
Someone got a consultant to work for "several WEEKS" for 200 lousy bucks?
So lemme get this right. They got stung badly by hiring someone who knew
virtually nothing about qmail install and didn't have the courage to admit
they knew nothing.
Having been burnt by a dishonest and incompetent consultant they are now...
> delighted to know that I have the job now.
handing that job over to someone who sounds like they have never done a
single qmail install?
Suggestions:
1. Get a real consultant (there are a number on www.qmail.org who can give
real references and have credibility that can be checked on this list)
2. Ask them to teach you the things you want to know about qmail and the
associated tools
3. Pay them real money for the site analysis, the installation, the
follow-up support and the training.
Presumably you get paid for the skills you do have, why not pay someone for
the skills you don't have yet?
Regards.
Text written by Mark Delany at 09:46 AM 3/25/99 -0800:
>> Ya know.... I just got back from a new consulting job. This place paid
>> someone several hundred dollars to install qmail and the person never
> ^^^^^^^
>> did get it working -- after several WEEKs of work. I'm sure you'll be
>
>Someone got a consultant to work for "several WEEKS" for 200 lousy bucks?
Probably more than 200 -- "several" is more than "a couple" (and even more
than "a few" -- the person may have made off with more than US$1000...).
Just a minor nit-pick. :)
-----------------------------------------------------------------
Kai MacTane
System Administrator
Online Partners.com, Inc.
-----------------------------------------------------------------
>From the Jargon File: (v4.0.0, 25 Jul 1996)
drop on the floor /vt./
To react to an error condition by silently discarding messages or
other valuable data. "The gateway ran out of memory, so it just
started dropping packets on the floor." Also frequently used of
faulty mail and netnews relay sites that lose messages.
Scott D. Yelich writes:
> It even comes default as an open relay!
Not if you follow the installation instructions. See that file named
INSTALL? It's mentioned prominently in README.
> I asked someone to give me an
> example of a standard unix program that called itself as part of its
> normal/standard use.
Some standard UNIX tools that run programs named on the command line:
chroot
env
find
nice
nohup
rsh
sh
su
time
xargs
It's hardly a surprise to see sh calling itself.
Obviously you weren't able to extrapolate from the example in step 10 of
INSTALL in rblsmtpd 0.70. A future release will have more examples.
---Dan
>Sendmail is not a Unix program. It is an NT program that someone
>ported to Unix twenty years ago.
Hmm...I can't quite see the humor in this, though it strikes me as
a bit funny, and not entirely untrue.
But, I'll ask anyway: did you really intend to suggest that Sendmail
was originally written for Windows NT over twenty years ago?
(I don't happen to know the history of Sendmail, and don't really care
all that much. All I know is how grateful I was to not have to buy the
huge ORA Sendmail book when buying other ORA books -- because I had
already decided to try qmail. If Sendmail is really that old, I'd
probably guess it comes from VAX/VMS, if not UNIX.)
tq vm, (burley)
From: <[EMAIL PROTECTED]>
> >Sendmail is not a Unix program. It is an NT program that someone
> >ported to Unix twenty years ago.
>
> Hmm...I can't quite see the humor in this, though it strikes me as
> a bit funny, and not entirely untrue.
>
> But, I'll ask anyway: did you really intend to suggest that Sendmail
> was originally written for Windows NT over twenty years ago?
Yes. Sendmail was written for NT 0.01 pre-beta in 1979. It was ported to
UNIX by Bill Gates in 1981.
--Adam
Does Allman know of this? hehehehe
Just to be sure.. i'm gonna name all my kids bill.. it must have some
magic power to it... you can srew up and never get blamed for anything
(Bill Gates - NT & OS/2 Bill Clinton.. a pattern is forming!)
Paul D. Farber II
Farber Technology
Ph. 570-628-5303
Fax 570-628-5545
[EMAIL PROTECTED]
On Thu, 25 Mar 1999, Adam D. McKenna wrote:
> From: <[EMAIL PROTECTED]>
>
>
> > >Sendmail is not a Unix program. It is an NT program that someone
> > >ported to Unix twenty years ago.
> >
> > Hmm...I can't quite see the humor in this, though it strikes me as
> > a bit funny, and not entirely untrue.
> >
> > But, I'll ask anyway: did you really intend to suggest that Sendmail
> > was originally written for Windows NT over twenty years ago?
>
> Yes. Sendmail was written for NT 0.01 pre-beta in 1979. It was ported to
> UNIX by Bill Gates in 1981.
>
> --Adam
>
>
>
>
[EMAIL PROTECTED] writes:
> But, I'll ask anyway: did you really intend to suggest that Sendmail
> was originally written for Windows NT over twenty years ago?
My intent was to point out how it is unlike most other Unix programs.
For example, there's lpr, lpd, and lpq. Sendmail folds all those
concepts into one program.
I like Henry Spencer's .signature quip:
"Those who do not understand Unix are bound to reinvent it, poorly."
--
-russ nelson <[EMAIL PROTECTED]> http://crynwr.com/~nelson
Crynwr supports Open Source(tm) Software| PGPok | There is good evidence
521 Pleasant Valley Rd. | +1 315 268 1925 voice | that freedom is the
Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | cause of world peace.
Russell Nelson <[EMAIL PROTECTED]> writes:
> [EMAIL PROTECTED] writes:
>> But, I'll ask anyway: did you really intend to suggest that Sendmail
>> was originally written for Windows NT over twenty years ago?
> My intent was to point out how it is unlike most other Unix programs.
> For example, there's lpr, lpd, and lpq. Sendmail folds all those
> concepts into one program.
Just like INN. At the time that INN was written, this was considered a
major feature (and indeed was a huge win in terms of performance). It's
also worth noting that sendmail started out as a much simpler program and
grew in complexity as mail did. It's as huge and complex as it is largely
because it has grown with mail and Unix and has never to my knowledge gone
through a ground-up rewrite (which it, by this point, needs, just like
most software that's grown randomly over the years).
> I like Henry Spencer's .signature quip:
> "Those who do not understand Unix are bound to reinvent it, poorly."
I hardly think Eric Allman doesn't understand Unix. :) But he, and
sendmail, come from a very early era of it.
--
Russ Allbery ([EMAIL PROTECTED]) <URL:http://www.eyrie.org/~eagle/>
Russ Allbery writes:
> > I like Henry Spencer's .signature quip:
>
> > "Those who do not understand Unix are bound to reinvent it, poorly."
>
> I hardly think Eric Allman doesn't understand Unix. :) But he, and
> sendmail, come from a very early era of it.
Well, I meant NT, and I think Henry did too.
--
-russ nelson <[EMAIL PROTECTED]> http://crynwr.com/~nelson
Crynwr supports Open Source(tm) Software| PGPok | There is good evidence
521 Pleasant Valley Rd. | +1 315 268 1925 voice | that freedom is the
Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | cause of world peace.
>[EMAIL PROTECTED] writes:
> > But, I'll ask anyway: did you really intend to suggest that Sendmail
> > was originally written for Windows NT over twenty years ago?
>
>My intent was to point out how it is unlike most other Unix programs.
>For example, there's lpr, lpd, and lpq. Sendmail folds all those
>concepts into one program.
Ah. Yes. Wish I'd listened to that "annoying" guy in Tech Pubs back in
early 1978, when he kept going on and on about this wonderful system
with things like "shells", "pipes", and so on. I had my hands full
wrapping my head around PRIMOS, after being used to TOPS-10, ITS, etc.
(Then again, having been partially exposed to Pr1me's efforts to "port"
UNIX to its hardware, I have fewer illusions about the so-called
"portability" of UNIX than most adherents. Though, IIRC, most of the
real problems were in the piles of C code that assumed every system
was a VAX....)
tq vm, (burley)
Enrico Mangano writes:
> fetchmail: SMTP listener doesn't like recipient address 'enr1co@out'
> fetchmail: can't even send to calling user!
> fetchmail: SMTP transaction error while fetching from popmail.iol.it
>
> What should i do??
Delete fetchmail. Use something that's more sane.
--
Sam
[EMAIL PROTECTED] writes:
> Hence "official patch" is rather an oxymoron. Use what you like, at
> your own risk. Personal responsibility is rather bracing--try it out.
Still, while I respect Dan's desire to control qmail, it wouldn't hurt
us to develop an official patch.
So, what goes in the official patch?
--
-russ nelson <[EMAIL PROTECTED]> http://crynwr.com/~nelson
Crynwr supports Open Source(tm) Software| PGPok | There is good evidence
521 Pleasant Valley Rd. | +1 315 268 1925 voice | that freedom is the
Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | cause of world peace.
Russell Nelson <[EMAIL PROTECTED]> wrote:
>
>Still, while I respect Dan's desire to control qmail, it wouldn't hurt
>us to develop an official patch.
I agree, but I think all of the patches included should be optional
and turned off by default. A single config file with a short
description of each patch and a mechanism for enabling it would be
nice.
>So, what goes in the official patch?
I don't use any of the patches, but the most popular ones seem to be
those that do spam control and improved logging.
-Dave
I have a couple of suggestions as well:
- the patch that deals with oversized dns packets
- the patch that limits the number of RCPT TO's
- the tarpitting patch
- the patch that limits the size of the bounce mail
- of course: one of the UCE patches (I like Sam's patch a lot, which already
combines some other patches)
- and a personal favorite of mine: the ldap patch (indeed witch an option to
compile it in or not)
Franky
> ----------
> From: Dave Sill[SMTP:[EMAIL PROTECTED]]
> Sent: Thursday, March 25, 1999 3:14 PM
> To: [EMAIL PROTECTED]
> Subject: Re: What goes in the official patch?
>
> Russell Nelson <[EMAIL PROTECTED]> wrote:
> >
> >Still, while I respect Dan's desire to control qmail, it wouldn't hurt
> >us to develop an official patch.
>
> I agree, but I think all of the patches included should be optional
> and turned off by default. A single config file with a short
> description of each patch and a mechanism for enabling it would be
> nice.
>
> >So, what goes in the official patch?
>
> I don't use any of the patches, but the most popular ones seem to be
> those that do spam control and improved logging.
>
> -Dave
>
Text written by Dave Sill at 09:14 AM 3/25/99 -0500:
>
>I agree, but I think all of the patches included should be optional
>and turned off by default. A single config file with a short
>description of each patch and a mechanism for enabling it would be
>nice.
I think this is an excellent suggestion. So the "official patch" would
essentially be like one of those "software samplers" that come with every
Boot magazine (for example), plus maybe an interactive shell script or
something that gives the info you mention.
>>So, what goes in the official patch?
>
>I don't use any of the patches, but the most popular ones seem to be
>those that do spam control and improved logging.
I'd add Russ' Open-SMTP to that list, and I'd suggest that the official
patch include daemontools, since it is not included in the standard Qmail
download. Other patches that I hear about a lot on this list (this is just
off the top of my head, and totally non-scientific) are: rblsmtpd,
qmail-uce, qmail-popbull and quite a few that I'm accidentally forgetting. :)
-----------------------------------------------------------------
Kai MacTane
System Administrator
Online Partners.com, Inc.
-----------------------------------------------------------------
>From the Jargon File: (v4.0.0, 25 Jul 1996)
can't happen
The traditional program comment for code executed under a condition
that should never be true, for example a file size computed as
negative ... Although "can't happen" events are genuinely infrequent
in production code, programmers wise enough to check for them habitu-
ally are often surprised at how frequently they are triggered during
development and how many headaches checking for them turns out to
head off.
On Thu, Mar 25, 1999 at 09:37:53AM -0800, Kai MacTane wrote (in his signature):
> From the Jargon File: (v4.0.0, 25 Jul 1996)
>
> can't happen
>
> The traditional program comment for code executed under a condition
> that should never be true, for example a file size computed as
> negative ... Although "can't happen" events are genuinely infrequent
> in production code, programmers wise enough to check for them habitu-
> ally are often surprised at how frequently they are triggered during
> development and how many headaches checking for them turns out to
> head off.
This is way, way off topic, but the programmers of Microsoft Outlook Express 5,
which was just released last week, apparently think that a mail message with
the Subject field missing from the headers "can't happen." Unfortunately, they
weren't wise enough to check for this anyway, and OE5 will crash it it receives
such a message. (The Subject header field has to be completely missing, not
just empty.)
Chris
Hi all,
Just wondering how much is involved in creating a "unique" bounce message
for a specific user. I wouldn't want to damage the regular process, just
had to it. Basically, if John Doe sends a message to a valid address,
bounce it anyway using this "modified" version of the bounce text/message.
What do you think ?
John
--
On Thu, Mar 25, 1999 at 11:00:08AM -0500, Jean Caron wrote:
> Hi all,
>
> Just wondering how much is involved in creating a "unique" bounce message
> for a specific user. I wouldn't want to damage the regular process, just
> had to it. Basically, if John Doe sends a message to a valid address,
> bounce it anyway using this "modified" version of the bounce text/message.
It depends on whether you want to append a specific message to the usual
bounce message generated by qmail, or generate a completely different
bounce message.
For the former, stick this into a .qmail file for the valid address:
|bouncesaying "some specific message" [ "$SENDER" = "John.Doe@whatever" ]
./Mailbox
This will return a normal qmail bounce ("Hi. This is the ...") to John Doe,
with "some specific message" as the reason for the bounce.
If you want to construct a completely specific bounce message, then you'll
have to check if the sender is John Doe, and inject a *new* message to him,
with a null envelope sender, and a From: header of [EMAIL PROTECTED],
and consisting of a body of your choice. Something in .qmail like:
|if [ "$SENDER" = "John.Doe@whatever" ] then; (echo "From:
[EMAIL PROTECTED]; echo ""; echo body) | qmail-inject -f '';fi
./Mailbox
This is not tested code, just a pointer to how it can be done.
--
System Administrator
See complete headers for address, homepage and phone numbers
Hi,
I'm new to the list (though not really new to qmail -- used it for
approx. three years now), so please forgive my possible unawareness.
Anyway, I have the following problem and I'd like to receive pros and
cons on the solutions I'm thinking of:
At my site most users use Netscape under MS-WinNT to read and write
mail. I also run a central mail server (qmail-1.01) where all
outgoing mail is relayed. I now have realized that by simply entering
a different address in Netscape's
"Edit/Preferences/Mail&Groups/Identity" a user may change his
"From:" and "Return-Path:" header lines line to what he likes.
As I understand RFC 822 this is not violating the standard but in this
case a "Sender:" field should reveal the true sender's identity
(agreed?).
Netscape however does not do this. Inserting a
"user_pref("mail.suppress_sender_header", false)" in Netscape's
prefs.js seems to have no effect under WinNT (at least for the version
I've tried).
So I'm thinking of letting qmail at our mail server correct for this
situation.
I doubt that the easiest remedy, to just bounce those messages (with
the wildcard patch and badmailfrom or the equiv. this patch uses),
will be acceptable to (some of) my users, so the idea is to let the
"Sender:"-field be added somewhere in the qmail-smtpd->qmail-queue
chain.
Now the "true" sender name will be ${TCPREMOTEINFO}@${TCPREMOTEHOST}
environment variable which is available to qmail-smtpd and also
(thanks to "execv") to qmail-queue. This requires an identd to run on
the NT box, of course.
So far, so good. The point where I'm undecided is the place where in
the qmail-smtpd->qmail-queue chain to add the header rewriting code.
I could do it either as a wrapper for qmail-smtpd or qmail-queue
(e.g. using Bruce Guenter's QMAILQUEUE-patch) the or by patching
either programs.
I'd prefer of course the wrapper solution as upgrading would be easier
(even though parsing the header lines would be somewhat tedious).
However, I'd like to bounce the mail if the identd-server is not
running on the other side (and hence $TCPREMOTEINFO will be empty).
This however seems hard to do with a wrapper. As my understanding
goes I would have been forced to bypass qmail-smtpd in this case and
write the bounce mail in the wrapper.
Patching qmail-smtpd would be better from this point of view.
Another solution could be to introduce a new environment variable
(QMAILSMTPBOUNCE) which if set forces (a minimally patched) qmail-smtpd
to bounce the message and e.g. write this variable's contents as
explanation to the user.
Any comments?
Thank you for listening.
Best regards,
Dirk
Dirk Alboth <[EMAIL PROTECTED]> writes:
| As I understand RFC 822 this is not violating the standard but in this
| case a "Sender:" field should reveal the true sender's identity
| (agreed?).
Should != Must. You can't stop people from lying.
Your only recourse is to cryptographically sign messages. Then the
recipients have some way to check the veracity of the putative sender.
I wouldn't bother hacking qmail when the cryptographic solution is
philosophically and practically superior.
| Now the "true" sender name will be ${TCPREMOTEINFO}@${TCPREMOTEHOST}
Pointless, since TCPREMOTEINFO is whatever the sender wants it to be.
It's for debugging, not security.
At 01:34 PM Thursday 3/25/99, Scott Schwartz wrote:
>Dirk Alboth <[EMAIL PROTECTED]> writes:
>| As I understand RFC 822 this is not violating the standard but in this
>| case a "Sender:" field should reveal the true sender's identity
>| (agreed?).
>
>Should != Must. You can't stop people from lying.
>
>Your only recourse is to cryptographically sign messages. Then the
>recipients have some way to check the veracity of the putative sender.
>
>I wouldn't bother hacking qmail when the cryptographic solution is
>philosophically and practically superior.
>
>| Now the "true" sender name will be ${TCPREMOTEINFO}@${TCPREMOTEHOST}
>
>Pointless, since TCPREMOTEINFO is whatever the sender wants it to be.
>It's for debugging, not security.
As an addendum to Scott's observations, TCPREMOTEHOST (or leastwise
TCPREMOTEIP) is recorded in the Received: header so you have certainty over
knowing which IP address originated the email. That's probably not useful
for a remote recipient, but assuming you have access to your
IPaddress->location information, you will be able to identify the sending PC.
Of course if a malicious insider has used some other persons PC, you wont
know from either Received: or Sender: headers. Certainly if someone accussed
me of sending an email based solely in Sender: or Received: I would get most
indignant (especially if I hadn't sent it :> )
So, as Scott says. Nothing shy of well deployed crypto-signatures help
identify the true sender.
Regards.
Dirk Alboth wrote/schrieb/scribsit:
> Now the "true" sender name will be ${TCPREMOTEINFO}@${TCPREMOTEHOST}
> environment variable which is available to qmail-smtpd and also
> (thanks to "execv") to qmail-queue. This requires an identd to run on
> the NT box, of course.
Here's what I do in similar cases, though it may not be applicable to your
case: If you have a Samba server, you can get ident info from the smbstatus
command. Basically grab the logged-in username from the line with
$TCPREMOTEIP in it and put it into TCPREMOTEINFO. Thanks to
tcp{server,client} this is even possible if the Samba and qmail server are
different machines. This can be a 10 line /bin/sh wrapper to qmail-smtpd.
> However, I'd like to bounce the mail if the identd-server is not
> running on the other side (and hence $TCPREMOTEINFO will be empty).
> This however seems hard to do with a wrapper.
Not at all. If a shell script offends thee, take a look at the rblsmtpd
code.
Stefan
Thank you to all who responded.
> At 01:34 PM Thursday 3/25/99, Scott Schwartz wrote:
> >Dirk Alboth <[EMAIL PROTECTED]> writes:
> >| As I understand RFC 822 this is not violating the standard but in this
> >| case a "Sender:" field should reveal the true sender's identity
> >| (agreed?).
> >
> >Should != Must. You can't stop people from lying.
Yes, my wording was not correct. RFC 822 says "MUST".
> >Your only recourse is to cryptographically sign messages. Then the
> >recipients have some way to check the veracity of the putative
> >sender.
I'm not concerned about strong authentication but rather to only let
out mails that comply with the rfc.
> >| Now the "true" sender name will be ${TCPREMOTEINFO}@${TCPREMOTEHOST}
> >
> >Pointless, since TCPREMOTEINFO is whatever the sender wants it to be.
> >It's for debugging, not security.
Sure, the one who has control over the machine can let port 113 answer
whatever he likes. In this respect, however, a company LAN is a bit
different to the internet, right? While I can to a certain degree
control over what services run on our machines I have little control
over what users enter at a menu prompt.
> As an addendum to Scott's observations, TCPREMOTEHOST (or leastwise
> TCPREMOTEIP) is recorded in the Received: header so you have
> certainty over knowing which IP address originated the email.
>
> Of course if a malicious insider has used some other persons PC, you
> wont know from either Received: or Sender: headers. Certainly if
> someone accussed me of sending an email based solely in Sender: or
> Received: I would get most indignant (especially if I hadn't sent it
> :> )
In my case it's not a malicious user who wanted to impersonate another
one but someone who uses another domain in the From: header. I assume
it's the person's private address. One may consider this as being
correct in the rfc822 sense: if he writes a private email then the
'identity of the person who wished this message to be sent' (talking
in rfc822 language) is not [EMAIL PROTECTED] but rather
[EMAIL PROTECTED] But the 'identity of the AGENT that sends the
message' is still [EMAIL PROTECTED] because by definition he is this
person whenever he uses the computer at work.
Of course checking the mail headers, in particular the received:
header, one can track the path of the message, if one would be
concerned to identify a malicious user. But as I said that's not the
point here.
Best regards,
Dirk
> Dirk Alboth wrote/schrieb/scribsit:
>
> Here's what I do in similar cases, though it may not be applicable
> to your case: If you have a Samba server, you can get ident info
> from the smbstatus command.
Good idea. I actually did also think of this but in our situation one
need not necessarily be connected to the Samba server.
> > However, I'd like to bounce the mail if the identd-server is not
> > running on the other side (and hence $TCPREMOTEINFO will be empty).
> > This however seems hard to do with a wrapper.
>
> Not at all. If a shell script offends thee, take a look at the rblsmtpd
> code.
Good point. You mean Russell Nelson's MAPS RBL patch, I think. I've
overlooked it. This is similar to the environment variable approach I
also considered.
Thanks for your help.
Best regards,
Dirk
>Russ Allbery <[EMAIL PROTECTED]> writes:
>> Bill Luckett <[EMAIL PROTECTED]> writes:
>
>>> Can anyone point me to some documentation on how to use fixcr. I've
>>> read that "you can simply run sh -c 'fixcr | qmail-smtpd' for your
>>> outgoing mail relay." but where do you put that command? In the startup
>>> script?
---
---
---
>Sorry to follow up to myself, but I should add that while I haven't been
>following the fixcr discussion that closely (it's not a problem that I
>have to deal with), the note "outgoing mail relay" seems to me to imply
>that you *don't* want to be running fixcr on arbitrary *incoming* e-mail.
>So you may want to only do this on the qmail-smtpd you're running on a
>special relay host or on a different port than smtp.
Is this so? Should fixcr not be used on incoming mail? We attach a lot of
UNIX text files--I sure hate to have to complicate my setup if it's not
necessary.
*************************************************************
Bill Luckett
Director of Information Systems
Phi Theta Kappa International Honor Society
Center for Excellence
Mississippi Education and Research Center
1625 Eastover Drive
Jackson, MS 39211
[EMAIL PROTECTED]
601.957.2241 ext 559
*************************************************************
Hello
I've setup a qmail to handle my virtual domain, and I want fastforward to
handle all mail for my virtual domain.
The virtual domain is netdata.nu and the hostname of my server is
linux.nystromska.soderkoping.se.
I got a unix account 'niklas' an that server and i put [EMAIL PROTECTED]:
niklas in my /etc/aliases. printforwards prints niklas@netdata
&[EMAIL PROTECTED] But when sending mail to
[EMAIL PROTECTED] I just got this in return
<[EMAIL PROTECTED]>:
Sorry, no mailbox here by that name. (#5.1.1)
When I put @netdata.nu: niklas in /etc/aliases it works, but I don't want
all email to my account.
Whats wrong?
from the log:
922391328.483286 info msg 26216: bytes 282 from
<[EMAIL PROTECTED]$
922391328.587358 starting delivery 7: msg 26216 to local
[EMAIL PROTECTED]
922391328.587381 status: local 1/10 remote 0/20
922391328.606761 delivery 7: failure:
Sorry,_no_mailbox_here_by_that_name._(#5.1.1$
922391328.606787 status: local 0/10 remote 0/20
922391328.617924 bounce msg 26216 qp 12958
922391328.618120 end msg 26216
922391328.618593 new msg 26218
922391328.618755 info msg 26218: bytes 892 from <> qp 12958 uid 86
922391328.623631 starting delivery 8: msg 26218 to local
[EMAIL PROTECTED]$
922391328.623842 status: local 1/10 remote 0/20
922391328.634741 delivery 8: success: did_1+0+0/
922391328.635019 status: local 0/10 remote 0/20
922391328.635362 end msg 26218
922392114.379017 status: exiting
BTW the domain netdata.nu ain't completly registret yet, but I've added it
to my DNS.
- - Niklas
What is the best way to use qmail-inject to send out a broadcast? Is
there a limit to the bcc's one can put on the end of the message header?
I have a list of about 2,200 people I e-mail a weekly newswire to. I am
looking at setting up ezmlm as a potential method to handle this, but
that will take a long time as it is quite complicated (requiring
interactions with an mysql database, specialized bounce handling [I
can't automatically have paying members unsubscribed], etc.).
I need a method to output my newswire (normally a 10k file) in rather
simple fashion. The current method, using a program called bulk_mail,
has been generating more and more failed e-mail ("503 Need MAIL before
RCPT" and other errors).
I've scanned the mailing list archives but seen little to suggest a
solution.
Thank you,
Jeff Hill
P.S. And thanks again to those who solved my double-carriage return
problem.
--
********* HR On-Line: The Network for Workplace Issues ********
** Ph:416-604-7251 -- Fax:416-604-4708 ** http://www.hronline.com **
At 04:26 PM Thursday 3/25/99, Jeff Hill wrote:
>What is the best way to use qmail-inject to send out a broadcast? Is
>there a limit to the bcc's one can put on the end of the message header?
Not practically.
>I have a list of about 2,200 people I e-mail a weekly newswire to. I am
Certainly 2K bcc: recipients is no problem at all in one qmail-inject
invocation. I've done 40K recipients, no problems.
>looking at setting up ezmlm as a potential method to handle this, but
>that will take a long time as it is quite complicated (requiring
>interactions with an mysql database, specialized bounce handling [I
>can't automatically have paying members unsubscribed], etc.).
If you have a reliable method for handling the bounces to update your
recipient list, then ezmlm may not buy you much. The way that ezmlm injects
mail is not significantly faster than multiple bcc's to qmail-inject. So if
you're looking at ezmlm solely from that perspective, I wouldn't bother.
Regards.
From: Mark Delany <[EMAIL PROTECTED]>
> If you have a reliable method for handling the bounces to update your
> recipient list, then ezmlm may not buy you much. The way that ezmlm injects
> mail is not significantly faster than multiple bcc's to qmail-inject. So if
> you're looking at ezmlm solely from that perspective, I wouldn't bother.
One small thing.. ezmlm does not allow BCC's. It will bounce the message
unless the list address is in To: or CC:. I think this is a spam prevention
feature. (?)
--Adam
Greg Moeller writes:
> Now, what happened last night was that qmail-send got stuck.
qmail-send creates a detailed log of its actions. I presume that you
followed the installation instructions, so you already know where to
find the log. What does the log say?
---Dan
Matthias Pigulla writes:
> So what about introducing a new qmail feature that allows to control the
> use of pipe commands?
This feature was introduced in qmail 0.73. One way to set it up is
drwxr-x--- root joegroup /home/joe
-rwxrw---- root joegroup /home/joe/.qmail
-rw-r----- root joegroup /home/joe/.qmail-direct
with joe-direct in .qmail and ./Mailbox in .qmail-direct. Then Joe can
change .qmail, but the x bit on .qmail will stop anything other than
forwarding.
---Dan
Steven Levis writes:
> Does anyone know at one point these were implemented?
Both tcpto and the quadratic retry schedule appeared in qmail 0.74. The
2-minute restriction on tcpto appeared in qmail 0.93. What are you
trying to change?
---Dan
D. J. Bernstein wrote:
>Steven Levis writes:
>> Does anyone know at one point these were implemented?
>
>Both tcpto and the quadratic retry schedule appeared in qmail 0.74. The
>2-minute restriction on tcpto appeared in qmail 0.93. What are you
>trying to change?
>
>---Dan
Trying to change the behavior of the retry schedule to handle hosts
which cause connections to hang on most, but not all, incoming SMTP
conversations.
The few connections that are accepted by the remote host seem to
clear the host-is-dead flag. At this point, all allocated qmail-remotes
get caught up retrying connections to this host.
Thanks. I may change the flag to be a counter of some sort, or some
other modification that will impact qmail's behavior the least.
===
Steven Levis
_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com
Hi.
A few weeks back I heard someone saying that qmail 2 will have a completely
different queue structure (the thread was about file names being i-node
numbers).
so:
- when will qmail 2 be released?
- what other changes will it incorporate?
- any other info about it?
- how do people who know that the queue will have a different structure
know about it? Or, if you can answer the previous questions, when did you
get the info?
Thanks.
David.
David Villeger wrote/schrieb/scribsit:
> - when will qmail 2 be released?
When it's finished.
> - any other info about it?
I guess it's going to be pretty cool! :-)
> - how do people who know that the queue will have a different structure
> know about it?
On his pages, Dan states that qmail 2 will have a zeroseek queue, IIRC.
There was some discussion about that here not so long ago and there's
[EMAIL PROTECTED] (with no traffic currently).
Stefan
Tim Pierce writes:
> My vacation program replied, and apparently that
> was enough to confirm my subscription.
A bounce is not a reply. All bouncers, including vacation programs,
should return messages to the envelope sender. The envelope sender for
an ezmlm confirmation request is a trash address.
---Dan
Hi there,
I know it's a little bit off the topic of this list. Please accept my
apology here. I just try to find out anybody out there has ever tried this
combination before.
Here is the script to start the daemon:
#! /bin/sh
ROG=sshd
LKDIR=/var/lock/sshd
mkdirs() {
[ -d $LKDIR ] || ( mkdir -p $LKDIR );
}
start () {
mkdirs
echo -n "Starting $PROG..."
supervise $LKDIR /usr/local/sbin/$PROG &
echo "done"
}
But the supervise somehow thinks the sshd is dead and keeps starting
it, giving out the following messages continuously:
Starting sshd...FATAL: Creating listener failed: port 22 probably
already in use!
....
The sshd program is downloaded and compiled from
http://www.ssh.fi/sshprotocols2/index.html
Thanks a lot in advanced.
--George Hong
System Admin
Dakotacom.net
Karl Lellman writes:
> I have talked to the firewall manufacturer about this, but they say that
> there is nothing they can do.
The firewall is converting TCP connection timeouts into successful TCP
connections with no data. That's incorrect behavior.
> Does anyone have any ideas or ways to get around this?
Connect directly to the remote host, rather than the firewall. Fix the
firewall.
---Dan
D. J. Bernstein writes:
> > I have talked to the firewall manufacturer about this, but they say that
> > there is nothing they can do.
>
> The firewall is converting TCP connection timeouts into successful TCP
> connections with no data. That's incorrect behavior.
>
> > Does anyone have any ideas or ways to get around this?
>
> Connect directly to the remote host, rather than the firewall. Fix the
> firewall.
This question actually comes up frequently on various technical mailing
lists and newsgroups. Unfortunately, there is no way within the POSIX
and BSD sockets API/SPI to make a program that can do this. Once a new
connection comes in, the program basically has to accept(), or else stop
selecting the listening fd for read, or close the listening fd. There
is no way to just ignore the connection without being constantly disturbed
but it being pending.
The problem is in the socket API design (and it's not the only problem
it has). You can prove me wrong with a program that _can_ do "the right
thing" with portable standard interface calls.
I do agree that it's incorrect behaviour. But since I don't see any fix
for it forthcoming, I think we just need to deal with the reality that
an established connection that summarily disconnects with no data is
effectively equivalent to connections that are refused and connections
that timeout. For SMTP that means the mail didn't go through.
Oh, I'd absolutely LOVE to be able to detect where a connection is coming
from even before I accept() it, and have a new API function to discard
the SYN so I can make it timeout, or maybe a new refuse() function so it
can force a connection refused scenario and pretend no one was listening
(basically to emulate that the daemon was not listening for that source).
I'd love to do that to known spammers and relays.
--
Phil Howard | [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED]
phil | [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED]
at | [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED]
ipal | [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED]
dot | [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED]
net | [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED]
Robert J. Adams writes:
> Is it possible to setup AutoTURN so that the user doesn't have to have a
> static IP?
ETRN and AutoTURN are designed for static IP addresses.
If you have a program that reliably maps dynamic IP addresses to
accounts, you can use that program with AutoTURN; but this is usually
difficult when someone else is in charge of the dynamic IP assignment.
---Dan
Hello
I had running qmail on an machine. This machine is now changed. I
copied all qmail stuff from the old to the new machine. And I checked
all permissions. All may run very well - but now incoming mails have
a delay up to an hour befor they are delivered to there mailbox.
Where can I search for this error? Someone an idea?
Sorry for my bad english.
Sigi.
Siegfried Kerkow wrote/schrieb/scribsit:
> all permissions. All may run very well - but now incoming mails have
> a delay up to an hour befor they are delivered to there mailbox.
Check /var/qmail/queue/lock/trigger. It should be
prw--w--w- 1 qmails qmail 0 Mar 25 23:03 trigger
Stefan
>
>Siegfried Kerkow wrote/schrieb/scribsit:
>>all permissions. All may run very well - but now incoming mails have
>>a delay up to an hour befor they are delivered to there mailbox.
>
>Check /var/qmail/queue/lock/trigger. It should be
>prw--w--w- 1 qmails qmail 0 Mar 25 23:03 trigger
Hmmm, this would be it. But I have no "trigger" file!
This I could not copy over from the old machine and I thought qmail
makes this new. But didn't look after.
How do I make this file?
Sigi.
>Stefan
>
>
On Fri, Mar 26, 1999 at 09:28:38AM +0100, Siegfried Kerkow wrote:
> >Check /var/qmail/queue/lock/trigger. It should be
> >prw--w--w- 1 qmails qmail 0 Mar 25 23:03 trigger
>
> Hmmm, this would be it. But I have no "trigger" file!
> This I could not copy over from the old machine and I thought qmail
> makes this new. But didn't look after.
>
> How do I make this file?
# cd /var/qmail/queue/lock
# mkfifo trigger
# chown qmails trigger; chgrp qmail trigger; chmod 622 trigger
--
System Administrator
See complete headers for address, homepage and phone numbers
Siegfried Kerkow wrote:
> >
> >Siegfried Kerkow wrote/schrieb/scribsit:
> >>all permissions. All may run very well - but now incoming mails have
> >>a delay up to an hour befor they are delivered to there mailbox.
> >
> >Check /var/qmail/queue/lock/trigger. It should be
> >prw--w--w- 1 qmails qmail 0 Mar 25 23:03 trigger
>
> Hmmm, this would be it. But I have no "trigger" file!
> This I could not copy over from the old machine and I thought qmail
> makes this new. But didn't look after.
>
> How do I make this file?
>
It might be worth checking the time on the machines involved. You may
find that there is an hour's time difference between your new qmail
machine and the others. This might be the problem.
Les.
>On Fri, Mar 26, 1999 at 09:28:38AM +0100, Siegfried Kerkow wrote:
>
>> >Check /var/qmail/queue/lock/trigger. It should be
>> >prw--w--w- 1 qmails qmail 0 Mar 25 23:03 trigger
>>
>> Hmmm, this would be it. But I have no "trigger" file!
>> This I could not copy over from the old machine and I thought qmail
>> makes this new. But didn't look after.
>>
>> How do I make this file?
>
># cd /var/qmail/queue/lock
># mkfifo trigger
># chown qmails trigger; chgrp qmail trigger; chmod 622 trigger
Many thanks for your help - it work's.
Sigi.
>--
>System Administrator
>See complete headers for address, homepage and phone numbers
>
>Siegfried Kerkow wrote:
>
>> >
>> >Siegfried Kerkow wrote/schrieb/scribsit:
>> >>all permissions. All may run very well - but now incoming mails
>have
>> >>a delay up to an hour befor they are delivered to there mailbox.
>> >
>> >Check /var/qmail/queue/lock/trigger. It should be
>> >prw--w--w- 1 qmails qmail 0 Mar 25 23:03 trigger
>>
>> Hmmm, this would be it. But I have no "trigger" file!
>> This I could not copy over from the old machine and I thought qmail
>> makes this new. But didn't look after.
>>
>> How do I make this file?
>>
>
>It might be worth checking the time on the machines involved. You may
>find that there is an hour's time difference between your new qmail
>machine and the others. This might be the problem.
Dear Les, you didn't read correct or my english is so bad. ;-)
All is running on the new server! From incoming mail to delivering to
the mailbox there is this long time. And second, all my machines are
synchronized with the atomic clock (germany) for all same time. ;-)
Sigi.
>Les.
>
>
>
Chris Green writes:
> I have created a file .qmail-maxine.green in /var/qmail/alias with
> just 'maxine' in it but that doesn't seem to work.
There's a big warning about this in the dot-qmail man page. There's also
ftp://koobera.math.uic.edu/www/qmail/faq/incominguser.html#alias-dots.
> It also seems a
> rather clumsy mechanism if I wanted to create a lot of aliases.
There are hooks for external alias mechanisms. You can use fastforward,
for example, if you prefer the /etc/aliases format.
> Should I be using
> the /var/qmail/users/assign file describe in qmail-users (but I don't
> have one at the moment).
You could set one up by putting
maxine:maxine:maxine.green:
into users/mailnames, running qmail-pw2u, and running qmail-newu. You
could then remove your alias. This would make maxine.green-* addresses
work the same way as maxine-* addresses.
> What I would really like to do is arrange that all mail for (regex)
> '.*maxine.*@isbd.demon.co.uk' would go to maxine, is there a simple
> way to do this?
You could insert something like
|case "_$LOCAL_" in _*maxine*_) forward maxine; esac
at the top of your existing ~alias/.qmail-default.
---Dan
Text written by Adam D. McKenna (and some other folks) at 07:31 PM 3/25/99
-0500:
>
>> >Sendmail is not a Unix program. It is an NT program that someone
>> >ported to Unix twenty years ago.
>>
>> Hmm...I can't quite see the humor in this, though it strikes me as
>> a bit funny, and not entirely untrue.
>>
>> But, I'll ask anyway: did you really intend to suggest that Sendmail
>> was originally written for Windows NT over twenty years ago?
>
>Yes. Sendmail was written for NT 0.01 pre-beta in 1979. It was ported to
>UNIX by Bill Gates in 1981.
Wow. You say that with such a straight face -- no smileys, nada. I'd
normally consider it deadpan humor and wind up on the floor laughing.
Unfortunately, MS' various departments and spokesthings have foisted
nonsense of a similar level of ludicrousness on the public too many times.
They say, in their deadpan way, "NT is at least as robust and scalable as
Linux", and "We integrated IE 4 with the operating system to benefit
consumers" and "We are winning our court case with the DoJ", and you can
only laugh uproariously at such things so many times before it sinks in:
these whackos are *serious!*
Then it stops being funny. It gets kind of ominous, in a Big-Brother-esque
kind of way, and that starts to infect your (or at least, my) view of
similar statements.
In fact, seeing such a thing on the Qmail list would probably be a bit
creepy, if I weren't sure you know better. :)
"Bill Gates? Isn't he the guy who invented the Internet, back in 1995?"
<shudder>
-----------------------------------------------------------------
Kai MacTane
System Administrator
Online Partners.com, Inc.
-----------------------------------------------------------------
>From the Jargon File: (v4.0.0, 25 Jul 1996)
hungus /huhng'g*s/ /adj./
[perhaps related to slang `humongous'] Large, unwieldy, usually unman-
ageable. "TCP is a hungus piece of code."
>Yes. Sendmail was written for NT 0.01 pre-beta in 1979. It was ported to
>UNIX by Bill Gates in 1981.
Okay, so it *was* intended to be a joke. It didn't work for me, because
I have this notion (not sure how valid) that NT is supposed to be somebody's
effort to create for Microsoft an OS along the lines of VAX/VMS (I think
the person came out of Digital, perhaps an ex-VMS engineer, and I know
I have his name somewhere in my archives of Internet memorabilia somewhere).
So, when I see someone saying "it was written for NT a long time ago",
I don't see it as obviously funny -- it might just be somebody
misremembering VAX/VMS as Windows NT.
(It's been explained to me, in private email, that Sendmail was written
for BSD. I'd always assumed it was originally targeted for some
variant of UNIX, but was *almost* gullible enough here to be convinced
it was originally written for VAX/VMS. Which is, I suppose, kind of
funny anyway.)
>"Bill Gates? Isn't he the guy who invented the Internet, back in 1995?"
No, that was our (US) Vice President, Algore.
tq vm, (burley)
>> On Thu, 25 Mar 1999 16:58:58 -0800,
>> Kai MacTane <[EMAIL PROTECTED]> said:
K> "Bill Gates? Isn't he the guy who invented the Internet, back in 1995?"
Nah, that was Al Gore.
K> <shudder>
Ditto.
--
Karl Vogel
ASC/YCOA, Wright-Patterson AFB, OH 45433, USA
[EMAIL PROTECTED] or [EMAIL PROTECTED]
Dijon vu - the same mustard as before. --bumper sticker
> K> "Bill Gates? Isn't he the guy who invented the Internet, back in 1995?"
> Nah, that was Al Gore.
Oh, ya'll want me to rant about mickeysoft?
just give the word...
Scott
From: Kai MacTane <[EMAIL PROTECTED]>
> Text written by Adam D. McKenna (and some other folks) at 07:31 PM 3/25/99
> -0500:
> >
> >> >Sendmail is not a Unix program. It is an NT program that someone
> >> >ported to Unix twenty years ago.
> >>
> >> Hmm...I can't quite see the humor in this, though it strikes me as
> >> a bit funny, and not entirely untrue.
> >>
> >> But, I'll ask anyway: did you really intend to suggest that Sendmail
> >> was originally written for Windows NT over twenty years ago?
> >
> >Yes. Sendmail was written for NT 0.01 pre-beta in 1979. It was ported to
> >UNIX by Bill Gates in 1981.
>
> Wow. You say that with such a straight face -- no smileys, nada. I'd
> normally consider it deadpan humor and wind up on the floor laughing.
>
> Unfortunately, MS' various departments and spokesthings have foisted
> nonsense of a similar level of ludicrousness on the public too many times.
> They say, in their deadpan way, "NT is at least as robust and scalable as
> Linux", and "We integrated IE 4 with the operating system to benefit
> consumers" and "We are winning our court case with the DoJ", and you can
> only laugh uproariously at such things so many times before it sinks in:
> these whackos are *serious!*
>
> Then it stops being funny. It gets kind of ominous, in a Big-Brother-esque
> kind of way, and that starts to infect your (or at least, my) view of
> similar statements.
>
> In fact, seeing such a thing on the Qmail list would probably be a bit
> creepy, if I weren't sure you know better. :)
>
> "Bill Gates? Isn't he the guy who invented the Internet, back in 1995?"
> <shudder>
Would you believe me if I told you that you grasped the exact sentiment that I
was trying to convey with that joke?
That was pretty... surreal :)
--Adam
"Adam D. McKenna" wrote:
>
> From: Kai MacTane <[EMAIL PROTECTED]>
> > Unfortunately, MS' various departments and spokesthings have foisted
> > nonsense of a similar level of ludicrousness on the public too many times.
> > They say, in their deadpan way, "NT is at least as robust and scalable as
> > Linux", and "We integrated IE 4 with the operating system to benefit
> > consumers" and "We are winning our court case with the DoJ", and you can
> > only laugh uproariously at such things so many times before it sinks in:
> > these whackos are *serious!*
> >
> > Then it stops being funny. It gets kind of ominous, in a Big-Brother-esque
> > kind of way, and that starts to infect your (or at least, my) view of
> > similar statements.
> >
> > In fact, seeing such a thing on the Qmail list would probably be a bit
> > creepy, if I weren't sure you know better. :)
> >
> > "Bill Gates? Isn't he the guy who invented the Internet, back in 1995?"
> > <shudder>
>
> Would you believe me if I told you that you grasped the exact sentiment that I
> was trying to convey with that joke?
>
> That was pretty... surreal :)
>
> --Adam
Sorry to continue this completely off-topic thread, but... I was
speaking to someone about Microsoft and he said that Microsoft invented
OO (object oriented programming). I mentioned to him that a statement of
such absuridty would put some people in a fury--he then continued to
defend the position--he was serious.
One thing Microsoft has contributed to science (social science, anyway).
It has proved that brainwashing on a massive scale is possible.
--
Joel Shellman
knOcean Interactive Corporation
http://corp.knOcean.com/
[EMAIL PROTECTED] wrote:
>
> >Yes. Sendmail was written for NT 0.01 pre-beta in 1979. It was ported to
> >UNIX by Bill Gates in 1981.
>
> Okay, so it *was* intended to be a joke. It didn't work for me, because
> I have this notion (not sure how valid) that NT is supposed to be somebody's
> effort to create for Microsoft an OS along the lines of VAX/VMS (I think
> the person came out of Digital, perhaps an ex-VMS engineer, and I know
> I have his name somewhere in my archives of Internet memorabilia somewhere).
>
Do you think about David Catler (SP?)
He was the man standing behind the development of VMS
[-snip-]
Uwe A. P. Wuerdinger
--
"One World, One Web, One Program" - Microsoft Promotional Ad
"Ein Volk, Ein Reich, Ein Fuehrer" - Adolf H.
Telefon: +49.8031/3 89 59-0
Telefax: +49.8031/3 89 59-19
WWW: http://www.connect-gmbh.de
Hi.
Does anyone have any info/measurements regarding lsmtp (from LSoft)
compared to qmail?
I realize they work on different OSes. Assuming equivalent hardware, is
lsmtp faster?
Let's say I send one email to 10000 recipients. Let's say 40% go to AOL.
AFAIK qmail will connect to AOL mail server 4000 times (4000
qmail-remote's, 4000 DNS requests, 4000 SMTP connections, ...).
Will lsmtp do the same or will it connect once to a AOL mail server and
inject the 4000 recipients?
Thanks.
David.
On Thu, 25 Mar 1999, David Villeger wrote:
> Hi.
>
> Does anyone have any info/measurements regarding lsmtp (from LSoft)
> compared to qmail?
>
> I realize they work on different OSes. Assuming equivalent hardware, is
> lsmtp faster?
>
> Let's say I send one email to 10000 recipients. Let's say 40% go to AOL.
>
> AFAIK qmail will connect to AOL mail server 4000 times (4000
> qmail-remote's, 4000 DNS requests, 4000 SMTP connections, ...).
>
> Will lsmtp do the same or will it connect once to a AOL mail server and
> inject the 4000 recipients?
LSMTPd will run on Solaris also, so not they can run on
identical hardware.
Qmail will indeed deliver 4000 copies seperately. This is the
only way to do VERP. (The DNS requests should be cached anyways). So,
the choice is:
1) Send one copy w/ LSMTPd, and save a little bit of time in
transferring the emails, and then spend alot of time manually
removing addresses that bounce from your list.
- OR -
2) Send 4000 copies w/ qmail, and let the system automatically remove
bouncers for you via VERP.
-Dustin
I installed daemontools and started to use cyclog, but such messages
started to pop up every few seconds.
cyclog: warning: unable to create @00000922434195, pausing: access
denied
cyclog: warning: unable to create @00000922434255, pausing: access
denied
cyclog: warning: unable to create @00000922434315, pausing: access
denied
...
I invoked cyclog (from Adam's howto) as
setuser qmaill cyclog -s5000000 -n5 /var/log/qmail/qmail-smtpd &
Anyone has seen such things, and how to fix it?
Thanks.
dp
Slightly unrelated (maybe), you should never use this
while ($line = <STDIN>)
but instead do
while (defined($line = <STDIN>))
because one day you'll be bitten by <STDIN> returning a single "0"
and you'll be dead.
perl -w should warn you about this BTW.
Florent
