qmail Digest 20 Oct 1999 10:00:01 -0000 Issue 795
Topics (messages 31832 through 31870):
tcprules error
31832 by: Edward Castillo-Jakosalem
31833 by: Anand Buddhdev
qmail in SCO
31834 by: Luis Bezerra
Re: Virtual domain problems
31835 by: Magnus Bodin
Re: Help! allowed rcpthosts
31836 by: Dave Sill
31863 by: Assela Pathirana
Re: ezmlm mailing lists
31837 by: Dave Sill
cyclog and date stamp problem
31838 by: Marcin Jaskowiak
31839 by: Russell Nelson
31840 by: Marcin Jaskowiak
31841 by: Russell Nelson
Winsock Failures Under RedHat 6.1
31842 by: Eden
31846 by: John K. Chester
31865 by: Sam
Re: mail appliance
31843 by: James Smallacombe
31844 by: James Smallacombe
31845 by: Russell Nelson
31856 by: Jon Rust
Re: qmail delivering in "waves"
31847 by: Markus Stumpf
Concurrencyremote for a specific host
31848 by: Stan Horwitz
31849 by: Ken Jones
31851 by: Markus Stumpf
31852 by: Charles Cazabon
31853 by: Dave Sill
31854 by: Markus Stumpf
31857 by: Giles Lean
31858 by: John R. Levine
31866 by: Russell Nelson
How to install amavis over qmail
31850 by: Albert Hopkins
VirtualSub-Domains?
31855 by: Randall Moseley
31870 by: Magnus Bodin
qmail, Linux, and NetApp/NFS
31859 by: Arnold, Josh A.
31860 by: Rani Assaf
31864 by: Russell Nelson
Solaris boot script.
31861 by: eric
31867 by: Magnus Bodin
Qmail Flooding a link.
31862 by: Mark Parker
Odd bouncing loop
31868 by: zip
Elaborative: percenthack
31869 by: Magnus Bodin
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 to all!
I just changed my tcp.smtp file from ip-based to domain-based. However
upon compiling, an error occurs. I think it has something to do with a
hyphen (-) in one of the domains. Is this allowed or is there an option
for this?
I would also like to know if my command is right:
# tcprules tcp.smtp.cdb tcp.smtp.tmp <tcp.smtp.new
Lastly, I don't have to restart qmail after compiling the cdb file ,
right?
Thanks once again and more power!
--
Edward Castillo-Jakosalem
Systems Administrator
On Tue, Oct 19, 1999 at 07:39:07PM +0800, Edward Castillo-Jakosalem wrote:
> Hi to all!
> I just changed my tcp.smtp file from ip-based to domain-based. However
tcpserver only allows IP-based control, not domain based.
> upon compiling, an error occurs. I think it has something to do with a
> hyphen (-) in one of the domains. Is this allowed or is there an option
> for this?
> I would also like to know if my command is right:
> # tcprules tcp.smtp.cdb tcp.smtp.tmp <tcp.smtp.new
That command is right.
> Lastly, I don't have to restart qmail after compiling the cdb file ,
> right?
tcp.smtp.cdb is read by tcpserver, so, no, you don't need to restart
qmail.
--
See complete headers for more info
Hello everybody,
Anyone knows qmail running in SCO UNIX?
--
-----------------------------
Lu�s Bezerra de A. Junior
[EMAIL PROTECTED]
SecrelNet Inform�tica LTDA
Fortaleza - Cear� - Brasil
Fone: 021852882090
-----------------------------
On Tue, Oct 19, 1999 at 09:00:13AM +0200, Magnus Bodin wrote:
> The fastforward-alternative
> ===========================
> This is the best alternative for a bigger virtual domain.
>
> Put "nowhere.mil:alias-nowhere" in virtualdomains
> and install the fastforward package
> <http://www.qmail.org/koobera/www/fastforward.html>.
>
> Then put every alias for nowhere.mil in /etc/aliases
> like this:
>
> [EMAIL PROTECTED]:[EMAIL PROTECTED]
> [EMAIL PROTECTED]:[EMAIL PROTECTED]
>
>
> Finally put the line
>
> | fastforward -d /etc/aliases.cdb
>
> in ~alias/.qmail-nowhere
That should have been ~alias/.qmail-nowhere-default
All previous (corrected) virtual domain cookbook examples is here, and will
be summarized in a bette form as soon as possible:
http://x42.com/qmail/doc/vdomains1.txt
http://x42.com/qmail/doc/vdomains2.txt
http://x42.com/qmail/doc/vdomains3.txt
--
magnus
-- MOST useless 1998 * http://x42.com/
Assela Pathirana <[EMAIL PROTECTED]> wrote:
>553 sorry, that domain isn't in my list of allowed rcpthosts (#5.7.1)
See:
http://Web.InfoAve.Net/~dsill/lwq.html#relaying
-Dave
Hi,
I got it fixed. Thanks for the information.
> >553 sorry, that domain isn't in my list of allowed rcpthosts (#5.7.1)
>
> See:
>
> http://Web.InfoAve.Net/~dsill/lwq.html#relaying
Yes, it gave me all the required answers and more!
Thanks
Assela
"Haifeng Guo" <[EMAIL PROTECTED]> wrote:
>I have install qmail and ezmlm+ezmlm-idx on my server,I have two server for
>mailing list,one is m1.a.com and the other is m2.a.com, and I want the
>mailing lists on both of the server can use the name a.com for sub ,unsub
>and send mail(for example: on the m1.a.com has a mailing list vb and on the
>m2.a.com has a mailing list vb ,I wish both of them can use [EMAIL PROTECTED] and
>[EMAIL PROTECTED] ,how to setup this,just like egroups.com),thanks.
Are the "vb" lists on m1 and m2 the same? It's not clear what results
you're trying to achieve.
-Dave
Hi,
Does anyone know how to make cyclog to put his logging activity time/date
stamp into log file?
I'm running qmail with above command :
"/usr/local/bin/accustamp \
| /usr/local/bin/setuser qmaill /usr/local/bin/cyclog /var/log/qmail &"
and the contents of log files are something like:
"940297677.122049 status: local 0/10 remote 0/20
940298143.250819 new msg 768173
940298143.250828 info msg 768173: bytes 627 from <[EMAIL PROTECTED]> qp 16
940298143.254849 starting delivery 1: msg 768173 to local
[EMAIL PROTECTED] status: local 1/10 remote 0/20".
So, i was wondering if it's possible to put a date info before query info
(just like in system logs).
I was searching for some command line parameters for cyclog but there is
no suitable ones, and for now on the only way is propably to modify
cyclog's source code.
But before that i just wanna know if anyone had a similar problems?
Hookahey!
Marcin Jaskowiak
Marcin Jaskowiak writes:
> Hi,
>
> Does anyone know how to make cyclog to put his logging activity time/date
> stamp into log file?
> 940298143.250819 new msg 768173
That *is* the time/date stamp. Filter the log file through tailocal
to see the time in the format you expect.
--
-russ nelson <[EMAIL PROTECTED]> http://russnelson.com
Crynwr sells support for free software | PGPok | Government schools are so
521 Pleasant Valley Rd. | +1 315 268 1925 voice | bad that any rank amateur
Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | can outdo them. Homeschool!
Russell Nelson wrote:
> That *is* the time/date stamp. Filter the log file through tailocal
> to see the time in the format you expect.
Thanks. It work.
Anyway, is that possible to log user name when someone is connecting via
pop3d (e.g. the name from 'USER <name>' string)?
Hookahey!
Marcin Jaskowiak
Marcin Jaskowiak writes:
> Anyway, is that possible to log user name when someone is connecting via
> pop3d (e.g. the name from 'USER <name>' string)?
Everything is possible with qmail. The easiest way to do that -- not
necessarily the best -- is to create a shell script, say, my-pop3d,
that looks like this:
#!/bin/sh
echo $USER | logger
/var/qmail/bin/qmail-pop3d
Run it in place of qmail-pop3d.
--
-russ nelson <[EMAIL PROTECTED]> http://russnelson.com
Crynwr sells support for free software | PGPok | Government schools are so
521 Pleasant Valley Rd. | +1 315 268 1925 voice | bad that any rank amateur
Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | can outdo them. Homeschool!
> Allright, I apologize for ahead of time to anyone I piss off with
this, but I am about to lose it. I have
> local mail working just fine with qmail, and am able to send mail to
other hosts from the server as well.
> However I am unable to recieve mail from remote clients. Every time I
attempt to send mail from another site
> I receive a Winsock Error:
Connection to eden.mine.nu. with Ip Address 24.93.40.80 failed from
mail.austin.rr.com-10.1.1.54
[EMAIL PROTECTED]
Server received Winsock error Connection timed out.
Running Redhat 6.1 i386, with sendmail/binmail removed. Would
appreciate any help, even flaming if it's contructive.
Thanks
At 06:08 PM 10/19/99 -0500, you wrote:
> > Allright, I apologize for ahead of time to anyone I piss off with
>this, but I am about to lose it. I have
> > local mail working just fine with qmail, and am able to send mail to
>other hosts from the server as well.
> > However I am unable to recieve mail from remote clients. Every time I
>
>attempt to send mail from another site
> > I receive a Winsock Error:
>
>Connection to eden.mine.nu. with Ip Address 24.93.40.80 failed from
>mail.austin.rr.com-10.1.1.54
>[EMAIL PROTECTED]
>Server received Winsock error Connection timed out.
>
>Running Redhat 6.1 i386, with sendmail/binmail removed. Would
>appreciate any help, even flaming if it's contructive.
>
>Thanks
>
>
This is spooky. Perhaps it's getting a bit too close to Halloween.
My qmail server has been in service for about 3 weeks now (installed on
RedHat 6.0 from Memphis RPM's, running on an old 386/40). Everything
works fine except for one server which has great trouble sending mail to my
server. It's running NTMail under Windows NT 4.0. On my end, I see a log
entry in /var/log/qmail-smtpd/, but no entry in /var/log/qmail/. On the
sending end, I see that the connection attempt was terminated by a Winsock
error.
Oh.... The spooky thing? 30 seconds after this email appeared in my inbox,
an email which the problem server has been trying to send me since 20:00
GMT yesterday appeared. Many of the emails it tries to send me do
eventually appear, usually some time the following day.
I wasn't going to send this to the list until I had gathered a bit more
info, but once I saw this post, I thought I'd mention it. Sorry for the
lack of hard data..... but I would like to hear if anyone's heard of a
similar problem.
----------------------------------------------------------------------------
-----
John K. Chester [EMAIL PROTECTED]
phone 212-792-2036 fax 212-253-4290
----------------------------------------------------------------------------
-----
John K. Chester writes:
> works fine except for one server which has great trouble sending mail to my
> server. It's running NTMail under Windows NT 4.0. On my end, I see a log
> entry in /var/log/qmail-smtpd/, but no entry in /var/log/qmail/. On the
> sending end, I see that the connection attempt was terminated by a Winsock
> error.
>
> Oh.... The spooky thing? 30 seconds after this email appeared in my inbox,
> an email which the problem server has been trying to send me since 20:00
> GMT yesterday appeared. Many of the emails it tries to send me do
> eventually appear, usually some time the following day.
Check the DNS and the ident port on the IP address this is coming from.
It's possible that the DNS is cocked up and the ident query stalls long
enough for NT to toss its cookies most of the time, and requeue the message
for another delivery attempt. But, once in a blue moon, the beast might
get distracted by something else, and by the time it finishes, you've timed
out and sent the banner.
--
Sam
Didn't I hear that someone put together a webmid module for qmail?
Anybody know anything about it?
On Mon, 18 Oct 1999, Jon Rust wrote:
> qmailadmin does not handle management of local accounts. It only
> handles management of virtuals. Not too bad really, but that requires
> you to have a pop prefix. Trying to avoid that...
>
> jon
>
> At 8:33 PM +0200 10/18/99, Markus Wuebben wrote:
> >On Sun, 17 Oct 1999, Jon Rust wrote:
> >
> >> I'm trying to build a mail "appliance" that I can install for
> >> customers who know nothing about UNIX and/or qmail. I suppose webmin
> >> will do for adding users, though a bit clumsy. Even so, that still
> >> leaves forwarding and vacation messages out. I'll try to write some
> >> scripts of my own for this purpose, but if someone wants to share,
> >> that would be great. Just looking for some no-frills, perl/shell CGI.
> >>
> >
> >
> >Check this out:
> >
> >
> >http://www.inter7.com/qmailadmin/
> >
> >
> >
> >Later,
> >Markus
>
>
Didn't I hear that someone put together a webmin module for qmail?
Anybody know anything about it? ^
On Mon, 18 Oct 1999, Jon Rust wrote:
> qmailadmin does not handle management of local accounts. It only
> handles management of virtuals. Not too bad really, but that requires
> you to have a pop prefix. Trying to avoid that...
>
> jon
>
> At 8:33 PM +0200 10/18/99, Markus Wuebben wrote:
> >On Sun, 17 Oct 1999, Jon Rust wrote:
> >
> >> I'm trying to build a mail "appliance" that I can install for
> >> customers who know nothing about UNIX and/or qmail. I suppose webmin
> >> will do for adding users, though a bit clumsy. Even so, that still
> >> leaves forwarding and vacation messages out. I'll try to write some
> >> scripts of my own for this purpose, but if someone wants to share,
> >> that would be great. Just looking for some no-frills, perl/shell CGI.
> >>
> >
> >
> >Check this out:
> >
> >
> >http://www.inter7.com/qmailadmin/
> >
> >
> >
> >Later,
> >Markus
>
>
James Smallacombe writes:
>
> Didn't I hear that someone put together a webmin module for qmail?
> Anybody know anything about it?
<li>Vyacheslav Ignatyuk wrote an alpha version of a <a
href="ftp://ftp.ant.ru/pub/qmail/qmail-manager.tar.gz">qmail
manager module</a> for <a
href="http://www.webmin.com">webmin</a>. It's a first alpha version,
so may be unstable.
--
-russ nelson <[EMAIL PROTECTED]> http://russnelson.com
Crynwr sells support for free software | PGPok | Government schools are so
521 Pleasant Valley Rd. | +1 315 268 1925 voice | bad that any rank amateur
Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | can outdo them. Homeschool!
At 2:40 PM -0400 10/19/99, James Smallacombe wrote:
>Didn't I hear that someone put together a webmid module for qmail?
>Anybody know anything about it?
Yes. I've played with it some, but it seems pretty rough. Doesn't
work with my install of vpopmail (3.4.9) for whatever reason.
Ken's recommendation for using qmailadmin for a local domain worked
like a champ. Instant email appliance. Just delivered to the customer
and they were very happy.
jon
I am still at this problem ...
In short: instead of keeping parallel deliveries around maxconcurrency values
qmail "explodes" to the maximum, goes down to about 15-25 deliveries,
stays there for a while and then "explodes" back to the maximum.
On Tue, Oct 05, 1999 at 08:08:37PM +0200, Markus Stumpf wrote:
> To illustrate the behaviour I put a gnuplot plot at
> http://www.lamer.de/maex/creative/software/qmail/deliver-stats.gif
I've spent a few hours debugging qmail-send and doing stress tests
and I *think* I have tracked the problem down:
a) While there are unprocessed messages in the queue, qmail does only few
deliveries (I still don't fully understand why, but I think it's due to
the "management flow").
b) Then comes the time, when all those unprocessed messages are processed,
qmail-send starts as much parallel deliveries as possible.
c) Too bad that few of those deliveries generate bounces. qmail injects
unprocessed messages to the queue and falls back to a)
In my tests I first had problems simulating the behaviour. I'd injected
emails with local "null" deliveries, and qmail always was around maximum
parallel deliveries. Then I did inject messages with two recipients
(one valid null delivery and one address that generated a bounce) and
the problem became evident.
Another test I did was kill -STOP qmail-send with a totally empty queue.
I then injected "good" emails and did a kill -CONT qmail-send. Even
during the "preprocessing phase"qmail used maximum deliveries
immediately and stayed at this value.
So I *think* it's due to the bounce handling.
Does anybody know why qmail behaves like in a) and if there is any way
around it?
Thanks
\Maex
--
SpaceNet GmbH | http://www.Space.Net/ | Yeah, yo mama dresses
Research & Development | mailto:[EMAIL PROTECTED] | you funny and you need
Joseph-Dollinger-Bogen 14 | Tel: +49 (89) 32356-0 | a mouse to delete files
D-80807 Muenchen | Fax: +49 (89) 32356-299 |
Hello fellow qmail users:
In the qmail documentation, there's information about that says I can use
the concurrencyremote control file to set the maximum number of concurrent
remote mail deliveries can occur. I am wondering if this can be set on a
per system basis. Specifically, I want to set the concurrencyremote level
down to a value of 5 for one specific system (because its being overloaded
with mail processing) and while allowing concurrencyremote to be set to a
value of 50 for all other systems that might receive mail from this
particular qmail server. Is this possible? If so, how?
Thanks
It can only be set on a per physical queue basis. If each "system"
uses the same physical queue, then the answer to your question is no.
If the "systems" use different physical queues, then you can do it.
Ken Jones
Stan Horwitz wrote:
>
> Hello fellow qmail users:
>
> In the qmail documentation, there's information about that says I can use
> the concurrencyremote control file to set the maximum number of concurrent
> remote mail deliveries can occur. I am wondering if this can be set on a
> per system basis. Specifically, I want to set the concurrencyremote level
> down to a value of 5 for one specific system (because its being overloaded
> with mail processing) and while allowing concurrencyremote to be set to a
> value of 50 for all other systems that might receive mail from this
> particular qmail server. Is this possible? If so, how?
>
> Thanks
On Tue, Oct 19, 1999 at 03:00:50PM -0400, Stan Horwitz wrote:
> particular qmail server. Is this possible? If so, how?
Take a look at
http://www.illuin.demon.co.uk/qmail/
there is a domain concurrency patch.
However this is from Aug 1997 so it is probably for qmail-1.01 and needs
adaption to qmail-1.03. Be sure to read the README file.
\Maex
--
SpaceNet GmbH | http://www.Space.Net/ | Yeah, yo mama dresses
Research & Development | mailto:[EMAIL PROTECTED] | you funny and you need
Joseph-Dollinger-Bogen 14 | Tel: +49 (89) 32356-0 | a mouse to delete files
D-80807 Muenchen | Fax: +49 (89) 32356-299 |
Stan Horwitz <[EMAIL PROTECTED]> wrote:
> In the qmail documentation, there's information about that says I can use
> the concurrencyremote control file to set the maximum number of concurrent
> remote mail deliveries can occur. I am wondering if this can be set on a
> per system basis. Specifically, I want to set the concurrencyremote level
> down to a value of 5 for one specific system (because its being overloaded
> with mail processing) and while allowing concurrencyremote to be set to a
> value of 50 for all other systems that might receive mail from this
> particular qmail server. Is this possible? If so, how?
It's not really a qmail issue. If a particular mailhost is bogging down
because it is accepting more SMTP connections than it can actually handle,
it's due to:
-poor configuration. Solution: tell the sysadmin.
-poor administration. Solution: tell the sysadmin.
-poor choice of MTA on their end. Solution: tell the sysadmin.
If a site can only handle 5 incoming SMTP connections at a time, their
system should be configured to only accept 5 simultaneous incoming SMTP
connections. A qmail instance attempting to deliver to that system can't
(and shouldn't try) to tell whether the system can actually handle a
connection that it accepts.
Charles
--
----------------------------------------------------
Charles Cazabon <[EMAIL PROTECTED]>
Any opinions expressed are just that -- my opinions.
----------------------------------------------------
Stan Horwitz <[EMAIL PROTECTED]> wrote:
>Specifically, I want to set the concurrencyremote level
>down to a value of 5 for one specific system (because its being overloaded
>with mail processing) and while allowing concurrencyremote to be set to a
>value of 50 for all other systems that might receive mail from this
>particular qmail server. Is this possible? If so, how?
"Everything is possible with qmail."
But "possible" is not always "easy", and this is one case where it's
not.
DJB's philosophy in this case is that a receiving system shouldn't
accept more SMTP connections than it can handle. It makes sense. Each
system has different capabilities, and either the receiver limits what
it accepts, or *every* *possible* sender has to maintain a database of
*every* *possible* recipient's capabilities: clearly impossible and
not a good idea.
However, should you chose to attempt a band-aid for a key remote site,
the approach would be to install a second qmail with a lower
concurrencyremote and redirect messages for the swamped site from the
primary qmail to the secondary.
-Dave
On Tue, Oct 19, 1999 at 01:23:18PM -0600, Charles Cazabon wrote:
> It's not really a qmail issue. If a particular mailhost is bogging down
> because it is accepting more SMTP connections than it can actually handle,
> it's due to:
> -poor configuration. Solution: tell the sysadmin.
> -poor administration. Solution: tell the sysadmin.
> -poor choice of MTA on their end. Solution: tell the sysadmin.
I don't agree.
1) There are mailservers that don't accept more than n concurrent
connections from the same IP (while they still have the capability for
open slots for other IPs).
However this causes them to first "see" the connection and then
dropping them. I wouldn't call that "poor [conf/admin/mta]".
This also leads to 2)
2) If a system is only accepting n connections a time, I have
concurrencyremote set to 4*n and I just received a bulk with 3*n
messages to that host, I will probably get n-m messages through and
2*n messages get blocked by either tcpto and/or the messages become
marked as "unreachable" and go to the backoff list.
Chances would be IMHO good to get them through smoothly if I could
set the max concurrency for that host to e.g. n/2. Then qmail
wouldn't even try to deliver more messages -> less errors, less
backoff delays, more throughput for the other mails in MY queue
(as they wouldn't have to wait for the others producing errors).
But I may be wrong ;-)
\Maex
--
SpaceNet GmbH | http://www.Space.Net/ | Yeah, yo mama dresses
Research & Development | mailto:[EMAIL PROTECTED] | you funny and you need
Joseph-Dollinger-Bogen 14 | Tel: +49 (89) 32356-0 | a mouse to delete files
D-80807 Muenchen | Fax: +49 (89) 32356-299 |
On Tue, 19 Oct 1999 15:39:28 -0400 (EDT) Dave Sill wrote:
> However, should you chose to attempt a band-aid for a key remote site,
> the approach would be to install a second qmail with a lower
> concurrencyremote and redirect messages for the swamped site from the
> primary qmail to the secondary.
Or use serialmail. This limits the concurrency to one and you'll have
to trigger serialmail from time to time, but it is easier I think to
install serialmail for a few band-aided sites than to have qmail
installations multiplying.
Regards,
Giles
>> However, should you chose to attempt a band-aid for a key remote site,
>> the approach would be to install a second qmail with a lower
>> concurrencyremote and redirect messages for the swamped site from the
>> primary qmail to the secondary.
>
>Or use serialmail. This limits the concurrency to one and you'll have
>to trigger serialmail from time to time, but it is easier I think to
>install serialmail for a few band-aided sites than to have qmail
>installations multiplying.
Actually, you can run multiple instances of serialmail out of the same
maildir, so if you want up to four deliveries, do something like this
in your trigger script
if [ there are messages to send ]
then
for i in 1 2 3 4
do
maildirsmtp blah blag ... &
done
wait # for all those serialmails to be done
fi
--
John R. Levine, IECC, POB 727, Trumansburg NY 14886 +1 607 387 6869
[EMAIL PROTECTED], Village Trustee and Sewer Commissioner, http://iecc.com/johnl,
Member, Provisional board, Coalition Against Unsolicited Commercial E-mail
Markus Stumpf writes:
> 1) There are mailservers that don't accept more than n concurrent
> connections from the same IP (while they still have the capability for
> open slots for other IPs).
> However this causes them to first "see" the connection and then
> dropping them. I wouldn't call that "poor [conf/admin/mta]".
Right, that's fine.
> 2) If a system is only accepting n connections a time, I have
> concurrencyremote set to 4*n and I just received a bulk with 3*n
> messages to that host, I will probably get n-m messages through and
> 2*n messages get blocked by either tcpto and/or the messages become
> marked as "unreachable" and go to the backoff list.
Okay, then what you're asking for is for qmail to special-case those
hosts which concurrently have TCP connections open and are not serving
new connections (either refusing connections or closing them
immediately, or giving a 4xx response to the smtp greeting). The
treatment that you want is to immediately retry mail when one of the
existing connections closes.
Or to put it a different way, you want qmail to recognize and remember
those hosts and when one of that host's connections closes, adjust the
retry schedule for the next message bound for that host. However, the
next message very likely will be retried in five minutes. So why go
to all the bother of recognizing and remembering those hosts just to
gain five minutes?
Of course, neither of us are solving the same problem as the original
poster posed. We're presuming that the destination host is
well-administered, and quickly rejects connections it cannot
immediately handle.
--
-russ nelson <[EMAIL PROTECTED]> http://russnelson.com
Crynwr sells support for free software | PGPok | Government schools are so
521 Pleasant Valley Rd. | +1 315 268 1925 voice | bad that any rank amateur
Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | can outdo them. Homeschool!
I have ./configure'd the amavis email virus program which says it supports
qmail, but there is no documentation on actually how to make it work with
qmail (that I can find anyway). Anyone have any experience with this?
--
Albert Hopkins
Sr. Systems Specialist
Dynacare, Inc
[EMAIL PROTECTED]
My goal is to send all incoming mail that qmail receives that match
either
a particular email and/or domains to be handled by an external web-based
mail system.
With just an entry in the rcphosts of 'me.stage.blah.com' and an
entry in the aliases file of '@me.stage.blah.com: |/path/to/mail.pl'
I have gotten it to work once then it stopped calling my external
application. I had my program send back a hard-error(100) when it
failed to match the email to a valid web mail user.
Any ideas?
Sincerely,
-Randall, dumb-founded-and-flustered
On Tue, Oct 19, 1999 at 01:16:46PM -0700, Randall Moseley wrote:
> My goal is to send all incoming mail that qmail receives that match
> either a particular email and/or domains to be handled by an external
> web-based mail system.
>
> With just an entry in the rcphosts of 'me.stage.blah.com' and an
> entry in the aliases file of '@me.stage.blah.com: |/path/to/mail.pl'
What does mail.pl do? Deliver or forward?
I mean, if you forward, I would do it like this:
control/rcphosts:
me.stage.blah.com
control/virtualdomains:
me.stage.blah.com:alias-mestage
~alias/.qmail-mestage-default:
| forward ${LOCAL}@otherdomain.com
Otherwise, put the delivery instruction lines into "~alias/.qmail-mestage-default".
/magnus
Based on everything I've been able to gather, the ultimate solution consists
of multiple qmail servers accessing a NetApp filer via NFS. In the past, I
had heard that Linux's NFS implementation was somewhat buggy. I'm sure that
it has since been made more robust, however I'm curious if anyone else is
using a similar setup on a large scale? I'd be interested in hearing about
any setups that involve multiple servers accessing a NetApp filer. Finally,
with cost being much less important than scalability, performance, etc. what
would typically be the bottleneck in a setup like this? The connection from
the servers to the filer will be gigabit ethernet. If it makes sense, I'll
throw quad xeon/1G ram/etc. on all the servers, but I suspect most of it
would be wasted. Thanks.
--
Josh Arnold
On 20 Oct 1999 01:34:44 +0200, Arnold, Josh A. <[EMAIL PROTECTED]> wrote:
> it has since been made more robust, however I'm curious if anyone else is
> using a similar setup on a large scale? I'd be interested in hearing about
We do... One toaster doing mail (more than 350000 mailboxes) and
serving users web pages via NFS... frontends are PCs (PII/128Mb)
running Linux 2.2.x
The whole thing is very stable (more than 4 months of uptime)... I've
just hacked qmail-getpw.c a little bit to hash on login and distribute
maildirs into multiple trees (so I could add other toasters)...
> would typically be the bottleneck in a setup like this? The connection from
network performance... Our toaster is connected through a gigabit
interface (actually 2)...
Regards,
Rani
Arnold, Josh A. writes:
> Based on everything I've been able to gather, the ultimate solution consists
> of multiple qmail servers accessing a NetApp filer via NFS.
Not completely obvious. It might be more efficient to have a server
dedicated to a subset of the mailboxes, and for the smtp front-ends to
run a variant of qmail-qmqpc to queue the mail on the appropriate
machine based on the recipient. I don't know -- I haven't yet had a
customer who needed that kind of solution who was starting from
scratch. Most of them have found that sendmail simply isn't up to the
task of dealing with more than a few tens of thousands of mailboxes,
and want me to transition them to qmail.
--
-russ nelson <[EMAIL PROTECTED]> http://russnelson.com
Crynwr sells support for free software | PGPok | Government schools are so
521 Pleasant Valley Rd. | +1 315 268 1925 voice | bad that any rank amateur
Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | can outdo them. Homeschool!
What are those of you who are running qmail on a Solaris machine
using for a startup script?
I am getting
Starting mail-transfer agent: qmail JOB/PID
[1] 154
Local delivery agent started...
Network delivery agent started.
setuser: fatal: unable to run tailocal: file does not exist
setuser: fatal: unable to run accustamp: file does not exist
setuser: fatal: unable to run cyclog: file does not exist
on boot. Tried this with /usr/bin/ksh and /bin/sh
Things work fine when manually starting it , even from /sbin/sh
The script follows..
-------------------------------------------------------------
PATH=/usr/local/bin:/usr/bin:/usr/sbin:/bin:/sbin:/var/qmail/bin
# User IDs
USERID=xx # Qmaild UID
GROUPID=xx # Nofiles GID
# Variables
CONCURRENT=40 # TcpServer MaxConns (40
def.)
LOGCOUNT=10 # Log Count
LOGSIZE=5000000 # Log Size
PORT=25 # SMTP Port
QMAILLOGDIR=/var/log/qmail # Qmail Log Directory
QMAILSMTPDLOCKDIR=/var/lock/qmail-smtpd # Qmail Lock Directory
QMAILSMTPDLOGDIR=$QMAILLOGDIR/qmail-smtpd # Qmail Smtpd Log Director
QMAILLOGUSER=qmaill # Qmaill User
QMAILPID=`ps -eo pid,comm | awk '{ if ($2 == "qmail-send") print $1}'` #
PID
TCPRULESD=/etc/tcprules.d # TcpRules Directory
SMTPD='rblsmtpd qmail-smtpd' # (Supervised) SMTP Daemon
# This must remain under TCPRULESD!!
CDB=$TCPRULESD/tcp.smtp.cdb # CDB File
## Now begin the tasks
case "$1" in
start)
echo "Starting mail-transfer agent: qmail JOB/PID "
# Start local delivery
csh -cf '/var/qmail/rc &'
echo "Local delivery agent started..."
# Start network delivery agent
supervise $QMAILSMTPDLOCKDIR \
tcpserver -v -c $CONCURRENT -x $CDB -u $USERID -g $GROUPID 0 $PORT \
$SMTPD 2>&1 | setuser $QMAILLOGUSER accustamp | \
setuser $QMAILLOGUSER tailocal | \
setuser $QMAILLOGUSER cyclog -s $LOGSIZE -n $LOGCOUNT $QMAILSMTPDLOGDIR &
echo "Network delivery agent started."
;;
etc...
On Tue, Oct 19, 1999 at 07:30:42PM -0500, eric wrote:
> What are those of you who are running qmail on a Solaris machine
> using for a startup script?
We are brave?
> The script follows..
> PATH=/usr/local/bin:/usr/bin:/usr/sbin:/bin:/sbin:/var/qmail/bin
Add a this line here:
export PATH
and you will find that the shell will know what e.g. tailocal is..
I otherwise would like to recommend looking into Dave Sill's script
and /var/qmail/rc that actually logs qmail-send into cyclog and not
syslog. You will find it more easy to start and stop qmail-send with
svc (supervise) than with your $QMAILPID.
http://Web.InfoAve.Net/~dsill/lwq.html#start-qmail
Regards,
magnus
--
MOST useless 1998 * http://x42.com/
Hi all,
The other day we came across an interesting scenario with our Qmail Server
(1.03).
The Server has been running smoothly for a month now. It operates purely as
a mail gateway and forwarder to a couple of domains, with no local mail or
POP3 mail. I am using TCPserver with the SMTP daemon.
During the course of the day we found our link to be very slow, and upon
investigation found that the Qmail server had spawned approximately 20 smtpd
processes all delivering mail to the same recipient. This appeared to be
flooding our link.
I looked into the mail log and found that 100 attempts to this recipient had
already been attempted, however none had been successful. Simply Starting
and stopping the Qmail and smtpd process did not work, as once these
processes were restarted the link utilisation went straight back to 100%.
I knew the recipients address so I went and found the file containing this
address in the /<Qmail root>/queue/remote directory and removed it from this
location.
This appeared to resolve the issue as I was then able to restart the Qmail
and smtpd processes without flooding our link.
When I viewed the file I moved from the /<Qmail root>/queue/remote
directory it simply had a single line with the recipients address, no
strange or repeated characters.
After having done this I located the sender of the email and requested they
resend the email while I monitored the system. This time the email passed
through without a problem.
Has anyone come across this sort of behaviour previously, or does anyone
know of any reason or protection from this happening again.
Are there other areas I should be looking at when troubleshooting this sort
of thing??
Out of interest the originating server of the email was and M$ Exchange 5.5
with MailSweeper installed on the SMTP interface.
Thanks
Mark P
Hi,
A Groupwise server (company123.com) that uses my qmail server
(mail.zipcon.net) as a relay just sent 6000+ messages identical to the one
included. It appears to me that the Groupwise server is not properly
handling verification that the bounced message has been delivered, or in
this case, also bounced. Either that, or the the rocketmail server is not
properly bouncing the message.
I tested a rocketmail mail server and found that the error message comes
after the message has been sent. I'm not familiar with the relevant RFCs,
maybe this error message itself is the problem.
Would someone mind taking a quick peek and clue me in if they see
something obvious. I will probably set up a temporary virtualdomain for
rocketmail.com and delivering the message to a local mailbox, but would
very much like to know why its happening. Thanks, Dan
---start of test to rocketmail.com mail server---
Connected to mx1.mail.yahoo.com.
Escape character is '^]'.
220 mta113.yahoomail.com ESMTP
mail from: <[EMAIL PROTECTED]>
250 ok
rcpt to: <[EMAIL PROTECTED]>
250 RCPT ok
data
354 go ahead dd
test
.
552 qdirdel.1 error 100:This user does not have a rocketmail.com account
([EMAIL PROTECTED])
quit
---end of test to rocketmail.com mail server---
Here is one of the bounce messages.
---start of bounce message---
>From [EMAIL PROTECTED] Wed Oct 20 05:35:46 1999 +0000
Status: R
X-Status:
X-Keywords:
Return-Path: <#@[]>
Delivered-To: [EMAIL PROTECTED]
Received: (qmail 21045 invoked from network); 20 Oct 1999 03:05:37 -0000
Received: from zipcon.net ([EMAIL PROTECTED])
by ran.zipcon.net with SMTP; 20 Oct 1999 03:05:37 -0000
Received: (qmail 11055 invoked by alias); 20 Oct 1999 02:59:31 -0000
Delivered-To: [EMAIL PROTECTED]
Received: (qmail 11052 invoked for bounce); 20 Oct 1999 02:59:31 -0000
Date: 20 Oct 1999 02:59:31 -0000
From: [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Subject: failure notice
Hi. This is the qmail-send program at zipcon.net.
I tried to deliver a bounce message to this address, but the bounce bounced!
<[EMAIL PROTECTED]>:
128.11.68.145 failed after I sent the message.
Remote host said: 552 qdirdel.1 error 100:This user does not have a rocketmail.com
account ([EMAIL PROTECTED])
--- Below this line is the original bounce.
Return-Path: <>
Received: (qmail 11034 invoked from network); 20 Oct 1999 02:59:28 -0000
Received: from company123.com (209.XX.XX.XXX)
by mail.zipcon.net with SMTP; 20 Oct 1999 02:59:28 -0000
Received: from COMPANY1#u#DOM-Message_Server by company123.com
with Novell_GroupWise; Tue, 19 Oct 1999 20:03:10 -0700
Message-Id: <[EMAIL PROTECTED]>
X-Mailer: Novell GroupWise 5.2
Date: Tue, 19 Oct 1999 20:03:10 -0700
Return-path: <>
From: [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Subject: Message status - undeliverable
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="=_93CAABFE.8FEE82B6"
This is a MIME message. If you are reading this text, you may want to
consider changing to a mail reader or gateway that understands how to
properly handle MIME multipart messages.
--=_93CAABFE.8FEE82B6
Content-Type: text/plain; charset=US-ASCII
Content-Disposition: inline
The message that you sent was undeliverable to the following:
davep (user not found)
Possibly truncated original message follows:
--=_93CAABFE.8FEE82B6
Content-Type: message/rfc822
Received: from mails.HHI.DE
([10.20.30.1])
by company123.com; Tue, 19 Oct 1999 20:02:32 -0700
Received: from 216.192.39.16 (sfr-tgn-sfu-vty16.as.wcom.net) by mails.HHI.DE
(5.65c/HHI-MX)
for <[EMAIL PROTECTED]>
id AA06749; Wed, 20 Oct 1999 04:54:59 +0200
Message-Id: <000013be5ae7$00000a94$000050b3@>
To: <[EMAIL PROTECTED]>
From: [EMAIL PROTECTED]
Subject: NO MORE REGRETS !
Date: Tue, 19 Oct 1999 19:27:47 -0700
Mime-Version: 1.0
Content-Type: text/html;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
X-Priority: 3
X-Msmail-Priority: Normal
<html>
<head>
<meta http-equiv=3D"Content-Language" content=3D"en-us">
<meta http-equiv=3D"Content-Type" content=3D"text/html; charset=3Dwindows-=
1252">
<meta name=3D"GENERATOR" content=3D"Microsoft FrontPage 4.0">
<meta name=3D"ProgId" content=3D"FrontPage.Editor.Document">
<title>email</title>
</head>
<body topmargin=3D"1" leftmargin=3D"1">
<p style=3D"line-height: 100%; margin: 0" align=3D"left"><font face=3D"Ari=
al" size=3D"4" color=3D"#0000FF"><b> </b></font></p>
<p style=3D"line-height: 100%; margin: 0" align=3D"left"><font face=3D"Ari=
al" size=3D"4" color=3D"#0000FF"><b>
Regrets come from making the wrong decisions</b></font></p>
<p style=3D"line-height: 100%; margin: 0" align=3D"left"> </p>
<table border=3D"0" width=3D"511" height=3D"730" cellspacing=3D"0">
<tr>
<td width=3D"501" height=3D"107" valign=3D"top" align=3D"left" bgcol=
or=3D"#FFFFCC" colspan=3D"2">
<p style=3D"line-height: 100%; margin-left: 5; margin-right: 2; ma=
rgin-top: 0; margin-bottom: 0" align=3D"left"> <p style=3D"line-heigh=
t: 100%; margin-left: 20; margin-right: 20; margin-top: 0; margin-bottom: =
0" align=3D"left"><font face=3D"Arial" size=3D"2">
The best way to void making the<b> wrong</b> decisions is to know =
the outcome
of your decision before
taking action. You don't need to be psychic to do it, all you nee=
--=_93CAABFE.8FEE82B6--
---end of bounce message---
On Tue, Oct 19, 1999 at 03:55:21PM -0700, Ted Lin wrote:
>
> If you wouldn't mind, I've read the man page, but I'm afraid I still don't
> fully understand the purpose of the percenthack control file. Can you shed
> any light on this?
(Here is a more elaborate answer.
This is also to be found at http://x42.com/qmail/doc/percenthack.txt
about the percenthack
=====================
This is a historic reminiscence from the days routed mail through serial
lines and uucp-connections more often than on leased lines.
It's use is to force routing of mail through one or several hosts, and to be
able to put this chain of routing hosts into the mailaddress itself.
You should know that this form of routed addressing is not recommended,
but it is unfortunately sometimes required when the local host is acting
as a gateway between two networks.
The function is this:
If the percenthack for a domain is enabled, then the mailhost is able to
expand mailaddresses that contains %-signs. A percent hacked mail address
is an address that consists of one username and one or more fully qualified
domain name (FQDN) of hosts that the mail shall be routed through.
Example 1:
example.com shall route mail for the domain uucp.example.com according to the
percent hack scheme. The only qmail configuration needed is to put
uucp.example1.com
in the file "controls/percenthack".
Now a mail sent to [EMAIL PROTECTED] will be rewritten to
"[EMAIL PROTECTED]" before any local delivery is done.
Conclusion: The mail is relayed (or forced to be routed) through
example.com. This relaying thing is not good if you sit on an open line, and
therefore percenthack should only be enabled within closed and controlled
networks.
The percenthack is not limited to one percent-sign. Only the last percent is
rewritten, and thus the address can include a whole chain of hosts that the
mail should be routed through.
Example 2:
We set up this example routing environment
host1.example.com host2.example.com
+------------------+ +-----------------+
| | | | /
|"uucp.example.com"+----+"int.example.com"+----+ internet
| | | | \
+------------------+ +-----------------+
to show how a two-step percenthack routing is done.
host1.example.com has the domain "uucp.example.com" in the
control/percenthack-file and in rcpthosts.
host2.example.com has the domain "int.example.com" in the
control/percenthack-file and in rcpthosts.
Now when a mail is sent to
[EMAIL PROTECTED]
(and the DNS for the sending party says that the MX for uucp.example.com is
host1.example.com) then the host1.example.com removes the @uucp.example.com
part and changes the last percent to a @, thus making the address now to
read
[EMAIL PROTECTED]
which forces it to be routed to host2.example.com which handles the
int.example.com mail.
So, in step 2, the host2.example.com rewrites the address by removing
@int.example.com and delivers the mail to
[EMAIL PROTECTED]
No part of this percenthack rewriting is written into the message headers;
it merely looks like this:
------------
>From [EMAIL PROTECTED] Wed Oct 20 03:30:46 1999
Return-Path: <[EMAIL PROTECTED]>
Delivered-To: [EMAIL PROTECTED]
Received: (qmail 16667 invoked from network); 20 Oct 1999 03:30:46 -0000
Received: from host2.example.com (10.0.1.12)
by mail.example.com with SMTP; 20 Oct 1999 03:30:46 -0000
Received: (qmail 10513 invoked from network); 20 Oct 1999 03:30:37 -0000
Received: from host1.example.com (10.0.1.11)
by host2.example.com with SMTP; 20 Oct 1999 03:30:37 -0000
------------------
Note that the percenthack rewriting is done before local delivery is
handled. This is important if the righthand part of the newly expanded
address is a domain that is handled by the current host.
>From the qmail-send(8) man page:
percenthack
List of domain names where the percent hack is applied.
If domain is listed in percenthack, any address of the
form user%fqdn@domain is rewritten as user@fqdn. user
may contain %, so the percent hack may be applied
repeatedly. qmail-send handles percenthack before
locals.
