qmail Digest 24 Jan 2001 11:00:01 -0000 Issue 1254
Topics (messages 55786 through 55845):
what does this mean?
55786 by: em`s
55787 by: Alex Pennace
Re: Patches
55788 by: Markus Stumpf
55791 by: Sumith Ail
55801 by: Charles Cazabon
Re: 502 unimplemented
55789 by: Markus Stumpf
55792 by: Uwe Ohse
55804 by: Justin Bell
spell check
55790 by: Rohit Gupta
Maildir in /etc/skel ?
55793 by: Pupeno
55796 by: Mark Delany
55828 by: Al Sparks
55830 by: Travis Leuthauser
Queue
55794 by: NDSoftware
55836 by: Keary Suska
Re: slow connection init
55795 by: Steve Woolley
55798 by: Mark Delany
55806 by: Henning Brauer
55807 by: Mark Delany
55816 by: Ian Lance Taylor
Secondary MX
55797 by: Travis Leuthauser
55799 by: Greg Owen
55800 by: Mark Delany
55802 by: Chris Johnson
Re: qmail-popup process not starting successfully at boot
55803 by: Charles Cazabon
Re: Subtle qmail bug? (was Re: Handling an MX record of 0.0.0.0 or 127.0.0.1)
55805 by: Matt Brown
55808 by: Dave Sill
55820 by: Peter Samuel
55821 by: Dave Sill
55824 by: Peter Samuel
55827 by: Scott Gifford
Delete bad messages in queue
55809 by: Alex Kramarov
virtual question?
55810 by: em`s
smtp-auth && supervise problem?
55811 by: Joe Appenzeller
55823 by: Joe Appenzeller
Problems with inject and qmail-smtpd
55812 by: Marcus Korte
55813 by: Alex Pennace
55825 by: Dave Sill
Re: QMail DOS
55814 by: Boz Crowther
Re: Special Routing setup
55815 by: Dave Sill
55829 by: Lieven Van Acker
Re: LOAD-BALANCING WITH QMAIL.
55817 by: Ian Matyssik
Setup of the big relay system.
55818 by: Ian Matyssik
/.qmail files
55819 by: Kelly Prophet
55826 by: Dave Sill
unsubscribe [EMAIL PROTECTED]
55822 by: Raymond Kirby
55834 by: qmail.artemas.reachin.com
qmail-users confusion
55831 by: Dave Sill
55832 by: Peter Samuel
55833 by: Dave Sill
Re: Possible problem with qmail-qmtpc patch
55835 by: Michael T. Babcock
Needing to inject ALOT of queued Email back into system.
55837 by: Greg Moeller
55838 by: Greg Moeller
55840 by: qmail.artemas.reachin.com
linking /usr/sbin/sendmail to /var/qmail/bin/sendmail ?
55839 by: Vegard Hansen
55841 by: qmail.artemas.reachin.com
55842 by: Harald Hanche-Olsen
Concurrencylocal and quota
55843 by: Ryszard Lach
large todo queue - HELP!
55844 by: Peter van Dijk
55845 by: Sumith Ail
Administrivia:
To unsubscribe from the digest, e-mail:
[EMAIL PROTECTED]
To subscribe to the digest, e-mail:
[EMAIL PROTECTED]
To bug my human owner, e-mail:
[EMAIL PROTECTED]
To post to the list, e-mail:
[EMAIL PROTECTED]
----------------------------------------------------------------------
Jan 23 19:03:30 server qmail: 980247810.020311 status: local 1/10 remote
0/20
Jan 23 19:03:30 server qmail: 980247810.032344 delivery 29: deferral:
Unable_to_find_alias_user!/
Jan 23 19:03:30 server qmail: 980247810.032445 status: local 0/10 remote
0/20
what does Unable_to_find_alias_user!/ means??
what can i do to fix this??
thanks//
e m ` s ([EMAIL PROTECTED])
U r l : http://ems.ath.cx
On Tue, Jan 23, 2001 at 07:05:05PM +0800, em`s wrote:
> Jan 23 19:03:30 server qmail: 980247810.020311 status: local 1/10 remote
> 0/20
> Jan 23 19:03:30 server qmail: 980247810.032344 delivery 29: deferral:
> Unable_to_find_alias_user!/
> Jan 23 19:03:30 server qmail: 980247810.032445 status: local 0/10 remote
> 0/20
>
> what does Unable_to_find_alias_user!/ means??
You haven't properly installed qmail.
> what can i do to fix this??
Properly install qmail.
Also, "server" is a poor choice for a host name. See
<http://www.rfc-editor.org/rfc/rfc1178.txt>.
On Tue, Jan 23, 2001 at 02:45:39PM +0530, Sumith Ail wrote:
> We are planning to install Qmail on a production server which will have
> around 500+ virtual domains. I am aware that some patches need to be
> applied to qmail before it can be used on a production server.
This is wrong.
> Can someone please let me know on what are the necessary patches to be
> applied. I am using the latest memphis RPM's of Qmail, daemontools and
> ucspi-tcp package. So I would like to know on which are the most
> required patches to these RPM's
You don't need any patches.
If you like modifications of some sort see
http://www.qmail.org/
and pick what you like.
\Maex
--
SpaceNet AG | http://www.Space.Net/ | Stress is when you wake
Research & Development | mailto:[EMAIL PROTECTED] | up screaming and you
Joseph-Dollinger-Bogen 14 | Tel: +49 (89) 32356-0 | realize you haven't
D-80807 Muenchen | Fax: +49 (89) 32356-299 | fallen asleep yet.
Hi
You mean that the memphis RPM's without any patches are fine to be run on
production servers.
Regards
-Sumith
----- Original Message -----
From: Markus Stumpf <[EMAIL PROTECTED]>
To: Sumith Ail <[EMAIL PROTECTED]>
Cc: <[EMAIL PROTECTED]>
Sent: Tuesday, January 23, 2001 4:58 PM
Subject: Re: Patches
> On Tue, Jan 23, 2001 at 02:45:39PM +0530, Sumith Ail wrote:
> > We are planning to install Qmail on a production server which will have
> > around 500+ virtual domains. I am aware that some patches need to be
> > applied to qmail before it can be used on a production server.
>
> This is wrong.
>
> > Can someone please let me know on what are the necessary patches to be
> > applied. I am using the latest memphis RPM's of Qmail, daemontools and
> > ucspi-tcp package. So I would like to know on which are the most
> > required patches to these RPM's
>
> You don't need any patches.
> If you like modifications of some sort see
> http://www.qmail.org/
> and pick what you like.
>
> \Maex
>
> --
> SpaceNet AG | http://www.Space.Net/ | Stress is when you
wake
> Research & Development | mailto:[EMAIL PROTECTED] | up screaming and
you
> Joseph-Dollinger-Bogen 14 | Tel: +49 (89) 32356-0 | realize you
haven't
> D-80807 Muenchen | Fax: +49 (89) 32356-299 | fallen asleep yet.
>
Sumith Ail <[EMAIL PROTECTED]> wrote:
> You mean that the memphis RPM's without any patches are fine to be run on
> production servers.
Yes. It wouldn't be a particularly useful piece of software otherwise.
This just re-illustrates Russell's point about people misinterpreting what
the "patches" section of qmail.org is all about.
How about "Third-party functional differentiation utilities"? :)
Charles
--
-----------------------------------------------------------------------
Charles Cazabon <[EMAIL PROTECTED]>
GPL'ed software available at: http://www.qcc.sk.ca/~charlesc/software/
Any opinions expressed are just that -- my opinions.
-----------------------------------------------------------------------
On Tue, Jan 23, 2001 at 10:56:29AM +0100, Stef Hoesli Wiederwald wrote:
> After that I said:
> quit
> and got a
> 451 timeout (#4.4.2)
> some minutes later...
First I thought it is your keyboard, now I tend to think it's either
your telnet of your terminal device driver thats broken. Or maybe it's
the TCP/IP Stack on that machine?
How is qmail started on that host, what do the logfiles say, what
version of qmail are you using, did you apply any modifications, what
OS is the machine running, ...
As long as you're hiding information we cannot even test the qmail
smtpd server. If you want help, provide information.
\Maex
--
SpaceNet AG | http://www.Space.Net/ | Stress is when you wake
Research & Development | mailto:[EMAIL PROTECTED] | up screaming and you
Joseph-Dollinger-Bogen 14 | Tel: +49 (89) 32356-0 | realize you haven't
D-80807 Muenchen | Fax: +49 (89) 32356-299 | fallen asleep yet.
On Mon, Jan 22, 2001 at 05:20:56PM +0100, Stef Hoesli Wiederwald wrote:
> We use qmail on one of our systems (How do I find out what version it
> is? Did not install it myself...).
Version:
1.03 has /var/qmail/bin/bouncesaying
1.02 qmail-send logs "status:" lines.
1.01 supports control/recipientmap. grep qmail-send for it.
No other version has that.
everything else means 1.00 or an even older release.
Regards, Uwe
On Tue, Jan 23, 2001 at 10:49:33AM +0100, Stef Hoesli Wiederwald wrote:
# sos:~> telnet <QMAIL HOST> 25
# Trying <QMAIL HOST IP>...
You must have a much better version of telnet than I
Here is what I get
telnet <QMAIL HOST IP>
Usage: telnet [-8] [-E] [-L] [-S tos] [-a] [-c] [-d] [-e char] [-l user]
[-n tracefile] [-b hostalias ][-r] [host-name [port]]
--
Justin Bell
|
Any spel check modules which can be incorporated in
web based email system using qmail...
|
Is it posible to create Maildir in /etc/skel/ (maildirmake /etc/skel/Maildir)
so whenever I add a user it is added with its Maildir ? (I'm working with
OpenBSD 2.7).
--
Pupeno: [EMAIL PROTECTED] www.pupeno.com.ar
On Wed, Jan 24, 2001 at 10:48:45AM +0000, Pupeno wrote:
> Is it posible to create Maildir in /etc/skel/ (maildirmake /etc/skel/Maildir)
> so whenever I add a user it is added with its Maildir ? (I'm working with
> OpenBSD 2.7).
Yes. (But it's an OS question you're asking, not a qmail question).
Regards.
--- Pupeno <[EMAIL PROTECTED]> wrote:
> Is it posible to create Maildir in /etc/skel/ (maildirmake /etc/skel/Maildir)
> so whenever I add a user it is added with its Maildir ? (I'm working with
> OpenBSD 2.7).
> --
> Pupeno: [EMAIL PROTECTED] www.pupeno.com.ar
I don't know. You could easily experiment and see if it works.
However, an alternative would be to modify the machine�s adduser script
(I think bsd uses a text shell script for adduser or its equivalent)
and add the maildirmake command to it.
=== Al
__________________________________________________
Do You Yahoo!?
Yahoo! Auctions - Buy the things you want at great prices.
http://auctions.yahoo.com/
Using FreeBSD, you can put a Maildir in your skel directory. I'm not
certain with OpenBSD though.
-Travis
>> Is it posible to create Maildir in /etc/skel/ (maildirmake
/etc/skel/Maildir)
>> so whenever I add a user it is added with its Maildir ? (I'm working with
>> OpenBSD 2.7).
>> --
>> Pupeno: [EMAIL PROTECTED] www.pupeno.com.ar
>
>I don't know. You could easily experiment and see if it works.
>However, an alternative would be to modify the machine�s adduser script
>(I think bsd uses a text shell script for adduser or its equivalent)
>and add the maildirmake command to it.
> === Al
Hi,
How i can delete a message in the queue because my qmail is crashed and my
queue is not cleaned and my customer receveid this message by 120 !
Thanks
Nicolas DEFFAYET, NDSoftware
http://www.ndsoftware.net - [EMAIL PROTECTED]
France: Tel +33 671887502 - Fax N/A
UK: Tel +44 8453348750 - Fax +44 8453348751
USA: Tel N/A - Fax N/A
Find the message in /var/qmail/queue/mess, the find for the message id under
queue: find /var/qmail/queue -name 'MSG_ID*' -exec rm -f {} \; or something
similar.
-K
"Do not meddle in the affairs of dragons, because you are crunchy and taste
good with ketchup."
> From: "NDSoftware" <[EMAIL PROTECTED]>
> Date: Tue, 23 Jan 2001 15:08:09 +0100
> To: <[EMAIL PROTECTED]>
> Subject: Queue
>
> Hi,
> How i can delete a message in the queue because my qmail is crashed and my
> queue is not cleaned and my customer receveid this message by 120 !
> Thanks
>
> Nicolas DEFFAYET, NDSoftware
> http://www.ndsoftware.net - [EMAIL PROTECTED]
> France: Tel +33 671887502 - Fax N/A
> UK: Tel +44 8453348750 - Fax +44 8453348751
> USA: Tel N/A - Fax N/A
>
>
I added both the -R and -H options and the initial connection
lag does not seem to have reoccurred. I will be trying the
-R and the -H individually to isolate the problem.
However, later in your note, you mentioned identd. I have
removed this service from my exposed email and web servers
because I heard they were security holes. I also thought identd
was only for other hosts trying to id processes on my box and
thus figured it was not needed. Could this be the problem?
Steve
----- Original Message -----
From: "Andrew Richards" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>; "'Steve Woolley'" <[EMAIL PROTECTED]>
Sent: Monday, January 22, 2001 2:35 PM
Subject: RE: slow connection init
> >flys (very fast). After aprox one day, any
> >connection into this server (sshd, telnet , pop,
> >smtp, etc) takes a while to initiate. Sometimes
> >more than 60 seconds -- which of course times out
> >most POP connections. Once connected, everything seems to
> >act normal (connections initiated quickly).
>
> Steve,
>
> Also take a look at the -R, -H and -l options to tcpserver - these
> relate to DNS and identd lookups - try using all three (see the
> man page) and see if the behaviour of the box changes. If so,
> investigate why - then either leave these options in, or address
> the issues these options work around.
>
> cheers,
>
> Andrew.
>
>
On Tue, Jan 23, 2001 at 09:05:10AM -0500, Steve Woolley wrote:
> I added both the -R and -H options and the initial connection
> lag does not seem to have reoccurred. I will be trying the
> -R and the -H individually to isolate the problem.
> However, later in your note, you mentioned identd. I have
> removed this service from my exposed email and web servers
> because I heard they were security holes. I also thought identd
> was only for other hosts trying to id processes on my box and
> thus figured it was not needed. Could this be the problem?
No. This is how it works:
Scenario 1: When your system establishes an SMTP session with a remote
system, the remote system may try and contact an ident server on your
system.
Scenario 2: When a remote system establishes an SMTP session with your
system, your tcpserver will try and contact the ident server on the
remote system. If the remote system is not running an ident server,
tcpserver has to wait for the timeout before proceeding. This is what
was happening to you.
By using the -R option on your tcpserver, you stop it trying to
contact the remote ident server.
You do not need to run an ident server, and indeed many people
don't. The impact applies to Scenario 1. Remote systems will be trying
to contact your ident server when you send mail out. Because you are
not running an ident server, they will timeout on that connection
prior to proceeding with the SMTP transaction. This is no big deal,
but it does add delay to some of your outgoing emails.
Regards.
>
> Steve
>
> ----- Original Message -----
> From: "Andrew Richards" <[EMAIL PROTECTED]>
> To: <[EMAIL PROTECTED]>; "'Steve Woolley'" <[EMAIL PROTECTED]>
> Sent: Monday, January 22, 2001 2:35 PM
> Subject: RE: slow connection init
>
>
> > >flys (very fast). After aprox one day, any
> > >connection into this server (sshd, telnet , pop,
> > >smtp, etc) takes a while to initiate. Sometimes
> > >more than 60 seconds -- which of course times out
> > >most POP connections. Once connected, everything seems to
> > >act normal (connections initiated quickly).
> >
> > Steve,
> >
> > Also take a look at the -R, -H and -l options to tcpserver - these
> > relate to DNS and identd lookups - try using all three (see the
> > man page) and see if the behaviour of the box changes. If so,
> > investigate why - then either leave these options in, or address
> > the issues these options work around.
> >
> > cheers,
> >
> > Andrew.
> >
> >
>
On Tue, Jan 23, 2001 at 02:22:24PM +0000, Mark Delany wrote:
> You do not need to run an ident server, and indeed many people
> don't. The impact applies to Scenario 1. Remote systems will be trying
> to contact your ident server when you send mail out. Because you are
> not running an ident server, they will timeout on that connection
> prior to proceeding with the SMTP transaction. This is no big deal,
> but it does add delay to some of your outgoing emails.
This could easily avoided in your firewall: for every incoming ident-lookup
reset the connection immediately.
In ipfilter this would be something like:
block return-rst in quick on [interface] from any to any port=ident
--
Henning Brauer | BS Web Services
Hostmaster BSWS | Roedingsmarkt 14
[EMAIL PROTECTED] | 20459 Hamburg
http://www.bsws.de | Germany
On Tue, Jan 23, 2001 at 04:05:38PM +0100, Henning Brauer wrote:
> On Tue, Jan 23, 2001 at 02:22:24PM +0000, Mark Delany wrote:
> > You do not need to run an ident server, and indeed many people
> > don't. The impact applies to Scenario 1. Remote systems will be trying
> > to contact your ident server when you send mail out. Because you are
> > not running an ident server, they will timeout on that connection
> > prior to proceeding with the SMTP transaction. This is no big deal,
> > but it does add delay to some of your outgoing emails.
>
> This could easily avoided in your firewall: for every incoming ident-lookup
> reset the connection immediately.
> In ipfilter this would be something like:
> block return-rst in quick on [interface] from any to any port=ident
Or more simply have a tcpserver process listen on the ident port and
run "/bin/true" for each connection.
tcpserver -HRl0 0 113 /bin/true
Regards.
"Mark Delany" <[EMAIL PROTECTED]> writes:
> Scenario 2: When a remote system establishes an SMTP session with your
> system, your tcpserver will try and contact the ident server on the
> remote system. If the remote system is not running an ident server,
> tcpserver has to wait for the timeout before proceeding. This is what
> was happening to you.
Not running an identd server need not lead to a noticeable timeout.
The remote system should refuse the connection attempt immediately,
tcpserver will get a RST packet and an ECONNREFUSED error, and will
continue immediately without waiting for the timeout period.
The timeout occurs because of firewalls. Firewalls drop the packets
intended for the identd server, so the tcpserver never gets the RST
packet and has to wait for the full timeout (default 26 seconds,
controllable by the -t option).
If the firewall is configured to pass through packets with a
destination port of 113, then remote servers won't time out.
Alternatively, a good feature to have on firewalls would be
``immediately refuse TCP connections to these port numbers.'' I don't
know if firewalls typically have that sort of feature.
Ian
Does qmail use the default queuelife and backoff algorithm for delivering
mail to a primary MX when it is acting as secondary? Or does it do something
special?
Thanks,
Travis Leuthauser
Technical Support
Broadband IP
> Does qmail use the default queuelife and backoff algorithm
> for delivering mail to a primary MX when it is acting as
> secondary? Or does it do something special?
It uses the default queuelifetime and backoff.
Note that having a concurrencyremote higher than the primary MX is
willing to handle can result in undue delays, because it'll start backing
off when the Primary says "No more!". If concurrencyremote is less than the
Primary will stop at, then it will run smoother.
--
gowen -- Greg Owen -- [EMAIL PROTECTED]
SoftLock.com is now DigitalGoods!
On Tue, Jan 23, 2001 at 08:20:01AM -0600, Travis Leuthauser wrote:
> Does qmail use the default queuelife and backoff algorithm for delivering
> mail to a primary MX when it is acting as secondary? Or does it do something
> special?
All remote mail is treated the same. In fact qmail doesn't really know
of the distinction you are making. It's simple a remote mail in the
queue that needs to be delivered.
If you want to make a distinction, you'll need to set up a separate
instance of qmail with a different queuelifetime parameter. There is
no mechanism to change the retry algorithm.
Regards.
On Tue, Jan 23, 2001 at 08:20:01AM -0600, Travis Leuthauser wrote:
> Does qmail use the default queuelife and backoff algorithm for delivering
> mail to a primary MX when it is acting as secondary? Or does it do something
> special?
It doesn't do anything special. It doesn't know it's a secodary MX; it just
knows it has a message it needs to deliver to a remote host, just like any
other message it needs to deliver to a remote host.
Chris
Keith Edwards <[EMAIL PROTECTED]> wrote:
>
> When our SUN box reboots pop3d does not start.
[...]
> In /nohup.out is the line
> 'env: No such file or directory'
> caused by the 'nohup /var/qmail/start-pop3d' command.
[...]
> =============================
> 4).
> Here is the script that is called
>
> #!/bin/sh
> #
> # /var/qmail/start-pop3d
> # Startup script for pop3d using tcpserver and vchkpw
> #
>
> exec env - PATH="/var/qmail/bin:/usr/local/bin:$PATH" \
> tcpserver -uNNN -gNNN 0 pop-3 /var/qmail/bin/qmail-popup \
> my.mailserver.com /home/vpopmail/bin/vchkpw \
> /var/qmail/bin/qmail-pop3d Maildir &
Change the start of that script to "exec /usr/bin/env - PATH=..." or
whatever the proper path is for you. That's what the error above means.
Charles
--
-----------------------------------------------------------------------
Charles Cazabon <[EMAIL PROTECTED]>
GPL'ed software available at: http://www.qcc.sk.ca/~charlesc/software/
Any opinions expressed are just that -- my opinions.
-----------------------------------------------------------------------
Scott Gifford <[EMAIL PROTECTED]> writes:
> Keary Suska <[EMAIL PROTECTED]> writes:
>
> > This would definitely be a bug of concern--even sendmail (yoiks!) knows how
> > to handle 0.0.0.0. But shouldn't qmail bounce the message as a possible MX
> > loop?
>
> It should, but does not. Putting it into ipme would cause it to.
This has been a feature of recent spam, which is probably why it's now
an issue. Several spam senders are now having sender addresses of
<spammer>@<spamdomain>, where <spamdomain> resolves via DNS to
'0.0.0.0'.
Eventually qmail rejects the message because it recognises that it's
looped around too much, of course.
-Matt
--
| Matthew J. Brown - Senior Network Administrator - NBCi Shopping |
| 1983 W. 190th St, Suite 100, Torrance CA 90504 |
| Phone: (310) 538-7122 | Work: [EMAIL PROTECTED] |
| Cell: (714) 457-1854 | Personal: [EMAIL PROTECTED] |
Matt Brown <[EMAIL PROTECTED]> wrote:
>This has been a feature of recent spam, which is probably why it's now
>an issue. Several spam senders are now having sender addresses of
><spammer>@<spamdomain>, where <spamdomain> resolves via DNS to
>'0.0.0.0'.
>
>Eventually qmail rejects the message because it recognises that it's
>looped around too much, of course.
Here's a possible fix. In control/virtualdomains:
[0.0.0.0]:alias-devnull
And in ~alias/.qmail-devnull-default
#
Which should throw away all mail to MX's resolving to 0.0.0.0.
-Dave
On Tue, 23 Jan 2001, Dave Sill wrote:
>
> Here's a possible fix. In control/virtualdomains:
>
> [0.0.0.0]:alias-devnull
>
> And in ~alias/.qmail-devnull-default
>
> #
>
> Which should throw away all mail to MX's resolving to 0.0.0.0.
Are you sure that will work? The envelope details won't mention
0.0.0.0, so how will qmail-send know to use that entry in
virtualdomains. (I haven't tested it yet, but I'm about to).
--
Regards
Peter
----------
Peter Samuel [EMAIL PROTECTED]
http://www.e-smith.org (development) http://www.e-smith.com (corporate)
Phone: +1 613 368 4398 Fax: +1 613 564 7739
e-smith, inc. 1500-150 Metcalfe St, Ottawa, ON K2P 1P1 Canada
"If you kill all your unhappy customers, you'll only have happy ones left"
Peter Samuel <[EMAIL PROTECTED]> wrote:
>On Tue, 23 Jan 2001, Dave Sill wrote:
>>
>> Here's a possible fix. In control/virtualdomains:
>>
>> [0.0.0.0]:alias-devnull
>>
>> And in ~alias/.qmail-devnull-default
>>
>> #
>>
>> Which should throw away all mail to MX's resolving to 0.0.0.0.
>
>Are you sure that will work?
Nope. That's why it's a "possible fix". I didn't have a handy
"resolves to 0.0.0.0" domain name to test.
>The envelope details won't mention
>0.0.0.0, so how will qmail-send know to use that entry in
>virtualdomains.
Beats me. :-)
>(I haven't tested it yet, but I'm about to).
Let us know...
-Dave
On Tue, 23 Jan 2001, Dave Sill wrote:
> Here's a possible fix. In control/virtualdomains:
>
> [0.0.0.0]:alias-devnull
>
> And in ~alias/.qmail-devnull-default
>
> #
>
> Which should throw away all mail to MX's resolving to 0.0.0.0.
My tests show that that won't work:
echo "[0.0.0.0]:alias-devnull" > /tmp/vd
cat /tmp/vd /var/qmail/control/virtualdomains > /tmp/nvd
cp /tmp/nvd /var/qmail/control/virtualdomains
chmod 644 /var/qmail/control/virtualdomains
echo # > ~alias/.qmail-devnull
chmod 644 ~alias/.qmail-devnull
killall -HUP qmail-send
/var/qmail/bin/qmail-remote \[0.0.0.0\] [EMAIL PROTECTED] [EMAIL PROTECTED] << EOF
From: [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Subject: testing
hello
EOF
qmail-remote returns with
rK0.0.0.0 accepted message.
Remote host said: 250 Message accepted for delivery
and the qmail log shows
980273407.745867 new msg 1112229
980273407.746179 info msg 1112229: bytes 366 from <[EMAIL PROTECTED]>
qp 8502 uid 8
980273407.751043 starting delivery 62: msg 1112229 to remote
[EMAIL PROTECTED]
980273407.751057 status: local 0/10 remote 1/20
980273467.837834 delivery 62: success:
209.217.125.238_accepted_message./Remote_host_said:_250_ok_980273466_qp_26250/
980273467.837854 status: local 0/10 remote 0/20
980273467.837860 end msg 1112229
Because 0.0.0.0 is not mentioned in the envelope RCPT TO header, the
virtualdomain rule does not apply.
--
Regards
Peter
----------
Peter Samuel [EMAIL PROTECTED]
http://www.e-smith.org (development) http://www.e-smith.com (corporate)
Phone: +1 613 368 4398 Fax: +1 613 564 7739
e-smith, inc. 1500-150 Metcalfe St, Ottawa, ON K2P 1P1 Canada
"If you kill all your unhappy customers, you'll only have happy ones left"
Matt Brown <[EMAIL PROTECTED]> writes:
> This has been a feature of recent spam, which is probably why it's now
> an issue. Several spam senders are now having sender addresses of
> <spammer>@<spamdomain>, where <spamdomain> resolves via DNS to
> '0.0.0.0'.
>
> Eventually qmail rejects the message because it recognises that it's
> looped around too much, of course.
Right, but it's a very effective (perhaps inadvertant) DOS tool. If
you can generate a stream of 10 messages/sec of these, it's the
equivalent of generating about 300 messages/sec --- a great way of
turning a puny dial-up connection into a mail server crushing machine.
We had a spammer sending a huge number of messages to users at this
address (<sigh> their fake bounce addresses are now getting on each
others' list...), which was causing our not-processed queues to hover
around 100, which was causing regular messages to be processed very
slowly.
Since qmail works around this simple mail loop for other address
referring to the local machine, it should do so for 0.0.0.0 as well.
------ScottG.
|
Hello. I got a small problem i hope you can help me.
A couple of days ago I have recompiled qmail with some patch/third
party enchancement/whatever_yuo_want_to_call_it, and the morning after I
saw alot of qmail-smtpd and qmail-queue processes running for 2,3,4 hours.
I thought that the patch did this, killed all of these processes and kept
on, thinking of reverting to prevous configuration in the evening. during
the day I have discovered, that I was getting alot of these processed, but
only when a certain server was trying to connect to me. I have blocked
that server in tcp.smtp and everything went on fine. i only later I have
discovered that is was some music lover that was trying to sent 5 mp3
files in the same time over and over again, and his server was sitting on
a wery slow link, so I was fooled to think that theer was a problem with
my server since I didn't see any traffic to corespond to 5 qmail-smtpd
processes.
Now I see, that these messages are still sitting in my queue, they
are already past queuelifetime, and they are not being flushed with "kill
-ALRM `pidof qmail-send`"
shoud I do something about them ? |
__________________________________________________
IncrediMail - Email has finally
evolved - Click
Here
|
hi i have a question about sending virtually, whatever u call that.. here's
the scenario..
I have a server which i got from some free dns hosting like dyndns.org
and i want to receive mail on that host like
host : hate.merseine.nu
i point it out to
hate.merseine.nu > 123.456.78.910
and now, i installed qmail to 123.456.78.910.
btw? 123.456.78.910 is not my primary mail srvr. i just want qmail to pass
it on to another server.
is this possible?
like i want to have a email like [EMAIL PROTECTED] > and i want to pass
it on virtually to [EMAIL PROTECTED] my own email.
is this possible?
thanks =)
curious kid =)
e m ` s ([EMAIL PROTECTED])
U r l : http://ems.ath.cx
Hi
I have qmail 1.03 running with vpopmail.
I patched qmail-smtpd with esmtp-tls and smtp-auth.
When I run it "stand-alone" (not supervised) it works perfectly,
authentification via vpopmail-passwdfiles etc.
when I want to run it via svscan / supervise, I can't authenticate myself.
My username/password is always rejected.
Any ideas why?
it works when I start qmail-smtpd by:
#/usr/local/bin/softlimit -m 100000000 /usr/local/bin/tcpserver -R -v -p -x
/etc/tcp.smtp.cdb -c 2000 -u qmaild -g
nofiles 0 smtp /var/qmail/bin/qmail-smtpd /home/vpopmail/bin/vchkpw
/bin/true /bin/cmd5checkpw /bin/true 2>&1
run-file for svscan (note: the command executed is exactly the same as
above)
#!/bin/sh
QMAILDUID=`id -u qmaild`
NOFILESGID=`id -g qmaild`
MAXSMTPD=`cat /var/qmail/control/concurrencyincoming`
exec /usr/local/bin/softlimit -m 100000000
/usr/local/bin/tcpserver -R -v -p -x /etc/tcp.smtp.cdb -c "$MAXSMTPD" -u
"$QMAILDUID" -g
"$NOFILESGID" 0 smtp /var/qmail/bin/qmail-smtpd /home/vpopmail/bin/vchkpw
/bin/true /bin/cmd5checkpw /bin/true 2>&1
Thanx
Joe
Resolved the problem myself (finally ...):
Was a startup-skrip problem:
problematic:
QMAILDUID=`id -u qmaild` (--> returns a number on my box)
NOFILESGID=`id -g qmaild`
exec /usr/local/bin/softlimit -m 100000000
/usr/local/bin/tcpserver -R -v -p -x /etc/tcp.smtp.cdb -c "$MAXSMTPD" -u
"$QMAILDUID " -g "$NOFILESGID" 0 smtp /var/qmail/bin/qmail-smtpd
home/vpopmail/bin/vchkpw /bin/true /bin/cmd5checkpw /bin/true 2>&1
working solution:
QMAILDUID=qmaild
NOFILESGID=nofiles
exec /usr/local/bin/softlimit -m 100000000
/usr/local/bin/tcpserver -R -v -p -x /etc/tcp.smtp.cdb -c "$MAXSMTPD" -u
"$QMAILDUID " -g "$NOFILESGID" 0 smtp /var/qmail/bin/qmail-smtpd
home/vpopmail/bin/vchkpw /bin/true /bin/cmd5checkpw /bin/true 2>&1
Later
Joe
Hi everybody,
I set qmail up regarding to LWQ.
The initial testing with injeting messages to qmail-inject gives the
follwing output and it does not work.
qmail-inject: fatal: qq trouble in home directory (#4.3.0)
In the log file:
2001-01-23 17:03:20.131989500 tcpserver: status: 1/20
2001-01-23 17:03:20.132080500 tcpserver: pid 4044 from ...
2001-01-23 17:03:20.134806500 tcpserver: ok 4044 ...
2001-01-23 17:03:20.134811500 tcpserver: warning: dropping connection,
unable to run /var/mail/bin/qmail-smtpd: access denied
2001-01-23 17:03:20.134860500 tcpserver: end 4044 status 28416
2001-01-23 17:03:20.134873500 tcpserver: status: 0/20
Has anybody an idea, what this messages mean?
Which owner:group/protection does qmail-smtpd need?
Every help is welcome.
Marcus
--
Sent through GMX FreeMail - http://www.gmx.net
On Tue, Jan 23, 2001 at 05:37:41PM +0100, Marcus Korte wrote:
> I set qmail up regarding to LWQ.
> The initial testing with injeting messages to qmail-inject gives the
> follwing output and it does not work.
>
> qmail-inject: fatal: qq trouble in home directory (#4.3.0)
>
> In the log file:
> 2001-01-23 17:03:20.131989500 tcpserver: status: 1/20
> 2001-01-23 17:03:20.132080500 tcpserver: pid 4044 from ...
> 2001-01-23 17:03:20.134806500 tcpserver: ok 4044 ...
> 2001-01-23 17:03:20.134811500 tcpserver: warning: dropping connection,
> unable to run /var/mail/bin/qmail-smtpd: access denied
> 2001-01-23 17:03:20.134860500 tcpserver: end 4044 status 28416
> 2001-01-23 17:03:20.134873500 tcpserver: status: 0/20
>
> Has anybody an idea, what this messages mean?
The permissions on your qmail setup are messed up. Run make check in
the qmail source directory.
Marcus Korte <[EMAIL PROTECTED]> wrote:
>In the log file:
>2001-01-23 17:03:20.131989500 tcpserver: status: 1/20
>2001-01-23 17:03:20.132080500 tcpserver: pid 4044 from ...
>2001-01-23 17:03:20.134806500 tcpserver: ok 4044 ...
>2001-01-23 17:03:20.134811500 tcpserver: warning: dropping connection,
>unable to run /var/mail/bin/qmail-smtpd: access denied
^^^^
That should be "qmail", no? Looks like a typo in your
/var/qmail/supervise/qmail-smtpd/run script.
I *strongly* recommend cutting and pasting the scripts...
-Dave
I was just reading that IBM is porting Linux apps to AIX. Maybe DOS is
next...
----- Original Message -----
From: "Russell Nelson" <[EMAIL PROTECTED]>
To: "Qmail Mailing List" <[EMAIL PROTECTED]>
Sent: Monday, January 22, 2001 10:20 PM
Subject: Re: QMail DOS
> QMail doesn't run under DOS.
>
> --
> -russ nelson <[EMAIL PROTECTED]> http://russnelson.com
> Crynwr sells support for free software | PGPok | "This is Unix...
> 521 Pleasant Valley Rd. | +1 315 268 1925 voice | Stop acting so
helpless."
> Potsdam, NY 13676-3213 | +1 315 268 9201 FAX | --Daniel J. Bernstein
>
Lieven Van Acker <[EMAIL PROTECTED]> wrote:
>mail from cust.mailserv --------> MR ---------> dest-mx
> | ^
> | |
> V |
> MAILPROCESSOR
>
>mail to cust.domain ------------> MR ----------> cust.postoffice
> | ^
> | |
> V |
> MAILPROCESSOR
>
>So all mail to and from a couple of mail-domains has to be routed via MR
>to MAILPROCESSOR (Virusscanning, other processing)
I can't really tell what you're trying to accomplish. How about
describing in words the paths you want incoming and outgoing messages
for the various classes of domains to take?
-Dave
Dave Sill wrote:
>
> I can't really tell what you're trying to accomplish. How about
> describing in words the paths you want incoming and outgoing messages
> for the various classes of domains to take?
>
> -Dave
OK, so here I go:
There are a number of customers, whose incoming and outgoing e-mail must be
rerouted via a central service that provides virus-scanning. As these
customers are connected each to various ISP's, and we want to provide the
virus-scanning service (and eventually other mail processing services), all
mail should be routed through one relay host (MYRELAY), the one I have to
setup. The virus-scanning is provided by a third party relay host
(PROCRELAY) who will accept smtp-connections from our relay host, process the
mail and route it back to the final destination.
Incoming mail:
1. any sender: mail sent to: [EMAIL PROTECTED]
2. mail is relayed to MYRELAY (MX record for customer.net)
3. (after checking that the message is not coming from PROCRELAY and thus
has to be processed before forwarded) mail is relayed to PROCRELAY
4. (after being processed) mail is sent to MYRELAY (MX record for
customer.net)
5. (message comes from PROCRELAY so can be safely forwarded to customers
postoffice/private relay)
Outgoing mail:
1. mail is sent from customers private mail-gateway to
[EMAIL PROTECTED]
2. mail is relayed to MYRELAY
3. (after checking that the message is not coming from PROCRELAY and thus
has to be processed before forwarded) mail is relayed to PROCRELAY
4. (after being processed) mail is sent out to MX for somewhere.net
I hope this clarifies the situation.
Because I have to route the mail on MYRELAY depending on where the previous
mail-hop was, I think I have to run 2 different mailprocesses, and use
portforwarding based on the smtp conversations client host. (there seems no
-inexpensive- way to let qmail-smtpd conversate with qmail-send)
This way I can use two different routing setups.
Lieven
Unfortunately no one did reply to me, so I found some resource on the web
and will try to investigate it.
Good luck.
qmailu writes:
> Hi Ian,
>
> Have you got this working ?? Noticed none had replied to this. Thought I'll
> get help from you. If you have , can you pls lemme know how you did this.
>
> Raghu
> ----- Original Message -----
> From: Ian Matyssik <[EMAIL PROTECTED]>
> To: <[EMAIL PROTECTED]>
> Sent: Friday, January 19, 2001 11:42 AM
> Subject: Re: LOAD-BALANCING WITH QMAIL.
>
>
>> Hello,
>>
>> I am writing to this message again. Just to confirm that the qmqp
>> supports round-robin natively. I was reading all about it and did not
>> understand. What I understood is if we keep mini-qmail on the clients and
>> have 4 servers for relaying, we just need to list all servers in
>> /var/qmail/control/qmqpservers and it will load balance them in
> round-robin
>> manner. If that is true what about if one of the relay servers goes down,
>> will it spoil something from the client side.
>>
>> Please confirm that or give some advice,
>> Regards,
>> Ian Matyssik.
>>
>> Ian Matyssik writes:
>>
>> > Hello,
>> >
>> > I am new on this list but have bin using qmail for 2 years. Now my
>> > company desided to expand mail relaying and my task is to find good
>> > sollution how to load-balance "qmqp" relays with nullmailer or qmail. I
>> > tried to look in the archive and found one thread on that topic but did
>> > not understand exactly if there is a patch for that or native capability
>> > of qmail allow that. Please help me on that. I just want to round-robin
> 4
>> > servers for now.
>> >
>> > Thank you,
>> > Ian Matyssik.
>>
>
>
> _________________________________________________________
> Do You Yahoo!?
> Get your free @yahoo.com address at http://mail.yahoo.com
>
Hello All,
I have a project of setting up 6 qmail servers just for relaying mail to
the Internet. Servers are Netra T1 Solaris boxes. And I will have to servers
which will generate actual mail. I need an advice from you people on how to
design it. I have an ideas but some things like loadbalancing and staff like
that is still in a gray area for me. Could some one share their experience
on that and on how to make this thing to be fast and productive + redundant.
Thank you.
Ian Matyssik.
|
Greetings,
I've just started the process of transferring servers. Not
really using the old data persay, but rather just moving from a sendmail system
to a qmail system.
One thing I have to have is the ability to pipe incoming email
for a specified user to a CGI script. I originally did this with the .procmailrc
files. Aswell it was possible with .forward files. I've read and am aware
that qmail *can* support the use of sendmail .forward files aswell as use
.procmailrc files through the .forward files - but personally I'd just assume
avoid it in it's entirity if at all possible.
So, basically what I am asking is if anyone is aware of a way
to pipe incoming email for a specified user to a location on the server. In this
case, a CGI-script.
Best Regards,
Kelly E. Prophet
GCN Tech
S3
|
"Kelly Prophet" <[EMAIL PROTECTED]> wrote:
>So, basically what I am asking is if anyone is aware of a way to pipe
>incoming email for a specified user to a location on the server. In
>this case, a CGI-script.
How about, in the appropriate .qmail file:
|/path/to/cgi-script
-Dave
unsubscribe [EMAIL PROTECTED]
I would think someone skilled enough to run a SMTP server would know how
to unsubscribe from a mailing list.
With ezmlm-style lists, you send mail to [EMAIL PROTECTED]
In this case, that would be [EMAIL PROTECTED]
- Sam
> unsubscribe [EMAIL PROTECTED]
>
Does users/cdb supplement regular users or is it an alternative to
them? In other words, if I have a users/assign and users/cdb with a
set of entries that doesn't include all valid local users, will mail
to one of the non-cdb users bounce?
The qmail-spawn man page says it won't:
For each recipient address, qmail-lspawn finds out which
local user controls that address. It first checks the
qmail-users mechanism; if the address is not listed there,
it invokes qmail-getpw.
The /var/qmail/docs/PIC*2local files say it won't:
| Is fred listed in qmail-users? No.
| Is there a fred account? Yes.
The qmail-lspawn.c code is kind of dense, but appears to follow the
documentation.
But when I added a dummy user (not in users/assign) to the one system
where I use qmail-users, mail to that user bounced. When I ran
qmail-pw2u and qmail-newu, mail to the dummy was delivered
locally. And when I removed the dummy entries from users/assign and
re-ran qmail-newu, mail again bounced:
2001-01-23 15:35:45.521585500 new msg 119332
2001-01-23 15:35:45.522563500 info msg 119332: bytes 931 from <[EMAIL PROTECTED]>
qp 27993 uid 49491
2001-01-23 15:35:45.850046500 starting delivery 2824054: msg 119332 to local
[EMAIL PROTECTED]
2001-01-23 15:35:45.850046500 status: local 1/60 remote 1/500
2001-01-23 15:35:45.874485500 delivery 2824054: failure:
Sorry,_no_mailbox_here_by_that_name._(#5.1.1)/
2001-01-23 15:35:46.200668500 status: local 0/60 remote 1/500
2001-01-23 15:35:46.824052500 bounce msg 119332 qp 19805
2001-01-23 15:35:46.866002500 end msg 119332
What am I missing?
-Dave
On Tue, 23 Jan 2001, Dave Sill wrote:
> Does users/cdb supplement regular users or is it an alternative to
> them? In other words, if I have a users/assign and users/cdb with a
> set of entries that doesn't include all valid local users, will mail
> to one of the non-cdb users bounce?
My experience is that if a match can't be found in users/cdb, then
qmail-getpw is called.
Does your assign/cdb file have a catchall entry?
+:alias:.......
--
Regards
Peter
----------
Peter Samuel [EMAIL PROTECTED]
http://www.e-smith.org (development) http://www.e-smith.com (corporate)
Phone: +1 613 368 4398 Fax: +1 613 564 7739
e-smith, inc. 1500-150 Metcalfe St, Ottawa, ON K2P 1P1 Canada
"If you kill all your unhappy customers, you'll only have happy ones left"
Peter Samuel <[EMAIL PROTECTED]> wrote:
>On Tue, 23 Jan 2001, Dave Sill wrote:
>
>> Does users/cdb supplement regular users or is it an alternative to
>> them? In other words, if I have a users/assign and users/cdb with a
>> set of entries that doesn't include all valid local users, will mail
>> to one of the non-cdb users bounce?
>
>My experience is that if a match can't be found in users/cdb, then
>qmail-getpw is called.
>
>Does your assign/cdb file have a catchall entry?
>
> +:alias:.......
Argh. Yes. I looked for one--I expected it to be at the end of the
file.
I also missed the part in the qmail-pw2u that says it does this:
A catch-all user, alias, controls all other addresses.
Sigh. Thanks, Peter.
-Dave
On Sun, 14 Jan 2001 20:25:44 +0100, Jurjen Oskam wrote:
> On Sun, 14 Jan 2001 13:39:01 -0500 (EST), Russell Nelson
> <[EMAIL PROTECTED]> wrote:
> >That's a misconfiguration. I'd rather that the email bounced than it
> >got delivered via SMTP silently. It could be that someone unaware of
> ...
> >delivered via qmtpd but it is failing to run for some reason. If we
> >fall back to smtp, they'll never know that it's failing unless
they're
> >watching their qmail logs carefully.
>
> But isn't that a bit in contradiction with the concept of backup
> MX'es?
If the goal is to deliver E-mail securely and quickly, falling back to
(a presumed functional) SMTP server fits the bill.
If the goal is to use QMTP "just because its there", then Russ' comments
make more sense.
--
Michael T. Babcock (PGP: 0xBE6C1895)
http://www.fibrespeed.net/~mbabcock/
We got hit by spam bigtime and our server got Very backlogged.
In order to get it alive in some form or another, we moved the mess dir of the
queue directory to a safe place,(and yes, I now know I should've moved the
todo/local/remote dirs too) ran queue-fix to tidy up the mess and
restarted qmail. So, the backlog of the server being down for hours is
cleared up and I want to put back the backlogged Email in chunks to it can
process it without horridly bogging down.
Now, how do I do it? I had assumed queuefix would've made it all better, but
it just cleans out files from todo and friends not puts anything back. (back
to that lesson thing I learned earlier :)
So, I'd like to shove the old mess directory a number at a time back into the
system and get it processed. There's around 300,000 Email to be processed.
(hence why I want to do it in pieces :)
Thanks in advance!
Greg
I'm just realizing what a <insert favorite nasty words> thing I've done by
just moving the mess dir. I eliminated all the envelope information. <siigh>
Guess that's what an 18 hour day trying to deal with 8 mail servers overloaded
with some idiot's urge to sell us a Diploma will do to one's brain. (oh, and
yes, we've called the RCMP on the idiot, let's hope he can be tracked)
(and I've only messed up one of the 8 servers, the others are running exim and
are dealing with their overload in their own chugging fashion, one of the
servers having around 8 gig of email in it's queue to process)
Anyway, I really hope there'll be some way to save some of the Email (try to
pull the To: info out? )
Greg
Did the spammer run a Rumpelstilksken on your system (for a in `cat
/usr/dict/words` ; do cat spam | mail $[EMAIL PROTECTED] ; done), or did the
spammer attempt to use you as a relay? Or do you have so many users that
when someone sends mail to all of them at once, it kills your servers?
I am surprised that you can not put the mess directory back in at
/var/qmail/queue/mess, then restart qmail. This should restore things
back to a chugging normal pace.
> Anyway, I really hope there'll be some way to save some of the Email (try to
> pull the To: info out? )
Ugh, if you have completely lost the envelope information, then you will
have to use '/var/qmail/bin/sendmail -t' on each and every one of the
messages, where the To/Cc is to an address in your rcpthosts. Do this
wrong, and a lot of duplicate mail will go out to mailing lists and what
not.
> Greg
- Sam
I read in a howto that you should link the old sendmail scripts in
/usr/sbin/sendmail to /var/qmail/bin/sendmail, but i havent got any sendmail
scripts under /var/qmail/bin/... Is this right?
I cant get qmail to listen to port 25, even if qmail is running.. Port 110
is listening but it closed connection after about 5 secs.
All the packages are installed and the scripts are running..
Is /var/qmail/bin/qmail-inject the script that takes over sendmails old
jobs?
Vennlig hilsen
Vegard Hansen
Systems developer
LOGIQ AS
Box 268
N-1752 Halden
Norway
Phone : 69190725
Fax : 69177019
Mobile : 41303033
E-mail : [EMAIL PROTECTED]
> I read in a howto that you should link the old sendmail scripts in
> /usr/sbin/sendmail to /var/qmail/bin/sendmail, but i havent got any sendmail
> scripts under /var/qmail/bin/... Is this right?
Actually, you should have /usr/sbin/sendmail (and, while you are at it
/usr/lib/sendmail) both be links to /var/qmail/bin/sendmail. This way,
legacy apps and scripts which pipe to sendmail will still work.
> I cant get qmail to listen to port 25, even if qmail is running.. Port 110
> is listening but it closed connection after about 5 secs.
Qmail needs another tool, such as inetd (ugh) or tcpserver (ucspi-tcp) to
listen on port 25.
> Is /var/qmail/bin/qmail-inject the script that takes over sendmails old
> jobs?
Sorta. qmail-inject is called by /var/qmail/bin/sendmail as a local
program you pipe outgoing mail to.
Sendmail is one big app that can, depending on its mood:
* Listen on port 25
* Update the alaias database
* Pipe local mail to remote machines
* etc.
Qmail, on the other hand, is a series of small apps, which makes it a lot
leaner, faster, and more secure than sendmail. Very useful if you get the
kind of mail traffic, say, egroups gets.
- Sam
+ "Vegard Hansen" <[EMAIL PROTECTED]>:
| but i havent got any sendmail scripts under /var/qmail/bin/...
| Is this right?
No. Your qmail installation is broken.
(BTW /var/qmail/bin/sendmail is a program, not a script.)
- Harald
Hi!
I have a problem with qmail-local, when one of mailboxes runs over quota limit
and a few new messages addressed to this mailbox come. Problem is that
qmail-lspawn spawns as many qmail-local processes, as concurrencylocal defines
and if there is more then concurrencylocal deliveries to this overfilled
mailbox - local deliveries get blocked (all qmail-lspawn processes are trying to
write messages to this mailbox). How can I solve this problem ?
TIA,
Siaco.
--
Ryszard Łach, Internet Designers S.A., Przedmiejska 6-10, 54-201 Wrocław
'echo "" |mail -s "send key pub" [EMAIL PROTECTED]' for my public GPG key
One of my big qmail boxes is not keeping up with the flow of mail:
messages in queue: 98620
messages in queue but not yet preprocessed: 71690
It's a dual PIII with 1GB of RAM.
concurrencylocal is 1024 (usage hardly ever goes above 3)
concurrencyremote is 256 (usage varies from 0 to about a 50 - remote
sites are slow)
smtpd concurrency (tcpserver -c) is 1024, usage 30-40
pop3 concurrency is 1024, usage 20-30
All logging thru multilog.
10:51AM up 8 days, 20:33, 3 users, load averages: 2.23, 1.84, 1.47
CPU states: 24.9% user, 0.0% nice, 35.8% system, 2.1% interrupt, 37.2% idle
iostat shows heavy traffic on the queue disk (a seperate 9gb scsi
disk).
My first idea is that the filesystem is not keeping up, and a bigger
queue-directory split would solve stuff. However, that's kinda hard to
implement now.
So my question is: any tips for increasing performance without
throwing the queue away, or tricks for getting rid of a todo-queue
quickly.
I am considering recompiling qmail for a bigger queuedir split, and
then applying queue-fix. Any objections or tips, or is this a bad
idea?
As an interesting detail, virtual domains are handled through the
alias user and fastforward, and then reinjected into the queue. This
probably makes it worse.
The todo-queue is *slowly* getting smaller (71288 now, compared to 71690
when I started typing), but the complete queue is growing (100121
now).
'Help!'
Greetz, Peter.
Hi,
Sorry can't help here but though would like to know how to setup the
concurency levels in tcpserver which is discussed below.
Regards
Sumith
----- Original Message -----
From: Peter van Dijk <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Wednesday, January 24, 2001 3:29 PM
Subject: large todo queue - HELP!
> One of my big qmail boxes is not keeping up with the flow of mail:
> messages in queue: 98620
> messages in queue but not yet preprocessed: 71690
>
>
> It's a dual PIII with 1GB of RAM.
>
> concurrencylocal is 1024 (usage hardly ever goes above 3)
> concurrencyremote is 256 (usage varies from 0 to about a 50 - remote
> sites are slow)
> smtpd concurrency (tcpserver -c) is 1024, usage 30-40
> pop3 concurrency is 1024, usage 20-30
>
> All logging thru multilog.
>
> 10:51AM up 8 days, 20:33, 3 users, load averages: 2.23, 1.84, 1.47
> CPU states: 24.9% user, 0.0% nice, 35.8% system, 2.1% interrupt, 37.2%
idle
>
> iostat shows heavy traffic on the queue disk (a seperate 9gb scsi
> disk).
>
> My first idea is that the filesystem is not keeping up, and a bigger
> queue-directory split would solve stuff. However, that's kinda hard to
> implement now.
>
> So my question is: any tips for increasing performance without
> throwing the queue away, or tricks for getting rid of a todo-queue
> quickly.
>
> I am considering recompiling qmail for a bigger queuedir split, and
> then applying queue-fix. Any objections or tips, or is this a bad
> idea?
>
> As an interesting detail, virtual domains are handled through the
> alias user and fastforward, and then reinjected into the queue. This
> probably makes it worse.
>
> The todo-queue is *slowly* getting smaller (71288 now, compared to 71690
> when I started typing), but the complete queue is growing (100121
> now).
>
> 'Help!'
>
> Greetz, Peter.
>
