[pfx] Email mit FG Funnel /Mailgun und eigenem Postfix Server

2025-04-05 Thread christian via Postfix-users

Hallo,
eine Kundin von mir hat eine Email die Sie seit längerem bei mir 
gehostet hat und alle Emails im IMAP Postfix bei mir liegen.


Jetzt hat sie sich bei FG Funnel ein Konto angelegt und will Newsletter 
und CRM mit dem System versenden. Also muss ich zu den bestehenden SPF, 
DKIM und DMARC Einträgen auch noch die von FG Funnel(bzw. die nutzen 
Mailgun) anlegen. Kann das überhaupt funktionieren? Wir haben jetzt die 
ersten Newsletter versendet und ein Teil davon landet im Spam Ordner der 
Empfänger.


Hat jemand von Euch damit Erfahrung?
Besten Dank für Tipps

Christian
___
Postfix-users mailing list -- postfix-users@postfix.org
To unsubscribe send an email to postfix-users-le...@postfix.org


[pfx] Re: postfix problem with resources

2025-04-05 Thread Wietse Venema via Postfix-users
natan via Postfix-users:
> Hi
> In old server i have:
> max_use=180
> 
> in new serwer I have:
> max_use=150

This determines how many processes Postfdix will run (150 in each
category). If the number is too large for your system (hardware
plus OS plus Postfix plus other prorams) then some operations will
fail due to a lack of resources.

> I don't know if it's that important

Very much. Try a smaller number until the problem does not come
back.

Wietse
___
Postfix-users mailing list -- postfix-users@postfix.org
To unsubscribe send an email to postfix-users-le...@postfix.org


[pfx] Re: spamhaus/abusix in rspamd or postfix

2025-04-05 Thread Matus UHLAR - fantomas via Postfix-users

On 23.03.25 15:16, lutz.niederer--- via Postfix-users wrote:

spamhaus and abusix have a query limit for free accounts. spamhaus does not 
seem to be a problem but abusix is 5000 queries/day. postfix as well as rspamd 
can query them.

postfix or rspamd: which one should be used to query them and why?


you can query them in both, but you should be using local caching 
non-forwarding DNS server so the same queries are not sent multiple times 
when receiving mail.


That way the same query from either won't be repeated.

--
Matus UHLAR - fantomas, uh...@fantomas.sk ; http://www.fantomas.sk/
Warning: I wish NOT to receive e-mail advertising to this address.
Varovanie: na tuto adresu chcem NEDOSTAVAT akukolvek reklamnu postu.
Save the whales. Collect the whole set.
___
Postfix-users mailing list -- postfix-users@postfix.org
To unsubscribe send an email to postfix-users-le...@postfix.org


[pfx] Re: Duplicated emails due to split by rcpt and OOM

2025-04-05 Thread Viktor Dukhovni via Postfix-users
On Thu, Apr 03, 2025 at 10:17:22PM +0200, Dmytro Alieksieiev via Postfix-users 
wrote:

> I faced issue with splitting emails via local filter and
> destination_recipient_limit = 1 under low memory condition which
> result in duplicated emails in queue.

Not duplicated, simply retained, if/when information that a recipient is
delivered did get recorded in the queue file.  What sort of "local
filter" are you talking about?

> 1. Postfix get email with multiple rcpts
> 
> 2. taking couple of rcpts (depending on settings, 3 by default) and creates
> new email in queue per each recipient

Why? Postfix does not by default do that, unless you configure
self-delivery.  What is the purpose of the internal delivery?

> 3. reach OOM

That's your actual problem.  Perhaps your filter uses unreasonable
memory, tune the filter and its concurrency, if that's a factor in
its memory requirements.

> 4. new emails stay in queue, but original email not get rcpt marked as
> "done", as result on next queue rerun they will be spitted again resulting
> in duplicated messages

What new emails?  Your problem description is much too fuzzy.

> Can be reproduced in docker with mem limit f.e. 100mb by send one email with
> 50 rcpts with next configuration:

With server systems routinely shipping with 256GB of RAM or more these
days, 100MB seems rather undersized.  Don't do that.

> 1. Extra main.cf:
> 
> smtpsplit_destination_recipient_limit = 1

What and why is "smtpsplit"?

> 2. Extra master.cf:
> 
> smtpsplit  unix  -   -   n   -   -   smtp
> -o syslog_name=postfix-split
> -o smtp_send_xforward_command=yes
> -o disable_mime_output_conversion=yes
> -o smtp_generic_maps=
> -o smtp_header_checks=

What is the intent here?  You should probably set a lower process limit
on this service.

> 127.0.0.1:5025 inet n  -   n   --  smtpd
> -o syslog_name=postfix-after-split
> -o content_filter=
> -o cleanup_service_name=cleanupsplit
> -o receive_override_options=no_unknown_recipient_checks
> -o smtpd_helo_restrictions=
> -o smtpd_client_restrictions=
> -o smtpd_sender_restrictions=
> -o smtpd_relay_restrictions=
> -o smtpd_recipient_restrictions=permit_mynetworks,reject
> -o mynetworks=127.0.0.0/8
> -o smtpd_authorized_xforward_hosts=127.0.0.0/8
> -o smtpd_tls_security_level=none
> -o smtpd_data_restrictions=

You may be processing each message through your milters twice.

-- 
Viktor.
___
Postfix-users mailing list -- postfix-users@postfix.org
To unsubscribe send an email to postfix-users-le...@postfix.org


[pfx] Re: Help with SASL Authentication Using /etc/sasldb2 in Postfix

2025-04-05 Thread Wietse Venema via Postfix-users
wouldsmina via Postfix-users:
> Hello,
> 
> I am facing an issue with configuring Postfix to use /etc/sasldb2. I have
> already set up SASL authentication, but authentication only works if
> /etc/sasldb2 is included in the $FILE variable into
> /usr/lib/postfix/configure-instance.sh file.

That looks like a Debian-specific script that populates 
/var/spool/postfix with copies of system files.

You can avoid that by using something like:

/etc/sasl2/smtpd.conf
pwcheck_method: auxprop
auxprop_plugin: sasldb

The actual pathname of the smtpd.conf file may differ.

Wietse
___
Postfix-users mailing list -- postfix-users@postfix.org
To unsubscribe send an email to postfix-users-le...@postfix.org


[pfx] Re: smtps starttls

2025-04-05 Thread Viktor Dukhovni via Postfix-users
On Thu, Apr 03, 2025 at 12:40:02PM +0200, Paul Neuwirth via Postfix-users wrote:

> that makes totally sense.
> 
> I am struggling with that scanner:
> https://siwecos.de/en/app
> 
> it says:
> 
> for
> 
> SMTPS Scanner [SMTPS_TLS]
> SMTP MSA(STARTTLS) Scanner [SMTP_MSA_TLS]
> SMTP(STARTTLS) Scanner [SMTP_TLS] 
> "TLS not supported"

Congratulations, you've found another useless scanner on the Internet,
why do you care???

> for mail.swabian.net

All you need to know is that your TLS support is just fine:

$ posttls-finger -cF /etc/ssl/cert.pem -lsecure -Lsummary 
"[mail.swabian.net]:587"
posttls-finger: Verified TLS connection established
to mail.swabian.net[80.152.201.148]:587: TLSv1.3 with
cipher TLS_AES_256_GCM_SHA384 (256/256 bits)
key-exchange x25519
server-signature RSA-PSS (2048 bits)
server-digest SHA256

$ posttls-finger -cwF /etc/ssl/cert.pem -lsecure -Lsummary 
"[mail.swabian.net]:465"
posttls-finger: Verified TLS connection established
to mail.swabian.net[80.152.201.148]:465: TLSv1.3 with
cipher TLS_AES_256_GCM_SHA384 (256/256 bits)
key-exchange x25519
server-signature RSA-PSS (2048 bits)
server-digest SHA256

> I have no idea, what version or cipher they test. that's why I tried
> with openssl s_client and I cannot find any issues. :-(
> 
> contacted their support, but doubt I'll get a reply.

Just ignore them.

-- 
Viktor.
___
Postfix-users mailing list -- postfix-users@postfix.org
To unsubscribe send an email to postfix-users-le...@postfix.org


[pfx] Re: smtps starttls

2025-04-05 Thread Matus UHLAR - fantomas via Postfix-users

On Thu, Apr 03, 2025 at 11:16:15AM +0200, Paul Neuwirth via
Postfix-users wrote:
> after an upgrade of postfix I am checking my config.
>
> I am in the impression, this should work:
> # openssl s_client -connect mail.swabian.net:465 -starttls smtp
> -tls1_3 CONNECTED(0003)
> Didn't find STARTTLS in server response, trying anyway...
> write:errno=32



On Thu, 3 Apr 2025 11:22:24 +0200
"Herbert J. Skuhra via Postfix-users"  wrote:

Either remove '-starttls smtp' or use port 587.


On 03.04.25 11:29, Paul Neuwirth via Postfix-users wrote:

ok, that means I was wrong with the impression, that this should work
on port 465.


"-starttls smtp" should work on ports 25/587 where plaintext is default.
(587 should require starttle before it allows anything).

on port 465, TLS should be the default and starttls should not make sense.


--
Matus UHLAR - fantomas, uh...@fantomas.sk ; http://www.fantomas.sk/
Warning: I wish NOT to receive e-mail advertising to this address.
Varovanie: na tuto adresu chcem NEDOSTAVAT akukolvek reklamnu postu.
He who laughs last thinks slowest.
___
Postfix-users mailing list -- postfix-users@postfix.org
To unsubscribe send an email to postfix-users-le...@postfix.org


[pfx] spamhaus/abusix in rspamd or postfix

2025-04-05 Thread lutz.niederer--- via Postfix-users
Hi!

spamhaus and abusix have a query limit for free accounts. spamhaus does not 
seem to be a problem but abusix is 5000 queries/day. postfix as well as rspamd 
can query them.

postfix or rspamd: which one should be used to query them and why?

Thanks!
-lutzn

___
Postfix-users mailing list -- postfix-users@postfix.org
To unsubscribe send an email to postfix-users-le...@postfix.org


[pfx] Duplicated emails due to split by rcpt and OOM

2025-04-05 Thread Dmytro Alieksieiev via Postfix-users

Hi Postfix community,

I faced issue with splitting emails via local filter and 
destination_recipient_limit = 1 under low memory condition which result 
in duplicated emails in queue.


From what I can see in logs:

- warning: process /usr/libexec/postfix/smtp pid 1039 killed by signal 9
- warning: private/smtpsplit socket: malformed response
- warning: /usr/local/libexec/postfix/smtpd: bad command startup -- throttling

From what I observe it goes as:

1. Postfix get email with multiple rcpts

2. taking couple of rcpts (depending on settings, 3 by default) and 
creates new email in queue per each recipient


3. reach OOM

4. new emails stay in queue, but original email not get rcpt marked as 
"done", as result on next queue rerun they will be spitted again 
resulting in duplicated messages


I expect that when email get queued as new email original email receive 
250 OK and mark email as "done", but looks like it not the case.


Maybe somebody know a secure way to avoid such situation? Or only 
possible way it to never face OOM, which honestly can be complicated at 
traffic spikes like email bombing, etc.


Can be reproduced in docker with mem limit f.e. 100mb by send one email 
with 50 rcpts with next configuration:


1. Extra main.cf:

milter_default_action = tempfail
smtpd_milters = inet:rspamd-proxy:11332
header_checks = regexp:/etc/postfix/header_checks
milter_header_checks = regexp:/etc/postfix/milter_header_checks
smtp_header_checks = regexp:/etc/postfix/smtp_header_checks
smtpsplit_destination_recipient_limit = 1

2. Extra master.cf:

smtpsplit  unix  -   -   n   -   -   smtp
-o syslog_name=postfix-split
-o smtp_send_xforward_command=yes
-o disable_mime_output_conversion=yes
-o smtp_generic_maps=
-o smtp_header_checks=
127.0.0.1:5025 inet n  -   n   --  smtpd
-o syslog_name=postfix-after-split
-o content_filter=
-o cleanup_service_name=cleanupsplit
-o receive_override_options=no_unknown_recipient_checks
-o smtpd_helo_restrictions=
-o smtpd_client_restrictions=
-o smtpd_sender_restrictions=
-o smtpd_relay_restrictions=
-o smtpd_recipient_restrictions=permit_mynetworks,reject
-o mynetworks=127.0.0.0/8
-o smtpd_authorized_xforward_hosts=127.0.0.0/8
-o smtpd_tls_security_level=none
-o smtpd_data_restrictions=
cleanupsplit  unix  n   -   n   -   0   cleanup
-o header_checks=

3. /etc/postfix/header_checks and /etc/postfix/smtp_header_checks content:

/^x-postfix-split:.*/   STRIP

4. /etc/postfix/milter_header_checks content:

/^x-postfix-split:\s+required/ FILTER  smtpsplit:127.0.0.1:5025

5. Configure Rspamd or other milter to add "X-Postfix-Split: Required"

--
Thanks,
Dmytro Alieksieiev
___
Postfix-users mailing list -- postfix-users@postfix.org
To unsubscribe send an email to postfix-users-le...@postfix.org


[pfx] Re: postfix problem with resources

2025-04-05 Thread natan via Postfix-users

Hi
In old server i have:
max_use=180

in new serwer I have:
max_use=150

I don't know if it's that important
I have no idea what could be wrong.
after restarting postfix everything seems to be back to normal so I 
don't know



W dniu 3.04.2025 o 10:32, natan via Postfix-users pisze:

Hi
Today I have this same problem with resources... for 4 houers works 
fine and I get many



Apr  3 10:07:07 node2 postfix/smtpd[426447]: warning: connect to 
private/policy-spf: Resource temporarily unavailable
Apr  3 10:07:07 node2 postfix/smtpd[426447]: warning: problem talking 
to server private/policy-spf: Resource temporarily unavailable



This is debian11 (systemd)
Linux node2 5.10.0-34-amd64
This machine is ber metal (64 vcory+ 128GB ram)
Load average: 0.20 1.36 3.10
ii  postfix    3.5.25-0+deb11u1 amd64

nobody   4125051  0.0  0.0  23216 16628 ?    Ss   09:48   0:00 
/usr/bin/python3 /usr/bin/policyd-spf


#cat /proc/4125051/limits
Limit Soft Limit   Hard Limit   Units
Max cpu time  unlimited unlimited    seconds
Max file size unlimited unlimited    bytes
Max data size unlimited unlimited    bytes
Max stack size    8388608 unlimited    bytes
Max core file size    0 unlimited    bytes
Max resident set  unlimited unlimited    bytes
Max processes 515047 515047   processes
Max open files    524288 524288   files
Max locked memory 65536 65536    bytes
Max address space unlimited unlimited    bytes
Max file locks    unlimited unlimited    locks
Max pending signals   515047 515047   signals
Max msgqueue size 819200 819200   bytes
Max nice priority 0    0
Max realtime priority 0    0
Max realtime timeout  unlimited unlimited    us


The difference between the new and old server is:
New:
-debian11 (systemd)
-in main.cf I add dkim:
smtpd_milters = inet:localhost:12301
non_smtpd_milters = inet:localhost:12301
milter_default_action = accept
milter_protocol = 6

Old serwer
-debian10(init.d)


In old debian10 (sysinit) all works fine and never get "Resource 
temporarily unavailable"

Both servers do not have selinux

Both servers have apparmor (distribution)
node2:~# journalctl | grep -i apparmor
Find nothing

node2:~# aa-status
apparmor module is loaded.
11 profiles are loaded.
10 profiles are in enforce mode.
   /usr/bin/freshclam
   /usr/bin/man
   /usr/sbin/clamd
   /usr/sbin/unbound
   lsb_release
   man_filter
   man_groff
   nvidia_modprobe
   nvidia_modprobe//kmod
   tcpdump
1 profiles are in complain mode.
   /usr/sbin/chronyd
4 processes have profiles defined.
2 processes are in enforce mode.
   /usr/bin/freshclam (1345)
   /usr/sbin/clamd (1144)
2 processes are in complain mode.
   /usr/sbin/chronyd (1374)
   /usr/sbin/chronyd (1389)
0 processes are unconfined but have a profile defined.


And Realy I dont have any problem whats going on

I must mention that my traffic is about 40 requests per second



W dniu 27.03.2025 o 20:32, Wietse Venema via Postfix-users pisze:

natan via Postfix-users:

W dniu 27.03.2025 o?14:44, Wietse Venema via Postfix-users pisze:

natan via Postfix-users:

Hi
I dont have selinux

Hi
I dot have and use apparmor to

Then perhaps your kernel does not have enough memory for file
handles, network connections, etc. Is this some kind of a 'minimal'
system like some people love to use for containers?

Wietse
___
Postfix-users mailing list -- postfix-users@postfix.org
To unsubscribe send an email to postfix-users-le...@postfix.org


--

___
Postfix-users mailing list -- postfix-users@postfix.org
To unsubscribe send an email to postfix-users-le...@postfix.org


--

___
Postfix-users mailing list -- postfix-users@postfix.org
To unsubscribe send an email to postfix-users-le...@postfix.org


[pfx] smtps starttls

2025-04-05 Thread Paul Neuwirth via Postfix-users
Hello,
after an upgrade of postfix I am checking my config.

I am in the impression, this should work:
# openssl s_client -connect mail.swabian.net:465 -starttls smtp -tls1_3
CONNECTED(0003)
Didn't find STARTTLS in server response, trying anyway...
write:errno=32
---
no peer certificate available
---
No client certificate CA names sent
---
SSL handshake has read 0 bytes and written 33 bytes
Verification: OK
---
New, (NONE), Cipher is (NONE)
This TLS version forbids renegotiation.
Compression: NONE
Expansion: NONE
No ALPN negotiated
Early data was not sent
Verify return code: 0 (ok)
---

but as you see, it doesn't - on port 25 it is working fine.

in master.cf:
smtpsinet  n   -   n   -   10   smtpd
  -o smtpd_tls_wrappermode=yes -o smtpd_sasl_auth_enable=yes
  -o smtpd_tls_security_level=encrypt

i think those are relevant from main.cf:
smtpd_tls_security_level =
smtpd_tls_wrappermode = no
smtpd_use_tls = no

what did I miss?

thank you
___
Postfix-users mailing list -- postfix-users@postfix.org
To unsubscribe send an email to postfix-users-le...@postfix.org


[pfx] Re: A question about the configuration of postscreen

2025-04-05 Thread Viktor Dukhovni via Postfix-users
On Sun, Apr 06, 2025 at 01:18:14AM +0200, Andreas Kuhlen via Postfix-users 
wrote:

> For better readability once more. Sorry for the first post which was a
> bit confusing because of its format!

This variant is not much better, at least not its text/plain variant,
only the HTML is correctly formatted.  Please avoid HTML in posts to
this list..

> I have a question regarding the configuration of postscreen. In my
> current master.cf file I have not allowed SASL authentication for SMTP:
> 

> I have a question regarding the configuration of postscreen. In my
> current master.cf file I have not allowed SASL authentication for SMTP:

> smtp inet n - y - - smtpd
>-o smtpd_sasl_auth_enable=no

You should leave essentially unchanged, just replacing "inet" with
"pass".  It is still smtpd(8) that might or might not do SASL, the
postscreen(8) service never implements SASL.

> To activate postscreen in the master.cf file I added the following passage:

> smtp  inet  n   -   y   -   1   postscreen
> smtpd pass  -   -   y   -   -   smtpd

Add
 -o smtpd_sasl_auth_enable=no

to the "smtpd pass" service.

> dnsblog   unix  -   -   y   -   0   dnsblog
> tlsproxy  unix  -   -   y   -   0   tlsproxy

> I commented out the upper part that prohibits SASL authentication for
> SMTP.

Changing the internal protocol from "inet" to "pass" does not
substantially change the fact that this smtpd(8) instance handles
incoming connections on port 25 after they're briefly inspected by
postscreen(8).

> Postscreen also works so far, the question remains whether the
> option line that prohibits SMTP authentication can also be set for
> postscreen?

See above.

> Like this:

> smtp  inet  n   -   y   -   1   postscreen
>-o smtpd_sasl_auth_enable=no
> smtpd pass  -   -   y   -   -   smtpd
> dnsblog   unix  -   -   y   -   0   dnsblog
> tlsproxy  unix  -   -   y   -   0   tlsproxy

No, the option goies "smtpd", just like its name implies.

-- 
Viktor.
___
Postfix-users mailing list -- postfix-users@postfix.org
To unsubscribe send an email to postfix-users-le...@postfix.org


[pfx] A question about the configuration of postscreen

2025-04-05 Thread Andy Kuhlen via Postfix-users
Hello list,

I have a question regarding the configuration of postscreen. In my
current master.cf file I have not allowed SASL authentication for SMTP:

smtp inet n - y - - smtpd -o smtpd_sasl_auth_enable=no

To activate postscreen in the master.cf file I added the following passage:

smtp inet n - y - 1 postscreen smtpd pass - - y - - smtpd dnsblog unix -
- y - 0 dnsblog tlsproxy unix - - y - 0 tlsproxy

I commented out the upper part that prohibits SASL authentication for
SMTP. Postscreen also works so far, the question remains whether the
option line that prohibits SMTP authentication can also be set for
postscreen? The configuration in main.cf does not need to be considered
here, I think.

Like this:

smtp inet n - y - 1 postscreen -o smtpd_sasl_auth_enable=no smtpd pass -
- y - - smtpd dnsblog unix - - y - 0 dnsblog tlsproxy unix - - y - 0
tlsproxy

___
Postfix-users mailing list -- postfix-users@postfix.org
To unsubscribe send an email to postfix-users-le...@postfix.org


[pfx] A question about the configuration of postscreen

2025-04-05 Thread Andreas Kuhlen via Postfix-users

For better readability once more. Sorry for the first post which was a bit 
confusing because of its format!

Hello list,

I have a question regarding the configuration of postscreen. In my
current master.cf file I have not allowed SASL authentication for SMTP:

*smtp inet n - y - - smtpd -o smtpd_sasl_auth_enable=no*

To activate postscreen in the master.cf file I added the following passage:

*smtp inet n - y - 1 postscreen smtpd pass - - y - - smtpd dnsblog unix 
- - y - 0 dnsblog tlsproxy unix - - y - 0 tlsproxy*


I commented out the upper part that prohibits SASL authentication for
SMTP. Postscreen also works so far, the question remains whether the
option line that prohibits SMTP authentication can also be set for
postscreen? The configuration in main.cf does not need to be considered
here, I think.

Like this:

*smtp inet n - y - 1 postscreen **-o smtpd_sasl_auth_enable=no*
*smtpd pass - - y - - smtpd dnsblog unix - - y - 0 dnsblog tlsproxy unix 
- - y - 0 tlsproxy*


___
Postfix-users mailing list -- postfix-users@postfix.org
To unsubscribe send an email to postfix-users-le...@postfix.org


[pfx] Re: Bests Database

2025-04-05 Thread John Hill via Postfix-users


On 3/31/25 3:22 AM, Matus UHLAR - fantomas via Postfix-users wrote:

On 29.03.25 17:41, John Hill via Postfix-users wrote:
In my last gig I ran Exchange, Oracle and various Debian servers. I 
retired and set up Postfix, bind9 etc to support my hobby domain. It 
keeps me involved.


I use Fail2Ban, lousy docs, good product. I upgraded from iptables to 
nftables. I have written a few regex filters passing variables to 
actions for nftables and Postfix that work rather well. I like 
MariaDB, not Oracle but, as I said, it keeps me involved.


I update nftables sets dynamically.  I did not like fail2ban 
reloading Postfix to update the access files.


Just FYI, this can be done with iptables as well, you just need to use 
ipsets which may need separate ipset command.


With the advice found here. I have changed access lists to lmdb. I 
had not used postmap on postscreen's cidr files previously. I read 
doing so would eliminate reloading on updates.


Used Ipsets a lot. Nftables will do the same thing, bu not quite as 
well. It loads updates to memory. I follow it with a rule set dump and 
write to a set table file for reloads.



Thx

--john

___
Postfix-users mailing list -- postfix-users@postfix.org
To unsubscribe send an email to postfix-users-le...@postfix.org