Postfix - antispam and relay package

biggsy

Hi Marcello,

Since updating to 2.10.2 pkg v.2.3.7 no mail relayed and seeing these messages in the log:

Dec  8 06:57:44 postfix/smtp[2973]: connect to 192.168.11.2[192.168.11.2]:25: Operation timed out
Dec  8 06:57:44 postfix/smtp[2973]: 7AC81B934: to=<me@my.domain>, relay=none, delay=1065, delays=1035/0.01/30/0, dsn=4.4.1, status=deferred (connect to 192.168.11.2[192.168.11.2]:25: Operation timed out)</me@my.domain>

Have restarted postfix many times and tried both "IPv4 only" and "all" options. Reinstalled, no change.

marcelloc

Can you post part of main.cf(interface section) and output of netstat -an | grep -i listen?

Pfsense 2.1?

biggsy

Thanks.

Yes, 2.1-RELEASE (amd64) built on Wed Sep 11 18:17:48 EDT 2013

Not sure what you mean by "interface section". Here is a sanitized version of main.cf:

/usr/pbi/postfix-amd64/etc/postfix/main.cf
#main.cf\
#Part of the Postfix package for pfSense
#Copyright (C) 2010 Erik Fonnesbeck
#Copyright (C) 2011-2013 Marcello Coutinho
#All rights reserved.
#DO NOT EDIT THIS FILE

mynetworks = /usr/pbi/postfix-amd64/etc/postfix/mynetwork_table
mynetworks_style = host
access_map_reject_code= 554
access_map_defer_code = 451
unverified_recipient_reject_code = 550
unknown_client_reject_code = 550
unknown_hostname_reject_code = 550
myhostname = <mail.my.domain>
bounce_queue_lifetime = 0s
postscreen_cache_retention_time = 30d
anvil_status_update_time = 1h
notify_classes = protocol
relay_domains = <first.domain> <second.domain>
transport_maps = hash:/usr/pbi/postfix-amd64/etc/postfix/transport
local_recipient_maps =
mydestination =
mynetworks_style = host
message_size_limit = 30000000
default_process_limit = 100
disable_vrfy_command = yes
strict_rfc821_envelopes = yes

#Just reject after helo,sender,client,recipient tests
smtpd_delay_reject = yes

# Don't talk to mail systems that don't know their own hostname.
smtpd_helo_required = yes
smtpd_helo_restrictions = check_helo_access pcre:/usr/pbi/postfix-amd64/etc/postfix/helo_check,
				reject_unknown_helo_hostname,
				reject_invalid_helo_hostname,
				reject_non_fqdn_helo_hostname,
				permit

smtpd_sender_restrictions = reject_non_fqdn_sender,
				reject_unknown_sender_domain,
				reject_unauth_pipelining,
				reject_multi_recipient_bounce,
				permit

# Allow connections from specified local clients and strong check everybody else.								
smtpd_client_restrictions = permit_mynetworks,
				reject_unauth_destination,
				check_client_access pcre:/usr/pbi/postfix-amd64/etc/postfix/cal_pcre,
				check_client_access cidr:/usr/pbi/postfix-amd64/etc/postfix/cal_cidr,
				reject_unknown_client_hostname,
				reject_unauth_pipelining,
				reject_multi_recipient_bounce,
				permit							

smtpd_recipient_restrictions = permit_mynetworks,
				reject_unauth_destination,
				reject_unauth_pipelining,
				check_client_access pcre:/usr/pbi/postfix-amd64/etc/postfix/cal_pcre,
				check_client_access cidr:/usr/pbi/postfix-amd64/etc/postfix/cal_cidr,
				check_sender_access hash:/usr/pbi/postfix-amd64/etc/postfix/sender_access,
				reject_non_fqdn_helo_hostname,
				reject_unknown_recipient_domain,
				reject_non_fqdn_recipient,
				reject_multi_recipient_bounce,
				reject_unverified_recipient,
				permit

inet_protocols = ipv4
inet_interfaces = <my.public.ip>
postscreen_disable_vrfy_command = yes
postscreen_non_smtp_command_enable = yes
postscreen_non_smtp_command_action = enforce
postscreen_pipelining_enable = yes
postscreen_pipelining_action = enforce
postscreen_bare_newline_enable = yes
postscreen_bare_newline_action = enforce
postscreen_greet_action = enforce
postscreen_access_list = permit_mynetworks,
			cidr:/usr/pbi/postfix-amd64/etc/postfix/cal_cidr
postscreen_dnsbl_action= enforce
postscreen_blacklist_action= enforce</my.public.ip></second.domain></first.domain></mail.my.domain>

and netstat output:


tcp4       0      0 <my.public.ip>.25       *.*                    LISTEN
tcp6       0      0 *.53                   *.*                    LISTEN
tcp4       0      0 *.53                   *.*                    LISTEN
tcp6       0      0 *.443                  *.*                    LISTEN
tcp4       0      0 *.443                  *.*                    LISTEN
tcp4       0      0 *.22                   *.*                    LISTEN
tcp6       0      0 *.22                   *.*                    LISTEN</my.public.ip>

biggsy

I've been running WireShark on the mail server. As expected, no connections to port 25.

May not be relevant but I just found these in syslog:


Dec  8 06:57:15 postfix/master[12027]: warning: ignoring inet_protocols parameter value change
Dec  8 06:57:15 postfix/master[12027]: warning: old value: "all", new value: "ipv4"
Dec  8 06:57:15 postfix/master[12027]: warning: to change inet_protocols, stop and start Postfix

marcelloc

That's why I have changed field info warning to restart postfix daemon after changing inet protocol or ip.

Netstat and config shows it listening on public address.

biggsy

@marcelloc:

That's why I have changed field info warning to restart postfix daemon after changing inet protocol or ip.

That's why I thought it might not be relevant. I have restarted many times. Even rebooted pfSense.

@marcelloc:

Netstat and config shows it listening on public address.

Yes, postfix is receiving emails but can't forward them to the mail server, it's just queuing them. That was working perfectly prior to the update.

Did you notice the "relay=none" in the log record? Prior to the update it was "relay=192.168.11.2[192.168.11.2]:25"

The relay host is set in the GUI for both domains to 192.168.11.2 and I can telnet 192.168.11.2 25 from an ssh session to pfSense without any problem.

biggsy

Netstat and config shows it listening on public address.

Now I see what you were saying.

I changed it to listen on loopback (instead of WAN) and it started delivering both queued and new emails.

I remembered you saying "choose at least wan loopback interfaces" when the package was very new but I have had postfix only listen on WAN since I first installed it. That was the only way I could get it to work before this latest update. I honestly don't remember the error I used to get if I listened on loopback.

Why would it need both WAN and loopback?

marcelloc

I've moved listening options from master.cf to main.cf on this latest update.

I'll test it here and update interface field info to warn about this behavior.

Thanks for your feedback.

biggsy

Without being restarted, is the postfix package somehow aware of changes in WAN IP?

It hasn't been a problem for me but I had thought about it this past couple of days.

marcelloc

@biggsy:

It hasn't been a problem for me but I had thought about it this past couple of days.

It will work fine if you listen on loopback and create a nat rule from wan address to 127.0.0.1

biggsy

@marcelloc:

It will work fine if you listen on loopback and create a nat rule from wan address to 127.0.0.1

no
Did you mean to say "nat rule"? I've never had a NAT - just a rule: pass any source in to destination WAN net on port 25.

Although, I'm not using postfix for outbound.

marcelloc

This is a common setup while using carp.

You configure the package(squid,snort,dansguardian,etc) to listen only on 127.0.0.1 and then create a nat rule to forward lan,wan,etc Carp address to 127.0.0.1.
This way you keep package running on both boxes without missing inet address startup errors.

If you configure this way on your wan, every time it changes ip, postfix will keep working as it's listening on lo0 and pfsense will keep the forward from wan address to it.

biggsy

Thank you, Marcello. Your explanation was very instructive and helpful.

I guess I've never had a problem because I think the only time my cable IP has changed was when I rebooted pfSense - and that's very infrequent.

elementalwindx

Wish OP would update the original post with newest info so I dont have to go thru 28 pages of whatever to see what the newest changes/versions/etc are of this package…

nekron

Hi …

I've some problem with Postfix Forwarder, it was installed on a new 2.1-RELEASE Install, the previous Version on a 2.1-BETA was functional (due HDD Crash i don't know the version

220 pfsense.xx ESMTP Postfix
HELO xxx
250 pfsense.xxx
MAIL FROM: <michael@xxx>
250 2.1.0 Ok
RCPT TO: <michael@xxx>
450 4.1.1 <michael@xxxx>: Recipient address rejected: unverified address: connect to 172.16.1.254[172.16.1.254]:25: No route to host</michael@xxxx></michael@xxx></michael@xxx>

172.16.1.254 port 25 is reachable from the pfsense-Box !

Best regards,
michael

expert_az

Hello marcelloc,

I'm getting these errors when attempting to query from search mail tab.

instead of query results getting lines below.This happening only query with todays date(15/01/2014),when selecting another date it's working ok.

Warning: sqlite_open(): file is encrypted or is not a database in
/usr/local/www/postfix.php on line 544 Warning: sqlite_query() expects parameter 1 to be resource, string given in
/usr/local/www/postfix.php on line 606 Warning: sqlite_fetch_array() expects parameter 1 to be resource, null given in
/usr/local/www/postfix.php on line 610 Warning: sqlite_fetch_array() expects parameter 1 to be resource, null given in
/usr/local/www/postfix.php on line 610 Warning: sqlite_fetch_array() expects parameter 1 to be resource, null given in
/usr/local/www/postfix.php on line 610 Warning: sqlite_fetch_array() expects parameter 1 to be resource, null given in
/usr/local/www/postfix.php on line 610 Warning: sqlite_fetch_array() expects parameter 1 to be resource, null given in

in addidion to this problem i'm getting error below when starting,restrating postfix.

php: /status_services.php: The command '/usr/local/etc/rc.d/postfix.sh stop' returned exit code '1', the output was '/usr/pbi/postfix-amd64/sbin/postconf: warning: /usr/pbi/postfix-amd64/etc/postfix/master.cf: unused parameter: user=postfix'

mauricioniñoavella

Hello friends, this is me showing this error. in Package Version 2.10.2 pkg v.2.3.7

I confirm if Postfix - antispam and relay package, can confirgurar with tls

this is the error

February 3 12:23:57 postfix / smtp [40769]: instead, send to (port submission/587) WITH STARTTLS
February 3 12:23:57 postfix / smtp [40769]: CLIENT wrappermode (port smtps/465) is unimplemented

So I have add to main.cf:

smtp_sasl_auth_enable = yes
relayhost = [smtp.xxx.xx]: 465
smtp_sasl_password_maps = hash :/ etc / postfix / sasl_passwd
smtp_sasl_tls_security_options = noanonymous
smtp_tls_security_level = encrypt
smtp_tls_mandatory_protocols =

Thanks for you colaboration.

Mauricio Niño Avella.

ics

Hi,

I have multifuntion printers used for scan to email.
Is it possible to configure postfix/postscreen to not reply "service currently unavailable" at the first connection of these MFP without allowing them to relay to internet ?
I mean I don't want to add the IPs in the MyNetworks list because it also allows these IPs to relay to internet.

Thanks

yaboc

hi Marcello

using this great tool and having some issues with the some good mails discarded by header checks because of mis-configured sender's server.

Feb 12 15:57:11 postfix/smtpd[10780]: connect from ms1.cernerasp.com[159.140.193.46]
Feb 12 15:57:11 postfix/smtpd[10780]: NOQUEUE: reject: RCPT from ms1.cernerasp.com[159.140.193.46]: 550 5.7.1 <taspmomail01.cmsext.com>: Helo command rejected: Host not found; from= infoviewreport@hvhc.orgto= me@mydomain.comproto=ESMTP helo= <taspmomail01.cmsext.com>Feb 12 15:57:16 postfix/smtpd[10780]: disconnect from ms1.cernerasp.com[159.140.193.46]

is there an easy workaround without disabling the spam checks? thanks

yaboc</taspmomail01.cmsext.com>/me@mydomain.com /infoviewreport@hvhc.org</taspmomail01.cmsext.com>

biggsy

Best thing to do is have the admin for that mail server fix the DNS entry but you could try adding this to your Access Lists, Header section:

/^From:.*@hvhc.org/ OK