Problem with Plesk SMTP service

[paddaone]

Hi,
I had the same problem and this fix works. Except that it comes back sometimes for unknown reason. I remove the unix... on both lines and it works again.
Why does it come back?

By default /etc/postfix/main.cf should consist following strings:
smtpd_milters = inet:localhost:12768
non_smtpd_milters = inet:localhost:12768

But looks like that your main.cf is modified:
smtpd_milters = inet:localhost:12768 unix var/spool/postfix/ctmilter/ctmilter.sock
non_smtpd_milters = inet:localhost:12768 unix var/spool/postfix/ctmilter/ctmilter.sock

To fix the issue you can try to remove unix var/spool/postfix/ctmilter/ctmilter.sock from /etc/postfix/main.cf and restart postfix. After that all should work fine.

 

[Nikolay.]

Why does it come back?

Do you have Outgoing Antispam extension installed? If so, I suggest removing it.

 

[paddaone]

Yep I do, removed it and it works.
Thanks for your help..

Do you have Outgoing Antispam extension installed? If so, I suggest removing it.

 

[RecreH]

Reopen this post because same issue is present now!!

By default /etc/postfix/main.cf should consist following strings:
smtpd_milters = inet:localhost:12768
non_smtpd_milters = inet:localhost:12768

But looks like that your main.cf is modified:
smtpd_milters = inet:localhost:12768 unix var/spool/postfix/ctmilter/ctmilter.sock
non_smtpd_milters = inet:localhost:12768 unix var/spool/postfix/ctmilter/ctmilter.sock

To fix the issue you can try to remove unix var/spool/postfix/ctmilter/ctmilter.sock from /etc/postfix/main.cf and restart postfix. After that all should work fine.

Hello:

Few months ago I was have the same issue, but prior patch was work perfectly!!, but now I'm having again the same issue (postfix smtp hangs).

I have see the main.cf and I have see the config lines has been modified!!, I have this:

smtpd_milters = , inet:127.0.0.1:12768 unix:/var/spool/postfix/ctmilter/ctmilter.sock
non_smtpd_milters = , inet:127.0.0.1:12768 unix:/var/spool/postfix/ctmilter/ctmilter.sock

Sintax is different, and have an comma before equal simbol. This work, but in few hours or one day, smpt hangs again!!!!!!!

I have try to restore the original fix lines as IgorG was write, but postfix smtpd can't work.

Please, how I can fix it now?.

Thanks!!.

 

[Lloyd_mcse]

I too had the extra comma at the beginning

eg:

smtpd_milters = , inet:127.0.0.1:12768

I also noted (while installing OpenDKIM) that the new default in Ubuntu 12.04 LTS is

/etc/default/opendkim

#SOCKET="local:/var/run/opendkim/opendkim.sock" # default

So one might assume that the default has changed for other milters?
Just a thought.

 

[VincentLippi]

smtpd_milters = , inet:127.0.0.1:12768 unix:/var/spool/postfix/ctmilter/ctmilter.sock
non_smtpd_milters = , inet:127.0.0.1:12768 unix:/var/spool/postfix/ctmilter/ctmilter.sock

Sintax is different, and have an comma before equal simbol. This work, but in few hours or one day, smpt hangs again!!!!!!!

I have try to restore the original fix lines as IgorG was write, but postfix smtpd can't work.

I've changed to this :

smtpd_milters = inet:localhost:12768 unix:/var/spool/postfix/ctmilter/ctmilter.sock
non_smtpd_milters = inet:localhost:12768 unix:/var/spool/postfix/ctmilter/ctmilter.sock

After, I restart in Plesk : Milter de Plesk (Postfix)
and the service : Serveur SMTP (Postfix) is green !

 

[Thomas F.]

So is this bug still not solved after all this time? This is ridiculous.

 

[jeanpaul]

i have the same problem with plesk parallels 12.
Oct 31 09:55:55 postfix/cleanup[12475]: warning: milter unix:/var/spool/postfix/ctmilter/ctmilter.sock: can't read SMFIC_BODY reply packet header: Connection timed out
milter-reject: END-OF-MESSAGE 4.7.1 Service unavailable - try again later;

 

[sosbilisim]

Cmilter eat cpu at %100 how to disable this ?

 

[jeanpaul]

i Think the problem comes from the packet Postfix 2.8.17

 

[zzz]

I have the same result on plesk 12.5 on centos 6.7

But my config is :

smtpd_milters = , inet:127.0.0.1:12768 unix:/var/spool/postfix/ctmilter/ctmilter.sock
non_smtpd_milters = unix:/var/spool/postfix/ctmilter/ctmilter.sock

Any ideas ?

 

[UFHH01]

Hi zzz,

to investigate your issue(s), please consider to add depending log - errors from your logs, add your current Plsk version ( incl. MU! ), your operating system and the MTA and it's version you use. It doesn't help much to "hijack" a thread, just because you didn't find a thread in the "right" forum. Plesk 11 is very different to Plesk 12.5 and solutions, tutorials and/or work-arounds may not at all fit your environment and resolutions may be absolutely useless for you.

 

[zzz]

Hi zzz,

to investigate your issue(s), please consider to add depending log - errors from your logs, add your current Plsk version ( incl. MU! ), your operating system and the MTA and it's version you use. It doesn't help much to "hijack" a thread, just because you didn't find a thread in the "right" forum. Plesk 11 is very different to Plesk 12.5 and solutions, tutorials and/or work-arounds may not at all fit your environment and resolutions may be absolutely useless for you.

Thank you for your reply

and Sorry for my "hijack" :-(

Consider my error :
" warning: milter unix:/var/spool/postfix/ctmilter/ctmilter.sock: read error in initial handshake "
And my config :
OS CentOS 6.7 (Final)
psa 12.0.18-cos6.build1200140606.15
postfix 2.8.17-14042513
courier-imap 4.15-cos6.build1200150114.15

Restarting postfix and postfix-milter don't resolve anything after few minutes the CPU of the messaging is over 100% and i need to restart the server to fix the issue.

Any ideas ?

 

[IgorG]

Any ideas ?

Try to use solution from http://kb.odin.com/en/117726

 

[zzz]

i had already tried :
ULIMIT="-n 4096"
with no result and i the alternative of disabling the code seems a little hard right ?

But if there is no other solution, i'm going to use it.

 

[UFHH01]

Hi zzz,

in some cases, you might have some standard CentOS settings, which will deny custom settings for non-root users for their open files limit. Please have a look at the files at "/etc/security/limits.d/" and check possible restrictions.

Please see as well your "limits.conf" and change the standard setting from 1024 to something more suitable, if you experience issues with the standard settings:

Before:

* hard nofile 1024
* soft nofile 1024

After:

* hard nofile 4096
* soft nofile 4096

 

[zzz]

Try to use solution from http://kb.odin.com/en/117726

In the file /etc/postfix/main.cf
i found :
smtpd_milters = , inet:127.0.0.1:12768 unix:/var/spool/postfix/ctmilter/ctmilter.sock
non_smtpd_milters = unix:/var/spool/postfix/ctmilter/ctmilter.sock

Should i comment this code ?

 

[zzz]

in this directory "/etc/security/limits.d/" i found the file :
90-nproc.conf
with this code :
# Default limit for number of user's processes to prevent
# accidental fork bombs.
# See rhbz #432903 for reasoning.

* soft nproc 1024
root soft nproc unlimited

Should i upgrade 1024 to 4096 ?

 

[UFHH01]

Hi zzz,

yes... that is what I suggested, if you would like to raise it for ALL users on your system. If you only would like to raise it for specific users, you have to define them with their username in front:

Example:

* soft nproc 1024
nginx         soft nproc          4096
apache        soft nproc          4096

In the file /etc/postfix/main.cf
i found :
smtpd_milters = , inet:127.0.0.1:12768 unix:/var/spool/postfix/ctmilter/ctmilter.sock
non_smtpd_milters = unix:/var/spool/postfix/ctmilter/ctmilter.sock

Should i comment this code ?

The KB states clearly:

If the issue persists, disable before-queue mail filters.

So please only disable it completely, if you can't solve your issue with other ways.