Postfix Queue Problem

[acrousset]

Me 2 i have this problem and i am tired to deal with my hosting provider that have no clues at all. (except restarting)
Always an email with no subject, no date that is in the mail queue that crashes all my mail services...
Something is wrong with Plesk 11 and this postfix / ctmilter issue...

ANY UPDATES FROM PLESK TO FIX THIS ONCE FOR ALL?

 

[H.W.B]

I have been analyzing this problem, and lately i noticed that there is a 0 bytes file in the tmp dir from the owner Postfix.

When i deleted that file, the problem was fixed!!!!

So there is a temp file creation in the TMP directory of the server.

Perhaps somebody at Parallels can now realy look in to it,

 

[Andrew_Pa]

Please can us provide exactly the solution you find, to give a try?
If we delete the folder, have we any problem?

I also confirm that the problem exist and it is very important!

 

[GravuTrad]

I have this problem too for few mails, but no file with 0 in size in the tmp directory...

 

[ScottGoddard]

Same Problem

Any update on this?

I have the same problem:

Every few days I get a message saying Server health parameter "Services > Mail server CPU usage" changed its status from "green" to "red". When I check the server the CPU usage has spiked massively (also the network bandwidth utilisation spikes at the same time???)

When I restart the system the usage is still high for a while but then settles down.

The only thing out of the ordinary I can see in the logs around the same time as this happens is ct-milter[1887]: open(/tmp//ctmiEiRCNb): Too many open files (Obviously the [1887] and ctmiEiRCNb part change each time but the rest is the same)

I have looked for the 0 byte file but cannot find anything

I have searched all over for a solution and this thread is the only mention of the problem I can find but no fix and no update since February :-(

Please help!

Server OS: CentOS 6.4 (Final) and Panel version: 11.5.30 Update #13 (upgraded from 10)

 

[g4marc]

same here!
since weeks, i try to find a solution for that!
i have this problem on a new and fresh server with PLESK admin-panel!
first, i thought, that something goes wrong during the installation-routine over my provider (strato), so, i installed everything again, but with the same result with postfix!

My machine:
-----------------------
BS Ubuntu 12.04.2 LTS
Panel-Version 11.0.9 Update #58, zuletzt aktualisiert am Sept 12, 2013 10:24 PM
Postfix version 2.9.6
-----------------------
i tried to set the firewall for postfix and DKIM, but nothings helps.
after e while (it takes a month, because it is my repalcemant machine, if something goes wrong on my other server. there is only one domain hosted), the smtp-server stops and i can't restart it, until i restart the whole machine.
only if i stop milter, i can start smtp-server, but if i restart milter after this, the smtp stops again.

here is the log:
------------------------------
warning: milter unix:/ctmilter/ctmilter.sock: can't read SMFIC_OPTNEG reply packet header: Connection timed out
warning: milter unix:/ctmilter/ctmilter.sock: read error in initial handshake
NOQUEUE: milter-reject: CONNECT from localhost[127.0.0.1]: 451 4.7.1 Service unavailable - try again later; proto=SMTP
------------------------------
it seems to me, there is an queue-overflow, that results on domainkeys-check...?
i have also the thing, that my server-condition/health goes up to red (over 3000% workload) when i resart the server...PSA shows it, at least...

is there a known bug in postfix 2.9.6 and should i update to 2.10.2 ?
i am wondering, that i can't find more users with the same issue.

my other server runs with 2.7.0 without any problems on same panel-version, but ubuntu 10.04.4

 

[g4marc]

i just have found this:
http://cafuego.net/2008/09/25/dkim-postfix

probably i have set the wrong port for milter/DKIM in my firewall...?

could somebody tell me, what are the right settings? port 53 on TCP is probably wrong...

 

[g4marc]

the problem is still there!
i am wondering, obviously this issue is not documented very well?
but i have it on a new installed server with plesk 11.05 and ubuntu 12.04!
also the load-problem of all services and CPU that is shown on PSA during the restart is everytime the same, even on a different hardware!

would be fine, to have support here!

 

[g4marc]

fine! since yesterday, i have the same problem on my other server with ubuntu 10.04!
error is:
postfix/smtpd[3098]: warning: milter unix:/ctmilter/ctmilter.sock: can't read SMFIC_OPTNEG reply packet header: Connection timed out
postfix/smtpd[3098]: warning: milter unix:/ctmilter/ctmilter.sock: read error in initial handshake

could somebody verify, that this problem occurs only with DIKM-akivation?
i deactivated this spamfilter in configuration 10 days ago on my other server, but i can't say exactly, that it works, because there is a lot less of traffic on this machine.

 

[g4marc]

is anybody out there? feels like all-round entertainment here =)

i have found this issue, if someboy cares...

there is a problem with the license-key, that is provided by my ISP!
obviously our key is not valid for "parallels Premium Outbound Antispam" but you can install the addon! logically, the configuration don't work. so i had uninstalled this tool, but the routine don't uninstall every part from that addon! you have to remove daily cronjob and this part in main.cf manually:

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

after this, postfix works, but i get this warning every 24 hours now:

/usr/lib/plesk-9.0/psa-pc-remote[11628]: PleskFilter: mi_stop=1
postfix/master[11818]: terminating on signal 15
postfix/master[12536]: daemon started -- version 2.7.0, configuration /etc/postfix
/usr/lib/plesk-9.0/psa-pc-remote[12296]: PleskFilter: mi_stop=1
postfix/smtpd[12626]: connect from unknown[unknown]
postfix/smtpd[12626]: lost connection after CONNECT from unknown[unknown]
postfix/smtpd[12626]: disconnect from unknown[unknown]
postfix/master[12536]: terminating on signal 15
postfix/master[12749]: daemon started -- version 2.7.0, configuration /etc/postfix

because of the autostart it isn't that big problem, but i suggest, that this warning has to do with the deinstallation of Outbound Antispam?

what i have to change, to eliminate this warning?


btw: it is a joke, that this essential addon "Parallels premium Outbount antispam" is not part of our licence-key! i can't find it at Paralles Product site anymore (Partner-products) even they sell it, we can't buy it, because Parallels only sell through CC. Our company don't have a CC and our ISP (STRATO) don't offer this addon!

our server was allready IP blocked from google, yahoo ASO, because of our clients contemt to have their anti-spam software up to date! the mailserver stops service, because of flooding and server IP was blacklisted. so nobody of our clients can send emails to google and it takes weeks, to get from the blacklist...it is annoying.

 

[dgtluc30]

+ 1 news for this problem ?

 

[ManuelK]

Hello

Got the same problem. Has this been fixed anyhow?

postfix/smtpd[14712]: connect from localhost[127.0.0.1]
postfix/smtpd[14712]: warning: connect to Milter service inet:127.0.0.1:12768: Connection refused
postfix/smtpd[14712]: NOQUEUE: milter-reject: CONNECT from localhost[127.0.0.1]: 451 4.7.1 Service unavailable - try again later; proto=SMTP
postfix/smtpd[14712]: disconnect from localhost[127.0.0.1]
postfix/smtpd[14712]: connect from static.5.228.76.144.clients.your-server.de[144.76.228.5]
postfix/smtpd[14712]: warning: connect to Milter service inet:127.0.0.1:12768: Connection refused
postfix/smtpd[14712]: NOQUEUE: milter-reject: CONNECT from static.5.228.76.144.clients.your-server.de[144.76.228.5]: 451 4.7.1 Service unavailable - try again later; proto=SMTP
postfix/smtpd[14712]: NOQUEUE: milter-reject: UNKNOWN from static.5.228.76.144.clients.your-server.de[144.76.228.5]: 451 4.7.1 Service unavailable - try again later; proto=SMTP
last message repeated 8 times


Thank you very much

kind Regards

 

[DimitriM]

+1 to this problem!

Config:
CentOS 6.5
Plesk 11.5.30

 

[SalvadorS]

Similar Issue here with a Debian 7.3 Box and Plesk 11.5.30 with all the microupdates

I fixed editing /etc/postfix/main.cf file comment these two lines:

smtpd_milters = , inet:127.0.0.1:12768
non_smtpd_milters = , inet:127.0.0.1:12768

I opened a post as I did not find this one: http://forum.parallels.com/showthre...-Service-unavailable-Problem-with-greylisting