Internet Mail gateway not processing messages

[easysoft]

Hi all,

We have a Scalix 10.0.1 installation on CentOS 4 which has been running very well for over a year now. I am using amavisd-new for spam filtering and clamd for virus scanning. Everything has been functioning perfectly for over a year now.

A few days ago we realised that outbound mail isn't being processed. Incoming mail is being correctly received. i didn't make any changes on the server.

1) Sendmail is running fine, and a command line "mail external@user.com" does indeed work. I can also telnet into the mail server on port 25 and issue MAIL-FROM and RCPT_TO commands fine.

2) All incoming mail is being received

3) All outbound mail is being held up in the Internet Mail Gateway queue - and this is indicated in the SAC. There are no errors in the IMG Logs. There are currently 2700 emails in the queue, which are displayed when I run omstat -q unix

4) I have run omcheck -s -d but there aren't any errors there

5) DNS is fine - including reverse DNS

I'm really stumped now. i have restarted Scalix (and even rebooted the whole server) but the outbound mail is still held up.

Any ideas on what might be wrong and what I can try to fix this ? Any help would be greatly appreciated.

Many thanks,

Arif Saleem

[easysoft]

Hi all,

Just some updates. I upgraded to Scalix 10.0.5 as I read about some Internet Mail Gateway problems that were fixed in 10.0.5. However the problem remains.

As you can see, IMG now has 2834 messages queued :

# /etc/init.d/scalix status | sort
Archiver Started 15:33:56 0
Background Search Service Started 15:33:56 0
Bulletin Board Server Started 15:33:56 0
CDA Server Started 15:33:56 0
Container Access Monitor Started NON-STOP
Database Monitor Started 15:33:56
Directory Relay Server Started 15:33:56
IMAP Server Daemon Started 15:33:56
Internet Mail Gateway Started 15:33:56 2834
Item Delete Daemon Started NON-STOP
Item Structure Server Stopped
LDAP Daemon Started 15:33:56
Licence Monitor Daemon Started NON-STOP
Local Client Interface Enabled 15:33:56 0
Local Delivery Started 15:33:56 0
Mime Browser Controller Started 15:33:56
Notification Monitor Started NON-STOP
Notification Server Started 15:33:56 0
Omscan Server Started 15:33:56 0
PC Monitor Started NON-STOP 0
POP3 interface Started 15:33:56 0
Print Server Started 15:33:56 0
Queue Manager Started NON-STOP
Remote Client Interface Enabled 15:33:56 6
Request Server Started 15:33:56 0
Service Router Started 15:33:56 0
Session Monitor Started NON-STOP
Shared memory daemon Started NON-STOP
*** shura
SMTP Relay Started 15:33:56
Test Server Started 15:33:56 0


Any help would really be appreciated.

[easysoft]

Hi All,

Still having this problem - though using omstat -q unix I can see there are a lot of 'missing messages' in the queue :

omstat : [SYS 2] No such file or directory
omstat : [OM 1209] Could not open a message file.
{many times}

and then :
6989843 jradcliff / internet MSG N New survey 21.08.07
etc

Is there some way I can just delete everything in the unix queue and let things start afresh ?

I can't list them all in the SAC to delete them - there are over 3000 now and the browser hangs when it tries to diplay them.

I have managed to start up omqdump but am a bit lost on how to use it to delete everything in the queue .... any help would be much appreciated.

Regards and thanks in advance,
Arif

[Valerion]

To delete messages off the queue, use D (not d) in omqdump

Try the following for me, please:

1) Try telnetting to port 25 on 127.0.0.1

2) run sendmail -bs
This will present you with a sendmail banner

do the following:

Code: Select all

EHLO localhost
MAIL FROM: <email address>
RCPT TO: <email address>
DATA
Subject: Test Mail

Testing ...
.
QUIT

Please preserve <> in the above. See if the mail gets correctly delivered. Also check if there are any errors in the sendmail log file (my machine it's /var/log/maillog).

[mikethebike]

Arif,

do not delete the messages from the queue, use the "M" (multiple move) option in omqdump to move them to the error queue. Then when you fix the problem you can move them back and not lose any messages.

Hav eyou looked in the fatal log? Are there any entries in there for the unix queue.

Mick

[easysoft]

Thanks for your suggestions .... but ...

Before reading these posts I had already used Z (zap) in omqdump to remove most of the messages in the unix queue. I restarted scalix after deleteing them, and then the remaining messages in the queue went through and now all outbound mail is going through normally.

I suppose I could have used the M command as suggested by Mick - maybe next time.
Thanks anyway for your help - it is greatly appreciated.

Regards
Arif

[computernay]

Just to add my two cents -
We had the same problem about a week ago. The Internet Mail Gateway queue seemed to be getting bigger and bigger. (around 5700 mails were in it before we caught it) I did some tracking using /var/opt/scalix/??/logs/audit and it appears that a spammer put our domainname on their spam. The servers on the other end (rightfully) detected it as spam and bounced it back to our server, but the bogus addresses the spammers used don't exist, so our server bounced it back to the other server, which detected it as spam, and so on and so forth. This continued for at least 24 hours before someone noticed our email wasn't getting sent. (actually, it was being delayed by about 12 hours waiting to get through the queue) I used SAC to delete most of the bounces going out. Once I got the queue down below 1000 or so the server was able to handle it and all was back to normal the next day. During this whole time, SAC showed 1 message being processed, so it appeared to be stuck, but really it just couldn't keep up with the load.
We use Scalix 11.1 on CentOS 5 x86_64, 3ghz P4, 1gb ram and one 320gb sata hard drive.

[christianb]

Hi

I've noticed several topics regarding the IMG queue on Scalix and would like to share our experience and see if this can help.

We've been using Scalix for over a year without problems. One month ago, it stopped sending outbound emails. All emails were received normally, and emails inside the domain were delivered normally.

Internet Mail Gateway queue was stuck as shown on SAC.

I noticed that restarting the server would work: the queue was emptied and would work fine for 5 to 10 hours. Than it would build up again. New restart worked.

I followed some tips from this post:

viewtopic.php?t=8941&highlight=queue

And ran omscan -A -a -f

It ran without errors, but the queue was still there. Than, at some point, I ran it after restarting, when the queue was empty, and it worked fine.

Since then, we've been one month without problems. But now it started again. I think the spam scenarion shown above might be the cause. The 5 - 10 hours working period seems to match a resend period from another server (it might try to send the email to us while the server is restarting, and since it can't contact it, the foreign server holds the bounce for a while for a retry.)

These stupid spammers use our domain and it seems some messages are in loop. But I have no technical skill to make sure...

Well, if that's the case, it seems a pretty easy way to bring down a Scalix server, isn't it? Is this kind of behavior universal to Mail servers? What can be done to avoid this?

I am about to install SpamAssassin, but wonder if this would solve the problems anyway, as emails don't seem to stop from entering and bouncing, they are just put on a Spam folder, right?

If anyone has any bright idea, please share it!

I've put replies on other posts and there's a suggestion to upgrade Scalix, but it doesn't seem the case:

viewtopic.php?t=9095&highlight=queue

Regards.