Internet Mail Gateway problem

[achan]

I'm new to Scalix and mail server.
After I have set the Scalix server for weeks. My mail server performance becomes slow. It may because I have setted the smtpd.cfg wrongly.

My users need to use my SMTP as the outgoing mail server from external.
Here is my new config of smtpd.cfg, should I configure as following:

#RELAY Log_Reject ALL
AUTH_MISMATCH reject ALL
AUTH_SUCCESS Accept ALL
AUTH_MISMATCH Reject LOCAL
AUTH_MISMATCH Reject ALL
ANONYMOUS Reject ALL

Also, when i'm using omstat -s, I found that there are many mails in "internet mail gateway", the result is here:

Service Router Started 09:36:07 0
Local Delivery Started 09:36:08 0
Internet Mail Gateway Started 09:36:08 338494
Local Client Interface Enabled 09:36:08 0
Remote Client Interface Enabled 09:36:08 1
Test Server Started 09:36:08 0
Request Server Started 09:36:08 0
Print Server Started 09:36:08 0
Bulletin Board Server Started 09:36:08 0
Background Search Service Started 09:36:08 0
CDA Server Started 09:36:08 0
POP3 interface Started 09:36:08 0
Omscan Server Started 09:36:08 0

I have tried to use omtidyallu command to delete all the outtray mails.
But the internet mail gateway still have large amount of mails, how can I clear the Internet mail gateway and improve the performance?

Thx

[ScalixSupport]

Hi. What are you trying to accomplish with the changes to the smtpd.cfg file? By default Scalix will only relay messages if the sender is either using AuthSMTP or if the sender is allowed by the RELAY accept lines.

Right now, you have a ton of messages backed up in your Internet Mail Gateway queue which may have come from someone spamming your site, someone on your site who is spamming the outside world or a looping message. If you type:

omstat -q unix

do the messages all have the same subject line?

Thanks,
Rachel

[achan]

hi, Rachel

I have tried "omstat -q unix", but the queue may be too long, as I have waited over half an hour, it still hasn't showed the result.
Can I sort the top 100 messages from the queue out, rather than show over 300,000 messages in the queue?

Thanks,
A.Chan

[ScalixSupport]

You should NEVER have a machine connected to the Internet acting as an open relay, this is BAD, BAD BAD.

The first thing you need to do is service sendmail stop. Then omoff -d0 unix.

This will stop people from connecting to your host which is what you want.

Next, you need to kill the queued messages. This can be done by scripting.

omstat -q unix > /tmp/msg.list

This will run for a loooong time. Come back when it is finished.

Cheers,

Sascha.

[jo_loop]

I have the same problem.
What is the next step?

[ScalixSupport]

Have you run omstat -q unix >/tmp/del.txt ?

Sascha.

[jo_loop]

yes i have ran it now.
Is there anything more i need to configure to prevent this from happening again?

[ScalixSupport]

Oh yes.

How many messages did you have in the queue? Reason I ask there are two ways to get rid of them, one involves scripting, the other does not.

Please post your smtpd.cfg here so we can fix it.

Cheers,

Sascha.

[jo_loop]

i hade 15000 messages in queue.
I have delete almost everyone of them.

Here is my smtpd.cfg


SMTPFILTER=TRUE
RELAY accept 127.0.0.1
RELAY accept 192.168.99.0
RELAY Log_Reject ALL

# extra rules added to prevent open relay usage
RECIPIENT Log_Reject *@*@*
RECIPIENT Log_Reject *%*
RECIPIENT Log_Reject *!*
RECIPIENT Log_Reject *#*@*

[jo_loop]

how do I disable bounce when a mail can not be deleivered because the address is missing?


/Johan

[ScalixSupport]

What MTA is listening on the interface connected to the Internet? sendmail, Postfix or Scalix SMTPD? You can tell by telnet'ing to it with the external IP Address on port 25.

Cheers,

Sascha.

[jo_loop]

Scalix smtp relay

[pL-Admin]

We are experiencing the very same problem.

Service Router 182902
Local Delivery 98120

and rising very quickly.

the configuration, scalix version and linux are unchanged for weeks

Where are the answers to the open questions in this thread? The one about disabling mail where address is missing and the one about the next step to kill all the damn msgs? Maybe this is only a test to see when our 22gb of free space run out?

regards
Peter & Marco

[ScalixSupport]

Use

omstat -q router >file1
omstat -q local >file2

This will give you a list of messages on a queue and their message id.

You can delete messages on the queue with:

omstat -q router -m message-id -j -R

Regards,
Don

[pL-Admin]

that would take quite some time with 200k msgs only in router ... and yes, I could write a script ...

we just injected QM_DONT_READ_MSG_AT_START=TRUE into general.cfg

and for the moment, this work-around seems to work

but .. where _are_ the 200k msgs now? gone with the wind?

regards,
Peter & Marco

/edit

Maybe QM_MAX_FAILURES would help, too (instead of the dont_read?)