Scalix Forums

Ok, first of all, I have fully read this topic that says how it should work:
http://www.scalix.com/community/viewtop ... 1833#11833

But I'm still getting problems!
I'm running RedHat FC4. I installed Mailman using the rpm via 'yum install mailman' and followed all the instructions in /usr/share/doc/mailman-2.1.8/INSTALL.REDHAT for post-install tasks. I created the mailing list, I received the mailing list emails saying I'm subscribed etc.

My problem is when I'm sending to the list, the email is getting stuck in the outbound queue saying "Deferred: Connection refused by domain.com."

So, I search the forums etc, and I find several responses saying thus:
"You need to have sendmail running as a daemon listening on 127.0.0.1:25."
http://www.scalix.com/community/viewtopic.php?p=807#807

Well, I have it set to running, and I can send mail, etc so I'm confused.

One possibility that I can think of is that in smtpd.conf I have this option set:

The reason I have this on is because I am using Mailscanner (Installed per instructions in this post: http://www.scalix.com/community/viewtop ... =7614#7614 .) Namely:

3) Leave Scalix as the primary SMTP listener and set SMTPFILTER=TRUE in smtpd.cfg.

So, my thought and question is, am I having this problem because of the SMTPFILTER setting? And if so, how can I get around this?

Thanks for the great email server and all your help!!!

Mito


PS. Also of note, I've always had a problem where all my mail is being received from 127.0.0.1 and I believe it is also due to this setting. If you know how to fix that as well, I'd love to hear it!

mito wrote:My problem is when I'm sending to the list, the email is getting stuck in the outbound queue saying "Deferred: Connection refused by domain.com."

Mmm ... weird. We will have to debug this step-by-step, then. It definatedly works for me on the same machine.

Run the following for me:

That should return something looking like the following. Check the hostnames are correct:


When you say "outbound queue" I assume you mean the sendmail mail spool, accessible via mailq. Also, is "domain.com" your domain or a user you're trying to send to's domain?

Try running where <ID> is the ID in the mailq command, and post the sendmail logs here for me.

mito wrote:PS. Also of note, I've always had a problem where all my mail is being received from 127.0.0.1 and I believe it is also due to this setting. If you know how to fix that as well, I'd love to hear it!

Can you post an example log of this as well?

Thanks

Valerion wrote:

mito wrote:My problem is when I'm sending to the list, the email is getting stuck in the outbound queue saying "Deferred: Connection refused by domain.com."

Mmm ... weird. We will have to debug this step-by-step, then. It definatedly works for me on the same machine.

Run the following for me:

Code: Select all

lsof -i:25

That should return something looking like the following. Check the hostnames are correct:

Code: Select all

COMMAND   PID USER   FD   TYPE DEVICE SIZE NODE NAME
sendmail 3572 root    4u  IPv4   8236       TCP localhost.localdomain:smtp (LISTEN)
omsmtpd  3774 root    4u  IPv4   8599       TCP <your-host-name>:smtp (LISTEN)

When you say "outbound queue" I assume you mean the sendmail mail spool, accessible via mailq. Also, is "domain.com" your domain or a user you're trying to send to's domain?

Try running

Code: Select all

sendmail -qI<ID> -v

where <ID> is the ID in the mailq command, and post the sendmail logs here for me.

mito wrote:PS. Also of note, I've always had a problem where all my mail is being received from 127.0.0.1 and I believe it is also due to this setting. If you know how to fix that as well, I'd love to hear it!

Can you post an example log of this as well?

Thanks :)

First of all, thank you for replying and trying to help me out!!!

I'll start off with this, because it's easy:

mito wrote:PS. Also of note, I've always had a problem where all my mail is being received from 127.0.0.1 and I believe it is also due to this setting. If you know how to fix that as well, I'd love to hear it!

After I posted this, I decided to do some more searches for this problem, now with the thought that it is related to the SMTPFILTER=TRUE option (previously I hadn't connected the two in my searches). This brought up this result, which helps to prove my theory, but the resolution isn't something that is really acceptable, where each message gives and error but is delivered anyways.

http://www.scalix.com/community/viewtop ... 7305#17305

Which logs are you wanting output from for this so I can show you those? (I'm still rather new at this whole email admin thing)


Next up, output of lsof -i:25 :

There were also a ton of established connections after this. The 192.168.21.31 is a 2nd NIC I have in the server connected to my internal network. Yes I changed the hostname to mydomain but it did show the correct hostname of the server (email goes to mydomain.com and the hostname of the server is mail.mydomain.com).

Valerion wrote:When you say "outbound queue" I assume you mean the sendmail mail spool, accessible via mailq. Also, is "domain.com" your domain or a user you're trying to send to's domain?!

When I say outbound queue, I do mean the sendmail queue, I believe... (as I said, still kinda new). Its the outbound mail queue as reported by MailWatcher, the web-interface to MailScanner that I am using (which I believe is the sendmail outbound mail spool/queue). Also, it is my domain, so in this message instead of domain.com I changed to using mydomain.com to help alleviate some confusion.



You said you have this running on the same machine. Does that mean you just have Scalix and Mailman on the same machine, or Scalix, MailScanner, and Mailman all on the same machine? I ask because I think the problem is that I'm throwing MailScanner into the mix as well, which is requiring that SMTPFILTER=TRUE option, which I think is causing my problems. Again, I'm new to this though, and that could be totally off-base...

Thanks again for your help!!!
Mito

Mmm ... this is strange. Can you try doing a



and see what you get back? Also try it to the IP address instead of the hostname and see if you get the same back.

Also, when you do the sendmail -q, do a

and see what gets displayed there.

The combination as you have it should still work fine.

Thanks again for your help, and especially for the very fast reply!

Valerion wrote:Mmm ... this is strange. Can you try doing a

Code: Select all

telnet mydomain.com 25

Sure thing, here we go! (Note that while I do have an internal-network card on my mail server, this is still connecting via the internet like everyone else...)



And here's the sendmail log. The result from sendmail -q was the same as before.




Thanks again!
Mito