borderline offtopic: local mail delivery

[kirberich]

I have an odd situation with my server here and I would appreciate it if someone could explain it to me.

I have a server that downloads mail from an external mail server.

Up until a few days ago, when I sent a mail to a local scalix users e-mail address (user@externalserver with sendmail, the mail was delivered directly to scalix, because apparently the mail went through the scalix message router.
(This is the part where I don't exactly understands how the mail routing works, but I figured this is the normal behaviour)

But now, when I do the exact same thing, the mail gets delivered to the external mail server.

I don't recall changing anything that could have resulted in this behaviour, so what could be the cause of this change?

[btisdall]

Does the modification date on your sendmail.cf offer any clue?

[kirberich]

The file was modified yesterday, that's not very surprising though, because I executed and omsendin to rewrite the file.

I didn't change any configuration files though, so that shouldn't have done anything, right?

Anyway the problems are accumulating now, users don't appear to be able to send and recieve external mail anymore. The are 260 mails stuck in the mail gateway that don't get delivered

Mailwasher complained that sendmail refuses the connection, that stopped when I restarted sendmail though.

I guess it's probably something stupid I did without even realizing it, but I really can't figure out what it is.

[kirberich]

Here's the only error message I can find:

SERIOUS ERROR Omscan Server (Omscan Server ) 07.17.07 16:50:44
[OM 11170] Could not obtain exclusive access to a queue.
-> Q_SetTimeOut
<- Q_SetTimeOut
-> Q_SetTimeOut
<- Q_SetTimeOut
-> Q_SetTimeOut
<- Q_SetTimeOut
-> Q_SetTimeOut
<- Q_SetTimeOut
-> Q_SetTimeOut
<- Q_SetTimeOut
-> Q_SetTimeOut
<- Q_SetTimeOut
-> Q_SetTimeOut
<- Q_SetTimeOut
<- /build/11.1.0/src/bin/scn/scn_msg.c:384[100,11170]

[btisdall]

I'll bet your antispam config is the issue - can you take mailwasher out of the loop & see what happens?

[kirberich]

Ok, I commented those lines out of the sendmail configuration file

Code: Select all

INPUT_MAIL_FILTER(`mailwasher_server',    `S=unix:/var/run/mwserver/mpd.sock, F=T, T=S:4m;R:4m')
define(`confINPUT_MAIL_FILTERS', `mailwasher_server')


recompiled the sendmail.cf with m4 and executed omsendin, which only replied this, but if I'm not mistaken that's expected behaviour.

Code: Select all

MAIL_CREATE_CONFIG set to 'no' in /etc/sysconfig/mail
NOTE: /etc/mail/sendmail.cf unchanged.  Scalix lines already active.


I restarted sendmail, scalix and scalix-tomcat and voila!

The queue is starting to empty.

Thanks a lot, though I still need to figure out what exactly caused the problem.

[kirberich]

Ok so the queue is empty now, but people still don't recieve mails and I appear to have some kind of mail delivery loop.

Something is seriously wrong here.

[btisdall]

Have you removed the 'SMTPFILTER' line from your smtpd.cfg & restarted the smtp relay daemon?

[kirberich]

No, I forgot about that. With that changed, mails get delivered now. Thanks a lot btisdall, this would have been a lot worse without your help.

I guess I will need to wait with reintegrating mailwasher till everybody went home, still, do you have any idea where in the configuration I might need to check for problems?

This is going to be a long day.

[kirberich]

The strange problems continue.

I can now recieve mails just fine, but sending external mails does not work

I also see these errors in my mail.err

Code: Select all

Jul 18 14:26:08 mailserver sendmail[30613]: l6ICQ8fH030613: SYSERR(root): ldap_init/ldap_bind failed to localhost in map ldapsx: Unknown error 325
Jul 18 14:26:53 mailserver sendmail[30689]: l6ICQrLm030689: SYSERR(root): ldap_init/ldap_bind failed to localhost in map ldapsx: Unknown error 325
Jul 18 14:28:07 mailserver sendmail[30851]: l6ICS7Bv030851: SYSERR(root): ldap_init/ldap_bind failed to localhost in map ldapsx: Unknown error 325
Jul 18 14:31:09 mailserver sendmail[31067]: l6ICV90a031067: SYSERR(root): ldap_init/ldap_bind failed to localhost in map ldapsx: Unknown error 325
Jul 18 14:32:26 mailserver sendmail[31192]: l6ICWPH4031192: SYSERR(root): ldap_init/ldap_bind failed to localhost in map ldapsx: Unknown error 325
Jul 18 14:33:57 mailserver sendmail[31673]: l6ICXvqj031673: SYSERR(root): ldap_init/ldap_bind failed to localhost in map ldapsx: Unknown error 325
Jul 18 14:43:41 mailserver sendmail[32362]: l6ICS7Bv030851: SYSERR(root): ldap_init/ldap_bind failed to localhost in map ldapsx: Unknown error 325
Jul 18 14:43:59 mailserver sendmail[32409]: l6IChxRe032409: SYSERR(root): ldap_init/ldap_bind failed to localhost in map ldapsx: Unknown error 325


(Edit: That part at least I managed to get fixed, the ldapmapper had crashed. Maybe that was the problem all along.)

[jmason1182]

OK I concur. I had the same smtp error logs about the ldap problems. I didn't even know there was an ldap mapper.... so I did a

Code: Select all

service ldapmapper restart

.


Now everything works as far as Mailwasher & sendmail.

[chris]

The strange problems continue.
(Edit: That part at least I managed to get fixed, the ldapmapper had crashed. Maybe that was the problem all along.)

Hi kirberich, so sendmail is dependent on the ldapmapper to send email. The reason for that is that sendmail uses the ldapmapper to determine whether the email is destined for a local or for a remote user. If it can't make that decision because it can't talk to the ldapmapper, mail will just pile up in the queue waiting for the ldapmapper to come back online.

So yes, this by way of confirming that your diagnosis sounds correct.

Chris