sendmail.cf: upgrade 10.0.5 to 11.0.4

[Franziska]

Hi

After upgrading I found diffenrences in the sendmail.cf file. In the "wiki" there is a advice, that the sendmail.cf will be replaced, because a new .m4 macro is used.

I checked the difference and found the following differences:
---
Old-sendmail.cf:
> C{G}scalix01.ari-armaturen.de
New-sendmail.cf:
> --- no entry ---
---
Old-sendmail.cf:
> DMari-armaturen.de
New-sendmail.cf:
> --- no entry ---
---
Old-sendmail.cf:
> Cwlocalhost scalix01.ari-armaturen.de
New-sendmail.cf:
> Cw
---
Old-sendmail.cf:
> # Scalix Mappers and Trusted User
> Komuser program /opt/scalix/bin/ommapsmtp
> Komxport program /opt/scalix/bin/ommapsmtp -x
> Tscalix
New-sendmail.cf:
> # Scalix LDAP mapper
Kldapsx ldap -h localhost -p 5757 -M simple
-T<TMPF> -1 -b "" -v scalixInstanceName -k (mail=%0)

---
Old-sendmail.cf:
### Scalix mailers
Mscalix, P=/opt/scalix/bin/unix.in, E=\n, F=PDFMXmnu, A=unix.in -s $h
Momxport, P=/opt/scalix/bin/xport.in, F=LMn, A=xport.in -s $h $u
New-sendmail.cf:
#######################################
### Scalix Mailer specification ###
#######################################

##### $Id: scalix.m4,v 1.1 2006/10/20 10:32:19 jch Exp $ #####

SScalixMapper
# do the LDAP lookup
R<$+><$+> $: <$(ldapsx $2 $: $)> <$1> <$2>

# if temporary failire, return original address for MTA to queue up
R<$* <TMPF>> <$+> <$+> $: $&{opMode} $| TMPF <$&{addr_type}> $| $2
R$* $| TMPF <$*> $| $+ $@ $3

# If matched by LDAP, resolve to the right Scalix mailer
R<$+> <scalix $*> <$+> $# scalix_xport $@ $1 $: scalix
R<$+> <$+> <$+> $# scalix_mime $@ $1 $: $2

# ... otherwise return the original address


Mscalix_mime, P=/opt/scalix/bin/unix.in, F=PDFMXmnu, E=\n,
A=unix.in -s $h
Mscalix_xport, P=/opt/scalix/bin/xport.in, F=LMn,
A=xport.in -s $h $u
---

So I knew, that the DM-macro is necessary for us, because we work with "transport"-Domains on our incoming mails.
(e.g.: *@ari-armaturen.de arrives on a external mailserver is switched to *@xxx.xxx.net and "re"switched to *@ari-armaturen.de on the Scalix-Server)
This worked fine in the past.

But when I put this macro into the new sendmail.cf and restarted sendmail, a problem occured:
The "relay" entry on the Scalix-server was not "scalix01" yet "xxx.xxx.net". And the server tried to send out the incoming mail. The sender got the replay, that his mail could not be delivered because the mailserver is not available. Of course...
(Example from /var/log/mail: Jun 10 02:13:24 scalix01 sendmail[1722]: l5A0DODb001712: to=<franziska.bee@shs.a
aimx.net>, delay=00:00:00, xdelay=00:00:00, mailer=esmtp, pri=123586, relay=mail
.shs.aaimx.net. [62.249.134.54]<-- "Wrong! Must be scalix01", dsn=2.0.0, stat=Sent (Ok: queued as 78AF1B37CE))

After a long time with try and error, I copied the old sendmail.cf back and restarted the sendmail. And see: It works.

---

In the FAQ-Wiki is the advice for the macro, but I did not found any documentation about changes, which are made bei the installer. Not good!!!

Now the question:
I want to know, where I can get a documentation for this changes?
The complexness and the difficult configuration of a sendmail needs this for the admin. You can not do anything here without publicising details.

[chris]

Hi Franziska,

By returning to the old sendmail.cf what you've basically done is avoided the new ldapmapper mechanism and switched back to the sendmail interface (which we've left in for legacy compatibility)

I assume that you set masquerade domain (DM macro) during server install?

Cwlocalhost scalix01.ari-armaturen.de is the macro for local-host-names.

If you ensure that this macro is set (or better add scalix.company.com to /etc/mail/local-host-names) you might have more luck.

I suspect you were seeing the relay=mail.shs.aaimx.net. because sendmail couldn't resolve scalix01? Is that possible?

I'm a bit confused as to what the actual problem is - can you describe what 1) doesn't work and what 2) should work?

Thanks,

Chris

PS: feel free to PM me with any config you can't sanitize.