[Logwatch-Devel] changes to sendmail service

Greg Matthews gmatt at nerc.ac.uk
Fri Oct 21 08:12:58 MST 2005


On Tue, 2005-10-18 at 22:21 -0700, Bjorn L. wrote:
> 
> Greg Matthews wrote:
> These two items ("WPO mailbox" in GroupWise, and unknown user in
> milter-ahead) are artifacts of the local sendmail configuration
> in your system.  At one point this type of site-specific matches
> were incorporated into the sendmail script, where they have been
> ever since (and are now mostly documented as such).

ok

> Then we added the Sendmail_MatchFilter, Sendmail_ReportFilter,
> and Sendmail_MilterHeadersToCount variables in the (currently)
> /usr/share/logwatch/default.conf/services/sendmail.conf file
> to deal with these local configuration issues.  So the log statements
> you see can be dealt with (counted, ignored, etc.) in your system
> any way you wish.  The intent is that all new log statements
> created by custom configuration of sendmail are handled with
> these three new variables.

I'll try to get my head around these and incorporate my stuff in there,
I'll let the list know if/when this gets done for milter-ahead.

> For the BlackHoles, I think that BlackHoles refers to the
> databases, and BlackHoled refers to the individual entries.  So
> I think that we want BlackHoled to be counted.  The detail level
> may need to be set to 10 to see all the entries.

yes but the total was always zero. With a higher detail level I could
see the total for a particular black hole but the grand total was always
zero. My change "fixed" that but its probably not 100% correct.

> > Can anyone explain the difference between "Rejected mail" and "Mail
> > Rejected" and why they cant be amalgamated?
> 
> The names may not be descriptive enough, but they've been like that
> for a long time.  Rejected mail means that the message was
> rejected due to local rules (such as access database).  Mail Rejected
> means that the message was rejected by the remote server during
> delivery.

ah ok. might be worth putting this in comments in the script?

> Thanks for the other corrections ("may be forged" string, missing
> fqdn, discard).  I've incorporated these.  And let me know if any
> of my other explanations don't match your expected output.

ok thanks. I'll let you know after I migrate to v7.

GREG

-- 
Greg Matthews           01491 692445
Head of UNIX/Linux, iTSS Wallingford



More information about the Logwatch-Devel mailing list