Re: tracking down the cause of an entry in syslog?
On Sat, 23 Jun 2001 03:08:39 -0800, Ethan wrote:
> On Sat, Jun 23, 2001 at 07:45:31AM -0300, Henrique de Moraes Holschuh wrote:
> > On Sat, 23 Jun 2001, Paul Wright wrote:
> > > >From /var/log/ippl/all.log:
> > > Jun 23 02:02:02 ICMP message type destination unreachable...
> >
> > Destination unreachable due to host down is fine, but due to illegal port
> > isn't...
> >
> > > I'm going to read up on procmail in hopes of finding out what port it is
> > > looking for, etc. I'm not convinced it's a bug as yet, but then I'm not
> > > convinced it's not.
> >
> > procmail doesn't use sockets at all AFAIK, the problem lies probably with
> > masqmail.
>
> procmail tries to notify biff about the new mail it just delivered, it
> does this by connecting to comsat on localhost. no comsat -> ippl log
> entry. i haven't been able to find a way to disable this so i just
> told ippl to ignore it:
>
> ignore icmp type 3 from localhost
>
> > > You can check this by sending a mail to:somebody@localhost
> >
> > Well, I have a 'full blown' MTA in my system (postfix), so it will just
> > (try to) deliver the email to user "somebody" and mail me back a complain if
> > it cannot. No weird ICMP messages.
>
> biff = no in main.cf for postfix will stop it from doing these
> connections, but procmail is another matter.
>
> btw: i use postfix with a ~/.forward that pipes my mail to procmail.
> i don't have masqmail.
>
Thank you Ethan. Very cluefull of you.
-ptw
--
Paul T. Wright <paul@cvanet.com>
-currently seeking employment-
Reply to: