Συντάκτης: Bevis R W King Ημερομηνία: Προς: exim-users Υ/ο: W B Hacker Αντικείμενο: Re: [exim] AUTH failure for local cron job output
Bill (et al)
>Doesn't look like 100% an authenticator problem (yet).
>
>You do not appear to be submitting on a submisison port, so are seen as
>a possible peer MTA. Are you set up to allow that?
Huh? It should be going directly into EXIM via /usr/sbin/sendmail
(which should be the exim replacement).
OK, what's /usr/sbin/sendmail....? OK, /etc/alternatives/mta which is in
turn /usr/sbin/sendmail.sendmail (?!?!)
Ah, right, a security update RPM for sendmail installed on 9th of August
appears to have trampled on the link replacing sendmail with the exim
equivalent. So real sendmail is being called and is looping it back to
the SMTP port on localhost.
That would definitely explain things.... wonder why it trampled on
the /etc/alternatives/mta link....
The clues were in the increased logging which really did show an SMTP
call to local host going in. This is what it said:
2006-10-16 15:04:24 SMTP connection from [127.0.0.1]:48110
I=[127.0.0.1]:25 (TCP/IP connection count = 1)
2006-10-16 15:04:24 ignoring AUTH=amanda@??? from
H=localhost (gabrielle.bevteccom.co.uk) [127.0.0.1]:48110
I=[127.0.0.1]:25 (client not authenticated)
2006-10-16 15:04:24 remote host address is the local host:
gabrielle.bevteccom.co.uk (while verifying
<amanda@???> from host localhost
(gabrielle.bevteccom.co.uk) [127.0.0.1]:48110)
2006-10-16 15:04:24 H=localhost (gabrielle.bevteccom.co.uk)
[127.0.0.1]:48110 I=[127.0.0.1]:25 sender verify defer for
<amanda@???>: remote host address is the local
host
2006-10-16 15:04:24 H=localhost (gabrielle.bevteccom.co.uk)
[127.0.0.1]:48110 I=[127.0.0.1]:25 F=<amanda@???>
temporarily rejected RCPT <brwk@???>: Could not
complete sender verify
2006-10-16 15:04:24 H=localhost (gabrielle.bevteccom.co.uk)
[127.0.0.1]:48110 I=[127.0.0.1]:25 incomplete transaction (RSET) from
<amanda@???>
2006-10-16 15:04:24 SMTP connection from localhost
(gabrielle.bevteccom.co.uk) [127.0.0.1]:48110 I=[127.0.0.1]:25 closed by
QUIT
Anyway, it does look to be solved. I'll keep watching it but this is
definitely looking positive.