[TriLUG] Sendmail oddities

Brian Henning brian at strutmasters.com
Wed Mar 22 13:40:37 EST 2006


Hi Gang,
   I think this is related to a buggered MailScanner install, but take a 
look and let me know if it sounds like a good hunch:

(FC3 server on relatively new hardware)

- During init, MailScanner tries to start but fails due to 
something-or-other.  Bad config, maybe; just haven't had time to futz 
with it.
- Later in init, sendmail starts with an [ OK ] result.

All seems well.  Furthermore:
from somewhere outside, telnet myserver.net 25 results in a normal 
conversation with sendmail, from MAIL FROM:<...> through DATA ... all 
the way to RSET and QUIT.  200-code success on every command.

HOWEVER...
The message from said entirely normal conversation seems, by all 
appearances, to be falling into the bit bucket--it does not end up in 
the recipient's mailbox.  Yet:

% service sendmail status
sendmail (pid blah blah blah) is running...

% service dovecot status
dovecot (pid blah blah blah) is running...

BUT:
when I go to the sendmail module of webmin on that server, the 
start/stop button at the bottom is showing the START option, not the 
STOP option, and once I click Start Sendmail in webmin, sendmail begins 
actually functioning correctly.

So:
- Why does sendmail indicate itself to be running to the service command 
(and the init script), and carry out valid conversations on port 25, if 
it's not really running [correctly]?
- Where do those messages go?
- And how come only webmin is alert enough to realize that sendmail is 
in fact /not/ running [properly], and is able to correct it?  What does 
it know that I don't? (obviously lots, but still...)

Thanks a ton,
~Brian


-- 
----------------
Brian A. Henning
strutmasters.com
336.597.2397x238
----------------



More information about the TriLUG mailing list