decrease log level for send_mail (was: Configuring for use with qmail)

On Wed, 31 Mar 2004, Marco Paganini wrote:

>
> > > > Marco: I think a more prominent log level for the error message would
> > > > be in order when the sendmail command fails, no?
> > >
> > > The problem is that sendmail will fork and always return a valid
> > > code...
> >
> > ... but some light-weight SMTP forwarders (nullmailer, ssmtp, etc.)
> > may not.
> >

> When I talk "sendmail" in this particular case, I'm referring to
> "/usr/sbin/sendmail", provided by many different MTAs.
>
> The problem with detecting a "failure" with sendmail is that sendmail
> delivers asynchronously. It will always return an "OK" code and deal with the
> problem later.

I think the situation is more diverse than that. Some lightweight
sendmail clones (ssmtp) do *not* fork, they deliver synchronously and
may return an interesting error message. Other may deliver
asynchronously but are able to detect some error conditions (malformed
message for instance) earlier and emit an interesting error code
immediatly.

I really think a non-zero error code from sendmail is:
- not so extraordinary (depending on your specific configuration)
- obviously very important to know/log when it happens

So simple suggestion: different log levels for successful vs failing
sendmail invocation.



> Of course, it can be configured to run synchronously, but that
>  would add undesirable delays to the process.

I don't understand this: add delays to which process ?

Moreover: why does this configuration choice matters to ASK ?


Cheers,

Marc.



-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id70&alloc_id638&op=click