On Aug 20 23:01:44, Gilles Chehade wrote:
> Dear misc@ and tech@,
> 
> We are getting closer to a stable version of OpenSMTPD and now would be
> the perfect time for you to start stress-testing and trying to crash it
> in various ways to make sure it is rock-solid.
> 
> Eric's recent commits brought to the daemon a brand new scheduler and mta
> logic which we started working on during the last hackaton, and that is
> much much much more efficient than what we had before. It just needs to
> be heavily tested ;-)
> 
> What's in -current now is very very close to what we want for the first
> release, minus minor bugfixes and one additional feature we'll work on.
> There will be no major feature written and committed for our very first
> release, so just make sure that you test it enough so that we can reach
> the quality you'd expect from us.


'smtpctl log brief/verbose' doesn't seem to make any difference,
either with incomming or outgoing mails, or various smtpctl actions
(such as smtpctl remove, pause/resume, ...). All of that gets logged
exactly the same with log brief and log verbose. Is this expected,
or just unimplemented yet?

Is 'log verbose' equivalent to running smtpd with '-v'?
Should this be explicitly stated in the documentation?


'smtpctl pause mda' indeed pauses local deliveries.
Sending one message to a local user with mda paused
results in the following queue:

# mailq
MDA|30850f19476f6e26|-|jsta6...@artax.karlin.mff.cuni.cz|h...@stare.cz|0|345600|0
MDA|30850f19476f6e26|-|jsta6...@artax.karlin.mff.cuni.cz|h...@stare.cz|0|345600|0
MDA|30850f19476f6e26|-|jsta6...@artax.karlin.mff.cuni.cz|h...@stare.cz|0|345600|0
MDA|30850f19476f6e26|-|jsta6...@artax.karlin.mff.cuni.cz|h...@stare.cz|0|345600|0
MDA|30850f19476f6e26|-|jsta6...@artax.karlin.mff.cuni.cz|h...@stare.cz|0|345600|0
MDA|30850f19476f6e26|-|jsta6...@artax.karlin.mff.cuni.cz|h...@stare.cz|0|345600|0
MDA|30850f19476f6e26|-|jsta6...@artax.karlin.mff.cuni.cz|h...@stare.cz|0|345600|0
MDA|30850f19476f6e26|-|jsta6...@artax.karlin.mff.cuni.cz|h...@stare.cz|0|345600|0

The one message is reported eight times in the queue, identically.
Is this expected?

(BTW, 'mailq' seems to be exactly the same as 'smtpctl show queue'.
If it is, should it be mentioned explicitly in the documentation?)

After 'smtpctl resume mda' the message gets delivered alright,
and the queue becomes empty.


Similarly for 'smtpctl pause mta', one outgoing email is reported as

# mailq
MTA|c3947cc0e3d01e0e|ENQUEUED|h...@stare.cz|n...@exi.ste.nt|0|345600|0
MTA|c3947cc0e3d01e0e|ENQUEUED|h...@stare.cz|n...@exi.ste.nt|0|345600|0
MTA|c3947cc0e3d01e0e|ENQUEUED|h...@stare.cz|n...@exi.ste.nt|0|345600|0
MTA|c3947cc0e3d01e0e|ENQUEUED|h...@stare.cz|n...@exi.ste.nt|0|345600|0
MTA|c3947cc0e3d01e0e|ENQUEUED|h...@stare.cz|n...@exi.ste.nt|0|345600|0
MTA|c3947cc0e3d01e0e|ENQUEUED|h...@stare.cz|n...@exi.ste.nt|0|345600|0

Here it is reported six times, identically.
On the next try, it is nine times:

# mailq
MTA|90377f4e50f2c732|ENQUEUED|h...@stare.cz|n...@exi.ste.nt|0|345600|0
MTA|90377f4e50f2c732|ENQUEUED|h...@stare.cz|n...@exi.ste.nt|0|345600|0
MTA|90377f4e50f2c732|ENQUEUED|h...@stare.cz|n...@exi.ste.nt|0|345600|0
MTA|90377f4e50f2c732|ENQUEUED|h...@stare.cz|n...@exi.ste.nt|0|345600|0
MTA|90377f4e50f2c732|ENQUEUED|h...@stare.cz|n...@exi.ste.nt|0|345600|0
MTA|90377f4e50f2c732|ENQUEUED|h...@stare.cz|n...@exi.ste.nt|0|345600|0
MTA|90377f4e50f2c732|ENQUEUED|h...@stare.cz|n...@exi.ste.nt|0|345600|0
MTA|90377f4e50f2c732|ENQUEUED|h...@stare.cz|n...@exi.ste.nt|0|345600|0
MTA|90377f4e50f2c732|ENQUEUED|h...@stare.cz|n...@exi.ste.nt|0|345600|0

Why is that?

After 'smtpctl resume mta' it tries to deliver the message
- which fails of course, the address is bogus. The log says


Aug 22 06:41:48 biblio smtpd[9558]: mta: route:exi.ste.nt[] error: 600 Domain 
does not exist
Aug 22 06:41:48 biblio smtpd[9558]: mta: route:exi.ste.nt[] error: 150 Can not 
connect to MX
Aug 22 06:41:48 biblio smtpd[9558]: mta: route:exi.ste.nt[] error: 600 Domain 
does not exist
Aug 22 06:41:48 biblio smtpd[9558]: mta: route:exi.ste.nt[] error: 150 Can not 
connect to MX
Aug 22 06:41:48 biblio smtpd[9558]: c3947cc0e3d01e0e: to=<n...@exi.ste.nt>, 
delay=1m17s, relay=exi.ste.nt, stat=LocalError (Can not connect to MX)


Why is each of the messages logged twice?
Is it really a "local error" that a domain
does not exist, or doesn't have an MX?

Besides these nitpicks, the new smtpd works well for me
- thank you guys!

        Jan

Reply via email to