Re: [Fwd: [ANNOUNCE] PostgreSQL RPM Yum repository is available] - Mailing list pgsql-www

On Tue, Jan 08, 2008 at 09:58:20PM -0400, Marc G. Fournier wrote:
> 
> Since nobody actually  understands *why* I need multiple headers, let me 
> explain:
> 
> Nothing in the headers below tells me *where* the problem is ... for instance, 
> according to the below, it was sent and received in the moderator queue at 
> around 16:30 on Monday ... it appears that at ~10:42am, it was then released 
> out of the queue and proceeded to be sent to Magnus, and delivered within a 
> couple of minutes ... so delivery time looks to be okay ...

If that's so, then that's the problem. Because the first copy of this email
was received *way* earlier than that. As in right after it was sent. Maybe
you can get a bunch of the original headers fromt he MBOX files in the
archives? (they have only one copy of it, but I'll bet it shows a
completely different delivery time)


> A second set of headers would at least let me see at what point the message 
> id's divert ... for instance, if "ESMTP id 281FC12A672F" stay'd constant, but 
> "ESMTP id 476F9DCC195" changed, then mx2.hub.org resent it to magnus multiple 
> times, and I can look at that ...

Yeah, I wish I did keep them. But it's not like I expected more copies to
arrive later :-( And they arrived far enough apart not to still be
available in the trash.


> According to mx2.hub.org itself, that ESMTP id sent to magnus *once*:

That corresponds to my logs. Actually, grepping for the messageid it only
turns up once.

Who generates the messageid? Devrim? majordomo? maia? Because I only see that
exact messageid once (not ESMTP id here, actual message id
"1199744552.30375.57.camel@localhost.localdomain") only shows up once in my
logs. So the different copies of the mail was delivered with different
messageids.



> Now, what we don't know ... those other 4 to 5 copies, do they all have the 
> same ESMTP id as this one, or did somehow 4 to 5 different copies get into the 
> queue, with seperate ESMTP ids ...

They don't have teh same ESMTP on my end at least, and not even the same
"actual" id. 


> Even doing a grep on magnus@hagander.net doesn't show any errors ... connection 
> to his host don't appear to have much in the way of delay ... that 28 second 
> one seems to be about the biggest one ...

Yeah, the delay is usually early in the chain, but usualliy not that bad.
But the total delay in the mod queue is the problem here I think - because
I certainly received the first copy a *lot* earlier than this one left the
moderator queue (if that's what's done at the step of that long delay)


//Magnus