Re: [Savannah-hackers-public] Problem with email delivery on fencepost?

[Bob Proulx]

Eli Zaretskii wrote:
> > Bob Proulx wrote:
> > Hmm...  debbugs is completely independent of eggs.  I don't see
> > anything wrong on debbugs.  It is receiving email and sending email.
> 
> If delivery of email on eggs stopped, then my email, sent through
> fencepost, would not get delivered to debbugs, right?

Hmm...  Initially I did not think so.  I almost posted saying that!
Because normally one would see this MX (mail exchange relay)
configuration and think it would go directly there.  (And I would have
been wrong!)

  $ host -t mx debbugs.gnu.org
  debbugs.gnu.org mail is handled by 10 debbugs.gnu.org.

But I decided to try a test of mail from fencepost to see how it went..

  Received: from fencepost.gnu.org ([2001:470:142:3::e]:34133)
        by eggs.gnu.org with esmtp (Exim 4.71)
        (envelope-from <address@hidden>)
        id 1igtjb-0003km-IX
        for address@hidden; Mon, 16 Dec 2019 12:00:39 -0500

That indicates that fencepost must be configured in the Exim
configuration to also relay outgoing mail through eggs.  I had not
known that.  So yes if eggs stalled in handling mail due to the disk
being full there then outgoing mail from fencepost would have also
been blocked behind it.

As long as I am here...  Mail from debbugs back to anyone @gnu.org
would definitely go through eggs too.  Same as for anyone else out in
the world.

  $ host -t mx gnu.org
  gnu.org mail is handled by 10 eggs.gnu.org.

> > Please send a message to address@hidden and ask about debbugs
> > being slow to respond to email.  One of Glenn, Michael, or Noam (or
> > potentially myself, but I don't know how the debbugs part operates)
> > will look at the problem there.  You need one of them to debug how
> > debbugs is working or not.
> 
> OK, will do next time something like this happens.

Now that we know in the above that outgoing mail from fencepost is
configured to relay through eggs and we know that eggs blocked for a
while due a spike in size of logfiles filling up the disk.  And we now
know that debbugs was also delayed in getting outgoing mail from
fencepost.  Therefore debbugs couldn't get the mail during the pause.
With all of the information in hand it all makes sense as to what
happened.

I like it when everything has an explanation that we can figure out. :-)

Bob