Re: [courier-users] Re: Automatically use secondary MX host if primary

Jay Lee wrote:

> One difficulty here is that you're offering suggestions to "improve
> Courier" rather than actually explaining what your problem is. That's
> somewhat offensive to a developer, see:
>
> http://www.catb.org/~esr/faqs/smart-...html#id3001405

I'm not claiming I found a bug. I'm suggesting a new feature.

> http://www.catb.org/~esr/faqs/smart-questions.html#goal

I believe I did exactly what is suggested but as I might be wrong I
apologize and will try to make myself clearer this time.

>
> Sam isn't going to want to change a single line of code until you
> explain what the problem actually is

The problems are two:

1. Courier gets some 4xx error. There are other MXs available but as far
as I can Courier tries the same MX several times if not all the time so
message delivery is unnecessarily delayed or delivery even fails.
Wouldn't it be better to try other MXs (same or different weights) in a
round robin or similar fashion in every delivery attempt (obviously
respecting weight precedence)?

2. Courier's connection to the other server times out so it defferes
message delivery. There are other MXs available but as far as I can
Courier tries the same MX several times if not all the time so message
delivery is unnecessarily delayed or delivery even fails. Wouldn't it be
better to try other MXs (same or different weights) in a round robin or
similar fashion in every delivery attempt (obviously respecting weight
precedence)?

As you can see there are two different problems for which I suggesting
the same solution. I will answer all the questions you proposed below
for the second problem as I have logs at hand for it.

> and why Courier should be the one to fix it.

Because the problem I see is that Courier isn't behaving as robustly as
it could in face of difficult situations real life presents to a MTA. I
repeat I'm not claiming Courier has a bug. I claiming it could have a
more robust behaviour. And to be entirely clear, I not saying Courier
isn't robust, I'm saying it could be even more robust.

> You should start by answering some of these questions: What do the
> client's MX records actually look like, are they equal weight or
> primary/secondary?

scorsese Maildir # dnsmx br.abnamro.com
30 plum03ap.abnamro.com
30 walnut001ap.abnamro.com
10 brsmtp02.br.abnamro.com
10 brsmtp04.br.abnamro.com
15 naxpf001.abnamro.com
15 naxpf002.abnamro.com
15 naxpf003.abnamro.com
15 naxpf011.abnamro.com
15 naxpf012.abnamro.com
15 naxpf013.abnamro.com

> What exactly happens when Courier tries to connect to the primary?

I get "Connection time out" errors.

> How about a session log of a server connection?

Here it is:

Jun 20 07:31:30 [courierd] newmsg,id=000F6086.42B69B17.0000146C: dns;
localhost (localhost [127.0.0.1])
Jun 20 07:31:30 [courierd]
started,id=000F6086.42B69B17.0000146C,from=<FAKE1@ empauta.com>,module=esmtp,host=br.abnamro.com,addr =<FAKE2@br.abnamro.com>
..
..
..
Jun 20 07:41:31 [courieresmtp]
id=000F6086.42B69B17.0000146C,from=<FAKE1@empauta. com>,addr=<FAKE2@br.abnamro.com>:
Connecti
on timed out
Jun 20 07:41:31 [courieresmtp]
id=000F6086.42B69B17.0000146C,from=<FAKE1@empauta. com>,addr=<FAKE2@br.abnamro.com>,status:
d
eferred

> What delivery failures are users getting?

None.

> Any custom configuration on your end?

None I remember.

If there is any further info that helps clarify my issue please let me
know so I will provide it.


Rodrigo Severo



-------------------------------------------------------
SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
from IBM. Find simple to follow Roadmaps, straightforward articles,
informative Webcasts and more! Get everything you need to get up to
speed, fast. http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click
_______________________________________________
courier-users mailing list
courier-users@lists.sourceforge.net
Unsubscribe: https://lists.sourceforge.net/lists/.../courier-users