Re: [exim-dev] Exim from mailnull by local "Auto-Submitted: …

On 6/8/05, Tim Wilde <twilde@???> wrote:

>No. Is UNIX stupid for giving you the 'rm -f /etc/passwd' option?

It would be stupid if it had a "rm_etc_password" command which is what
I would compare "ignore_bounce_errors" to. I hope you notice the
distinction.


On 6/8/05, Tim Wilde <twilde@???> wrote:
> On Wed, 8 Jun 2005, Tony Marques wrote:
>
> > I hope people here will excuse me if I think Exim is stupid for having
> > such an option.
>
> It isn't ignore_bounce_errors_after that's a problem, it's a combination
> of ignore_bounce_errors_after and auto_thaw. As others have pointed out,
> this is a problem with the way these servers are being configured, not
> with Exim itself. The default options (auto_thaw of 0s and
> ignore_bounce_errors_after of 10w) are perfectly sane, and will result in
> exactly one retry, 10 weeks later, not constant retries.

Hah, there isn't a "ignore_bounce_errors" command (either my misread
or a result of a "typo/quick typing"), there is a
"ignore_bounce_errors_after" command. That makes all the difference
in the world, and I can finally see more of a reason for such an
option.

Yet, wouldn't you only just need an "auto_thaw" of 10w to do what you
described? You still wouldn't need a "ignore_bounce_errors_after".
Also wouldn't both your example and my example both result in one
retry every 10w, repeatedly? What other option determines a bounces
life?

Of course this is the first time I've encountered any of these options
and I'm not familiar with their behaviors. I'm just reading them
literally and imagining what they describe.

In the case where a server is sending bounces repeatedly every two
hours wouldn't "auto_thaw 2h" be enough to do it? On the other hand
an "ignore_bounce_errors_after 0" result it in retrying as if it
encountered a 45x error and perhaps also retry every two hours? --
strike that... I've just read the message about i_b_e_after option
deleting after one additional retry. We're left with auto_thaw alone.


> These are both very sensible options to have, the only problem is someone
> configuring them incorrectly, and possibly setting it as a default in a
> product that includes Exim. You need to complain to the server admins
> and/or whomever is distributing exim configured this way by default, not
> here.

Well, that would be entirely reasonable if 45x errors didn't exist.
As it is, they are questionable at best.


> Tim Wilde

Thanks Tim. Things are more little more sensible now.
"ignore_bounce_errors" wasn't sensible.


> Tim Jackson wrote 10:14 am
>
> So ignore_bounce_errors_after set to non-zero will cause *one*
> additional bounce retry, approximately 2 days after the first rejection,

Ah, even more sense!

"ignore_bounce_errors_after" causing at most one additional bounce is
entirely sensible -- anything that eventually discards a bounce is
fine.

> Exim to have, since a well configured system very rarely generates
> bounces anyway (at least not to external users; and internal mail is a

Yes, that is one of the reasons why I thought if this was a bug, could
explain why it would go unnoticed.

> A more problematic option is likely to be auto_thaw; this *will* cause
> repeated retries of rejected bounces.

That is likely what I've been describing.

Now why would you need an auto_thaw if there is an i_b_e_after.
Perhaps auto_thaw should not apply to frozen bounce messages, would
that be what i_b_e_after be for?


> Tim Wilde

Again, thanks Tim. Things are even more sensible now. auto_thaw all
messages is bad.