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

On Wed, 8 Jun 2005 09:55:31 -0700
Tony Marques <tymes10@???> wrote:

> > i suspect the behaviour you are observing is coming from hosts that
> > have ignore_bounce_errors set to anything not zero, and have
>
> Well, an "ignore_bounce_errors" set true should be enough then...
> there isn't a need for any cron jobs or anything to thaw as it would
> never get frozen and retrying is normal.

I should point out that I think Tamasz was referring to
ignore_bounce_errors_after, which doesn't do quite what it sounds like,
and certainly doesn't cause incessant retries. From the manual:

"After a permanent delivery failure, bounce messages are frozen, because
there is no sender to whom they can be returned. When a frozen bounce
message has been on the queue for more than the given time, it is
unfrozen at the next queue run, and a further delivery is attempted. If
delivery fails again, the bounce message is discarded. This makes it
possible to keep failed bounce messages around for a shorter time than
the normal maximum retry time for frozen messages."

FWIW, this option is set to 2 days by default.

So ignore_bounce_errors_after set to non-zero will cause *one*
additional bounce retry, approximately 2 days after the first rejection,
and then discard the bounce. This is not an unreasonable option for
Exim to have, since a well configured system very rarely generates
bounces anyway (at least not to external users; and internal mail is a
matter of policy), and it is possible that someone misses a bounce
because their mailbox is full or something. (The systems you are
fighting with are not well- configured if they are generating bounces
because they don't like the attachment name or whatever; again, this is
not Exim default behaviour)

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

FWIW, you are not the only one suffering from brain-dead hosts like
this. Just this morning I was browsing some logs and found some
incessant retries from an Exim server like you describe. They too
carried the X-AntiAbuse headers I was on about earlier, and now you
mention it, I do think I see this with a fair degree of frequency too.

Whilst you're right in saying that the presence of auto_thaw does give
users a long rope to hang themselves with, I wouldn't interpret the
presence of lots of machines doing this as "lots of users get it
wrong"; as I said in my post earlier today, I think it's far more
likely to be a poor choice that is imposed on users of a particular
packaging or configuration system.


Tim