Hi
> I can't comment on that from practical experience, but what do you mean by "exim would fail"? Surely the worst that would happen is that the mail
would be temporarily refused, and tried again later (at least by bona fide
senders - some spammers don't bother to retry, which can make the occasional
temporary failure quite beneficial ;-)
Do you have clamav set up?
> The way that we have it set up, if spamd fails - which it does occasionally, though not often - then the mail is accepted without being
spam-rated. So the spam load goes up temporarily, until the spamd gets
restarted (there's a cron task that checks it and resuscitates it if
necessary, but it's very rarely necessary).
Could you post that part of your configuration?
Thats just the point I am looking for - I dont know how to bypass the spam
detection if spamd is offline.
> I don't believe that will happen, at least we've not seen any evidence of it in practice. More likely, if something goes wrong near the end of the
DATA phase (which is what we are talking about), the risk is that the same
mail will be delivered twice, i.e one attempt to send to us is accepted but
the far end thinks it failed, so they send it again later.
Is it possible to shift spam and virus detection behind the data phase?
First of all I do not want to reject any mail and second, I need to be able
to select and deselect that scans depending on the receipients address. This
seems simply impossible at data phase.
