On Sun, 2 Aug 1998, Hans Morten Kind wrote:
> Half my mainlog consisted of errors like
> SMTP error [...] after MAIL FROM: [...] 421 Disk full
> exim tried delivery every minute to these hosts. Is there any reason why
> exim is not using the retry-database in such cases? The disk-full
> situations are reported at different stages during the SMTP-dialogue,
> some servers reply 4xx after MAIL FROM:, some after DATA.
This is an example of a logical problem which I have noted in the past,
but for which I haven't found a solution. At present, the receipt of a
4xx response to MAIL FROM or RCPT TO or DATA does not delay other
messages to the same host, because Exim thinks it is a problem that is
specific to this one message, and not to the host as a whole. (For
example, it could be "quota exceeded" for the receiving user, or it could
be some DNS delay in trying to verify the MAIL FROM argument.)
> I patced the transport smtp.c to do retry_add_item when a 4xx code was
> received
Of course this could be done. However, you then have the problem that
all mail to a host gets delayed if there is a temporary problem with one
specific message.
I suppose the only thing that could be done would be to keep a database
of hosts for which 421 errors had been received, and to delay retrying
if too many were received in a given interval, or some rule like that.
That is obviously a non-trivial bit of work.
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
--
*** Exim information can be found at
http://www.exim.org/ ***
