On Wed, 5 Jul 2000, Gary Palmer wrote:
> If I read this correctly, if a mail server I secondary for (using
> exim) goes down and stays down for more than the retry interval, then
> all new mail won't be queued, but rather bounced pretty quickly. This
> isn't the desired behaviour (to my mind at least). I'd like new mail
> to wait 5 days (or however long the retry period is) just like happens
> with sendmail. Is there any (clean) way to do this?
No, though you can make it behave slightly differently by setting
delay_after_cutoff = false
in the SMTP transport. I guess this is a fundamental philosophical
question, but to my mind, as a user, if I send a message to a host that
has been dead for 5 days I *like* to be told "this host has been dead
for a long time" as quickly as possible, rather than have to wait
another 5 days. And as a postmaster, I like to get the stuff off the
queue as quickly as possible.
Obviously you take a different view, but I'm afraid Exim can't be made
to do per-message retrying like that. For those kinds of errors its
retrying is firmly host-based.
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
