https://bugs.exim.org/show_bug.cgi?id=3118
--- Comment #4 from Maxim Galaganov <max@???> ---
(In reply to Jeremy Harris from comment #3)
> I was thinking that if a forced-fail was a feasible result of an expansion
> test,
> then that test can be used as a condition on routers that would avoid the
> need for
> fallback_hosts to be used. Certainly a match_ip can be used in a test on a
> router.
The use of fallback hosts in my case is not related to the use of socks proxy
and is unavoidable. Connection to the recipient MX needs to be made via proxy.
But if that MX is unavailable, then the message must be relayed to a fallback
host without using proxy, because the fallback host is unroutable via proxy (in
a 10.0.0.0/8 network).
And when in router, you cannot predict whether the destination MX is
unavailable and choose the separate fallback transport upfront, instead of
relying on fallback_hosts.
Moreso, retry_add_item() is short-circuited to return without doing anything in
our installation, apart from fallback hosts where it is fully functional. So
all other servers are oblivious to retry configuration, and fallbacks need to
be used after first attempt fails.
--
You are receiving this mail because:
You are on the CC list for the bug.
--
## subscription configuration (requires account):
##
https://lists.exim.org/mailman3/postorius/lists/exim-dev.lists.exim.org/
## unsubscribe (doesn't require an account):
## exim-dev-unsubscribe@???
## Exim details at
http://www.exim.org/
## Please use the Wiki with this list -
http://wiki.exim.org/
