Here's an issue I noticed with fallback_hosts and it might be a bug.
The command works correctly if it fails to deliver to the main host it
will send the message to the fallback host, unless the problem sending
to the mail host is because of dns resolution. Then it just retries
rather that sending to the fallback host.
For example:
process_and_forward:
driver = manualroute
domains = +preprocess_domains
condition = ${if match{$h_X-Spam-Class:}{SPAM-HIGH}{no}{yes}}
transport = process_and_forward_smtp
route_list = *
${lookup{$domain}dbm{/etc/exim/control/run/preprocess.db}{$value}{}}
fallback_hosts = retry.ctyme.com
In this case the route_list is looked up from a database. Let's say it
looks up mail.domain.com. If that host is down then the message is sent
to retry.ctyme.com, but - if the host name mail.domain.com doesn't
resolve then it just retries rather than sending the message to
retry.ctyme.com.
I think this is a bug. If the host doesn't resolve then it should send
it to the retry server.
For what it's worth the reason this comes up is that I do front end spam
filtering for thousands of domains. It's a somewht common occurrance
that someone will change their DNS and the host name that I'm forwarding
to goes away. So what I want to have happen is that if the host name
goes bad I want to transfer the message to my retry server and let it
buffer these messages until the customer figures out how to fix the DNS.
BTW - a good way to make Exim faster is to have a main server that tries
one tome to send a message and if it fails it laterals the message to
another server. That keeps the queue size small on the main server and
increases speed.
--
## List details at http://lists.exim.org/mailman/listinfo/exim-users
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/
