Koozali.org: home of the SME Server
Legacy Forums => Experienced User Forum => Topic started by: ptb on April 05, 2002, 09:46:01 PM
-
I've searched back through the archives and haven't found anything that explains the remote host configuration, so I apologize if I'm asking an obvious question.
I'm trying to use the ETRN command to fetch email from a remote host. I've configured the SME server to do so, and it appears to be doing as expected.
The remote host is running sendmail 8.11.6, so ETRN is supported. I can see the ETRN command being accepted by the remote host (by looking at /var/log/maillog)
However, no mail is being retrieved, apparently because email is being delivered to the user at the remote host (where I can retrieve it via IMAP or POP3).
My question is: how do I configure the remote host to hold the email in the queue so that it can be retrieved? Here's the situation:
Remote host:
xyzzy.com
www.xyzzy.com
mail.xyzzy.com
SME server:
office.xyzzy.com
MX:
mail.xyzzy.com 10
Currently, the MX record points at mail.xyzzy.com. Once I confirm that ETRN is working, I want to change it so that the highest priority is office.xyzzy.com and second priority is www.xyzzy.com, i.e.
Remote host:
xyxxy.com
www.xyzzy.com
mail2.xyzzy.com
SME server:
office.xyzzy.com
mail1.xyzzy.com
MX:
mail1.xyzzy.com 10
mail2.xyzzy.com 20
I _think_ the answer is to change the virtusertable on the remote host so that mail sent to user@xyzzy.com is resent to user@office.xyzzy.com, i.e.
@xyzzy.com %1@office.xyzzy.com
Will this have the desired effect? Is there a different/better way to do it?
I have one more wish/requirement that I'm not sure is possible to meet: can I make the mail accessible on the remote host, in the event that that office.xyzzy.com becomes unavailable for an extended period of time? Since it appears that the mail must remain in sendmail's queue, it doesn't appear to be immediately available.
But, if a change to the virtusertable (removing the above entry) would cause sendmail to deliver the email locally (even though it is already in the queue), that would be an acceptable fall-back strategy -- presuming that the users in question already have accounts on the remote host.
Has anyone implemented something like this?