Re: [EXIM] "full virtualization"?

トップ ページ
このメッセージを削除
このメッセージに返信
著者: Lutz Pressler
日付:  
To: exim-users
CC: Philip Hazel
題目: Re: [EXIM] "full virtualization"?
> I don't think I quite understand. Exim already stores
> sender_host_address in the spool header file for messages that have come
> from a remote host. Is this what you are wanting, or is there something
> else I have missed?

Now I don't think I understand.. I don't find any sender_host_address
entries in *-H files here, only host_adress - and that is the remote
address. What I would need is the connect local address.

>Are you just wanting to be able to say
>
> interface = $sender_host_address
>
> in the smtp transport?

Yes, if sender_host_address is the connected local address.

>> In principle this should not be difficult to do, I think. The more
>> complicated part will be the code that reuses SMTP connections to the
>> same destination which has to differentiate by source address now, too.


> That would be hard, because you can't discover the sender_host_address
> without reading the message's header file, and Exim does not do that at
> the right time. I don't think this could be done at all easily without
> major nasty hacks to the code.... thinks ... well, what might work is to
> key the records in the waiting database by interface as well as by
> remote IP address. If you did that, it might be reasonably
> straightforward, now I think about it.


>> What does it do at the moment if multiple SMTP transports with different
>> "interface" options are used? Have not looked yet.


> Each SMTP transport has its own "waiting" database.



-- 
  _              |  Lutz Pressler          |  Tel: ++49-551-3700002
 |_     |\ |     |  Service Network GmbH   |  FAX: ++49-551-3700009
 ._|ER  | \|ET   |  Hannah-Vogt-Strasse 1  |   mailto:lp@SerNet.DE
Service Network  |  D-37085 Goettingen     |  http://www.SerNet.DE/




--
*** Exim information can be found at http://www.exim.org/ ***