On Mon, 2008-01-21 at 09:16 +0200, Sim Zacks wrote:
> Are there any server-side security possibilities involved in dbmail?

  Yes.

>  For 
> example, is it at all possible that a received email attachment will 
> actually execute on the server?

  That case is not likely, but there are other ways to get things to
execute, eg. if you found and exploited a buffer overflow or point of
sql injection.

>  Or could a mail with specific headers 
> cause the dbmail to execute shell commands or run an application?

  It's not designed to execute commands from the headers, but that
doesn't mean there will never be any errors found that would allow that
to happen, or even that libraries dbmail uses (eg. gmime) wouldn't have
an error and allow that.

> I'm trying to decide if it makes sense to put the dbmail on my 
> production server or if I should lock it down on its own server.

  It would be more secure to keep it isolated, though you can do that on
one physical box as well (eg. in a virtual machine, or maybe even
running chroot).  If you have users on the shared server, or any other
access (eg. a webserver running), or use a shared database, then the
installation procedures (eg. file permissions and whatnot) are of much
greater consequence.  But security is complex, there's no
one-size-fits-all answer.

-- 
Jesse Norell
Kentec Communications, Inc.
[EMAIL PROTECTED]
_______________________________________________
DBmail mailing list
[email protected]
https://mailman.fastxs.nl/mailman/listinfo/dbmail

Reply via email to