On Wed, Mar 05, 2014 at 02:11:23PM +0100, Michael Neumann wrote:
> Hi,
> 

Hi,


> I am having an OpenSMTPD configuration which relays local mail to
> itself, failing after 100 (?) recursions.
> 

yup, the safe-guard loop detection kicked in at this point.


> The problem is that a periodic cron-job sends mail to root@`hostname`,
> but the OpenSMTPD instance on `hostname` does not have a rule to accept
> mail for `hostname`. `hostname` in this case is for example
> mail2.mydomain.de. As mail2.mydomain.de does not have a MX record (only
> mydomain.de has a MX record, pointing to mail.mydomain.de), it tries to
> relay the mail directly to mail2.mydomain.de, that is, it tries to relay
> the mail to itself. Luckily there is a limit and after 100 recursions it
> stops.
> 

mh, is your instance declared as a backup MX for the other domain ?
if it's not declared as a backup MX, then the resulting loop is pretty
much ... what you asked for :-)


> It's easy to fix by adding a rule for mail2.mydomain.de, but I am just
> curious wheather it's good practice or useful to relay to ourself, or
> whether we want to forbid that?
> 

yes, in my opinion the correct fix is at the config level.

if the smtp / dns configuration allows for a loop, trying to fix it at
any other level than the configuration is going to break setups. this
is the case with our safe-guard which will prevent opensmtpd from running
in an environment where it takes more than 100 hops to reach destination.

if you have ideas, i'll be open to them

-- 
Gilles Chehade

https://www.poolp.org                                          @poolpOrg

-- 
You received this mail because you are subscribed to [email protected]
To unsubscribe, send a mail to: [email protected]

Reply via email to