Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Set a sender email address for bounced messages

3 views
Skip to first unread message

Ajit

unread,
Sep 11, 2008, 2:31:58 PM9/11/08
to ajit...@hotmail.com
Hi! I am aware that for bounced messages sendmail sets a null sender.
We have a requirement that for bounced messages there should be a
valid sender email address. Is there a straightforward way to set a
sender for boucned messages? I have tried explaining that this is not
desirable but have been asked to investigate.

I am hoping that someone can provide me with a method to this and/or
also with suitable arguments as to why this may not be good.

Thanks

Ajit

Grant Taylor

unread,
Sep 11, 2008, 3:25:47 PM9/11/08
to
On 09/11/08 13:31, Ajit wrote:
> Hi! I am aware that for bounced messages sendmail sets a null sender.
> We have a requirement that for bounced messages there should be a
> valid sender email address. Is there a straightforward way to set a
> sender for boucned messages? I have tried explaining that this is not
> desirable but have been asked to investigate.

I'm not aware of a way to change the source address of bounced messages,
short of modifying source code.

Doing what you are asking to do is a */VERY/* */BAD/* idea. The null
reverse path sending address is used to detect and loops of bounced
messages. In short, receiving email servers will not bounce a message
if it is from the null reverse path, rather they will discard the
message (if they did not actually reject it).

Imagine if someone sends you a message that your server accepts and then
decides to bounce back to the sending server. If you send the bounce
with something other than the null reverse path address it is very
possible (and almost likely) that the message you bounce will be
re-bounced back to you. If along with sending bounces with something
other than the null reverse path address you don't have correctly
functioning bounce detection you will end up bouncing the bounce of your
bounce. Now imagine if you ended up bouncing a message back to another
server that decided not to use the null reverse path address in their
bounces? In this situation you would be completely down to relying on
the number of Received: headers in the message to detect loops. This
could (and would) quickly spiral out of control.

In short, DO NOT DO IT!

> I am hoping that someone can provide me with a method to this and/or
> also with suitable arguments as to why this may not be good.

If you can't persuade the PHBs that this is a bad idea, get the ""order
in writing signed by (at least) your supervisor to cover your back side.
(I'd also seriously consider changing jobs.)

Grant. . . .

Kari Hurtta

unread,
Sep 12, 2008, 4:33:41 AM9/12/08
to
Grant Taylor <gta...@riverviewtech.net> writes in comp.mail.sendmail:

> On 09/11/08 13:31, Ajit wrote:
> > Hi! I am aware that for bounced messages sendmail sets a null
> > sender. We have a requirement that for bounced messages there should
> > be a valid sender email address. Is there a straightforward way to
> > set a sender for boucned messages? I have tried explaining that this
> > is not desirable but have been asked to investigate.
>
> I'm not aware of a way to change the source address of bounced
> messages, short of modifying source code.

There is mailer flag for it. "SENDMAIL INSTALLATION AND OPERATION GUIDE"
is your friend:

| g Normally, sendmail sends internally generated
| email (e.g., error messages) using the null return
| address as required by RFC 1123. However, some
| mailers don’t accept a null return address. If
| necessary, you can set the g flag to prevent send‐
| mail from obeying the standards; error messages
| will be sent as from the MAILER‐DAEMON (actually,
| the value of the $n macro).


> Doing what you are asking to do is a */VERY/* */BAD/* idea. The null
> reverse path sending address is used to detect and loops of bounced
> messages. In short, receiving email servers will not bounce a message
> if it is from the null reverse path, rather they will discard the
> message (if they did not actually reject it).
>
> Imagine if someone sends you a message that your server accepts and
> then decides to bounce back to the sending server. If you send the
> bounce with something other than the null reverse path address it is
> very possible (and almost likely) that the message you bounce will be
> re-bounced back to you. If along with sending bounces with something
> other than the null reverse path address you don't have correctly
> functioning bounce detection you will end up bouncing the bounce of
> your bounce. Now imagine if you ended up bouncing a message back to
> another server that decided not to use the null reverse path address
> in their bounces? In this situation you would be completely down to
> relying on the number of Received: headers in the message to detect
> loops. This could (and would) quickly spiral out of control.

Number of Received: -lines does NOT dectect this loop. Notice that bounce
is new message.

There is nothing what will prevent bounce loops, if envelope sender
address is sent MAILER-DAEMON@{hostname}

> In short, DO NOT DO IT!
>
> > I am hoping that someone can provide me with a method to this and/or
> > also with suitable arguments as to why this may not be good.
>
> If you can't persuade the PHBs that this is a bad idea, get the
> ""order in writing signed by (at least) your supervisor to cover your
> back side. (I'd also seriously consider changing jobs.)
>
>
>
> Grant. . . .

/ Kari Hurtta

Ajit

unread,
Sep 16, 2008, 7:27:47 AM9/16/08
to
On Sep 12, 4:33 am, Kari Hurtta <hurtta
+comp.mail.sendm...@siilo.fmi.fi> wrote:
> Grant Taylor <gtay...@riverviewtech.net> writes in comp.mail.sendmail:

Hi all! Thanks for the responses. I have already explained the bounce
loop to the Powers that be. I will try out the "g" flag. I do intend
to cover my back before I present them a solution.

Once again, thanks

Ajit

Ajit

unread,
Sep 17, 2008, 1:13:47 PM9/17/08
to
On Sep 16, 7:27 am, Ajit <ajit.da...@gmail.com> wrote:
> On Sep 12, 4:33 am, Kari Hurtta <hurtta
>
>
>
> +comp.mail.sendm...@siilo.fmi.fi> wrote:
> > Grant Taylor <gtay...@riverviewtech.net> writes in comp.mail.sendmail:
>
> > > On 09/11/08 13:31, Ajit wrote:
> > > > Hi! I am aware that for bounced messages sendmail sets a null
> > > > sender. We have a requirement that for bounced messages there should
> > > > be a valid sender email address. Is there a straightforward way to
> > > > set a sender for boucned messages? I have tried explaining that this
> > > > is not desirable but have been asked to investigate.
>
> > > I'm not aware of a way to change the source address of bounced
> > > messages, short of modifying source code.
>
> > There is mailer flag for it. "SENDMAIL INSTALLATION AND OPERATION GUIDE"
> > is your friend:
>
> > | g Normally, sendmail sends internally generated
> > | email (e.g., error messages) using the null return
> > | address as required by RFC 1123. However, some
> > | mailers don't accept a null return address. If
> > | necessary, you can set the g flag to prevent send‐
> > | mail from obeying the standards; error messages
> > | will be sent as from the MAILER‐DAEMON (actually,
> > | the value of the $n macro).
>
> > > Doing what you are asking to do is a */VERY/* */BAD/* idea. The null
> > > reverse path sending address is used to detect and loops of bounced
> > > messages. In short, receiving email servers will notbouncea message

> > > if it is from the null reverse path, rather they will discard the
> > > message (if they did not actually reject it).
>
> > > Imagine if someone sends you a message that your server accepts and
> > > then decides tobounceback to the sending server. If you send the
> > >bouncewith something other than the null reverse path address it is
> > > very possible (and almost likely) that the message youbouncewill be

> > > re-bounced back to you. If along with sending bounces with something
> > > other than the null reverse path address you don't have correctly
> > > functioningbouncedetection you will end up bouncing thebounceof
> > > yourbounce. Now imagine if you ended up bouncing a message back to

> > > another server that decided not to use the null reverse path address
> > > in their bounces? In this situation you would be completely down to
> > > relying on the number of Received: headers in the message to detect
> > > loops. This could (and would) quickly spiral out of control.
>
> > Number of Received: -lines does NOT dectect this loop. Notice thatbounce
> > is new message.
>
> > There is nothing what will preventbounceloops, if envelope sender

> > address is sent MAILER-DAEMON@{hostname}
>
> > > In short, DO NOT DO IT!
>
> > > > I am hoping that someone can provide me with a method to this and/or
> > > > also with suitable arguments as to why this may not be good.
>
> > > If you can't persuade the PHBs that this is a bad idea, get the
> > > ""order in writing signed by (at least) your supervisor to cover your
> > > back side. (I'd also seriously consider changing jobs.)
>
> > > Grant. . . .
>
> > / Kari Hurtta
>
> Hi all! Thanks for the responses. I have already explained thebounce
> loop to the Powers that be. I will try out the "g" flag. I do intend
> to cover my back before I present them a solution.
>
> Once again, thanks
>
> Ajit

Hi! I have tried setting the g flag for all the mailers and also
created a new error mailer but the bounce message still comes with a
null sender. Obviously I am not defining the flag properly. Would you
be able to tell me how to set the g flag.

Thanks

Ajit

0 new messages