Hi there. New here 
I work for an ISP and we are trying to get an issue resolved. We run a mail server which our customers can use. This is running RedHat 4 enterprise with Exim 4.43 and exiscan-acl patch revision 28 installed.
Currently the server is setup to accept all emails sent to domains we administrate and then it send that mail off to the clients. All clients have primary mx records, we are just the backup.
This is causing a problem now with backscatter. i.e. as our server has accepted the email, we now try and send that mail to our client, who rejects it as there is maybe not a user there by that name.
To get around this I'd like our server to verify if the user exits on our clients server BEFORE we accept and hold that mail. If not we need to reject it right there and then.
Now I know exim can be configured to verify the recipient on the local server before accepting mail but what about the next mail server in line? Can this be done?
I can't test this on our live server as there are a few hundred companies running through it. I have set up a 3rd exim server as a test and will be testing everything from there. Currently the important parts of the config are below:
I've added the following bit as well, but as far as I know this only checks the local user database? Am I right here?
This last part I've added just before the warn conditions in the first statement. relay_to_domains is of course a file which contains all our customers domains.
Has anyone got any ideas that could help me out here?
Thanks!
I work for an ISP and we are trying to get an issue resolved. We run a mail server which our customers can use. This is running RedHat 4 enterprise with Exim 4.43 and exiscan-acl patch revision 28 installed.
Currently the server is setup to accept all emails sent to domains we administrate and then it send that mail off to the clients. All clients have primary mx records, we are just the backup.
This is causing a problem now with backscatter. i.e. as our server has accepted the email, we now try and send that mail to our client, who rejects it as there is maybe not a user there by that name.
To get around this I'd like our server to verify if the user exits on our clients server BEFORE we accept and hold that mail. If not we need to reject it right there and then.
Now I know exim can be configured to verify the recipient on the local server before accepting mail but what about the next mail server in line? Can this be done?
I can't test this on our live server as there are a few hundred companies running through it. I have set up a 3rd exim server as a test and will be testing everything from there. Currently the important parts of the config are below:
Code:
acl_check_rcpt:
accept
local_parts = postmaster
domains = +local_domains
warn
condition = ${if eq {$ACL_NONCUSTOMER}{yes}{yes}{no}}
delay = NONCUSTOMER_RCPT
warn
domains = !+relay_to_domains
set ACL_UNAUTHORISED = yes
delay = UNAUTHORISED_RCPT
acceptI've added the following bit as well, but as far as I know this only checks the local user database? Am I right here?
Code:
accept
domains = +relay_to_domains
endpass
verify = recipientThis last part I've added just before the warn conditions in the first statement. relay_to_domains is of course a file which contains all our customers domains.
Has anyone got any ideas that could help me out here?
Thanks!
