On Thu, Aug 05, 2004 at 10:34:23PM -0600, Shane Boulter wrote: > Greetings, > > I'm in the process of wrapping up a new qmail-ldap server for our > company. I'm using qmail-ldap and the qmail-scanner program. I really > liked the qmail-scanner program however I didn't like how long it took > for a client to send it an email. So following the qmail-scanner people > recommendations I install two instances of qmail-ldap. One outside > facing and one inside. Basically all emails come though the outside > facing qmail-ldap instance and are forwarded using smtproutes to the > inside qmail-ldap instance. This allows for no delays to clients in > sending emails. > > I was extremely happy with this setup until I entered our new host in > the dns as a mx record. I then started to get all these emails to fake > email addresses. eg [EMAIL PROTECTED] > > I realize that by setting the RCPTCHECK variable it should just bounce > these but after reading the notes on how RCPTCHECK works I finally > understand why it doesn't work for my situation. As I don't have any > domains in my outside facing qmail-ldap control/locals it never checks > for these users to really exist. I can't have my domains in my outside > control/locals as all I want this instance to do is forward the email > using smtproutes to my inside qmail-ldap instance. > > Is there a reason why the RCPTCHECK doesn't go against the > control/rcpthosts? And the big question is could it be programmed to be > more like the SENDERCHECK option where I can turn it to check against > the control/rcpthosts file? >
You need to set this up as a cluster. The antivirus gateway is accepting mails and does also a local delivery but because the mailHost attribute is set to your internal mailserver it will forward the mails. With this you can use RCPTCHECK. There is also a simpler solution with a mini-qmail setup on the av gateway which does not need any changes in the ldap db. The actual setup is left open as an exercise for the reader :) -- :wq Claudio
