On 8/26/2011 3:53 PM, lance raymond wrote: > Sorry for the previous, I don't wish to make things complicated, so > focusing on my in-house server. I have moved my application server > to use this server and I see mail is going out, to the main google, > yahoo, etc. but get one deferred on an .edu server. The error is > "host mail-mx1-prod-v.cc.nd.edu > <http://mail-mx1-prod-v.cc.nd.edu>[129.74.250.243] said: 451 4.1.8 > Possibly forged hostname for myIP (in reply to RCPT TO command)
Does the mail eventually go through? If yes, then end. The error mentions "hostname for IP". This suggests a dns mismatch somewhere. Share your actual domain name and server IP to get suggestions. Maybe your HELO name (main.cf: smtp_helo_name, default $myhostname) doesn't have an A record pointing back to your server. Or maybe your IP has a "generic" hostname. Or no hostname at all. > > Now, the problem is simply that 'ourdomain.com > <http://ourdomain.com>' is hosted at google and we cant relay though > him. Should be able to if you set up client auth http://www.postfix.org/SASL_README.html#client_sasl > Our application sends mail from noem...@mydomain.com > <mailto:noem...@mydomain.com>. > Now I guess there doing a comparison > to the mx on mydomain.com <http://mydomain.com> which points to > google then where the mail came from. So the easy question is am I > correct? Probably incorrect. You should contact postmaster at nd.edu and find out why they defer your mail. Or post the unaltered "postconf -n" and log entries so we can examine your dns records. > Next, can I setup my inhouse postfix to send mail from > mydomain.com <http://mydomain.com> or will I have issues later (this > could be just the 1st). Yes, possible to do this without issues. Proper DNS entries (including spf records including google + your local host) are a big step towards getting this resolved. > > If not, I guess my other option is looking for all the places mail > get's sent and change the name to like noem...@sub.mydomain.com > <mailto:noem...@sub.mydomain.com> and then use that. Not likely to change this particular problem. -- Noel Jones
