Slashdot Mirror
Classed as Spam by Large-Scale Free Email Servers?
bartle asks: "I run my own personal domain that serves all of my email needs of myself and a few friends. In general this has worked out pretty well but there's a fairly significant limitation: if I send an email to a Hotmail or Yahoo account that I've never contacted before it tends to get filed as spam. This means that if I'm writing someone out of the blue I need to send an email from a free service which kind of defeats the purpose of running ones own email server. My domain has a SPF record, the IP resolves, and it doesn't appear to be on any blacklists. I can not find any documentation on what hoops I need to jump through before Hotmail and Yahoo will consider my mail legitimate. I understand that there's a general paranoia about publishing information that could assist spammers but this attitude seems to be leaving do-it-yourselfers out in the cold. Does anybody have any ideas? Are there guidelines or protocols I can follow to make my email non-spam?"
You're on a dynamic IP address, aren't you?
:\
:P
Most major mail provider bl dynamic IP's. The way around this is to smarthost against a known, static-IP mail server. In short, smarthost outbound mail. Inbound mail is fine, and you'll be all fixed up.
Not flaming you, but any good smtp faq or mailing list would have told you this, and Sendmail's FAQ answered it for me 3 years ago.
Just wondering how we get to an ask slashdot from a simple mail administration question. Google really *would* have answered this.
I know, I know, I'm new here. Next meme.
Karma: Chameleon (mostly due to the fact that you come and go).
Most likely your IP is listed at those sites as belonging to a range given out to ISP's for reistribution to their customers. You are probably rejected for that reason, because 'normal' domestic users don't have mail servers, or so these parties seem to wrongfully think. You can configure your mail server to send out the mail through our ISP's smtp server (smarthost).
Likely it's because when the other end does a reverse DNS record lookup, and your hostname and the PTR record don't match up. Usually this ends up resulting in receiving fine but problems sending.
Try setting up your ISP's SMTP server as your outgoing mail relay. In other words, when you send mail to your SMTP server, instead of looking up the remote host, doing an MX record lookup, etc., just send to your ISP's SMTP server. They should be configured to accept anything from your IP (you are their customer after all), and it only requires one extra hop for your email on the way out the door.
Instead of:
Email client -> Your SMTP -> MX record lookup -> Destination
it becomes
Email client -> Your SMTP -> ISP SMTP -> MX record lookup -> Destination
After doing this, from your point of view, nothing will have changed, and you can learn to sysadmin on a small scale to your heart's content.
A lot easier than getting an ISP to change the PTR record to your hostname.
- I don't need to go outside, my CRT tan'll do me just fine.
Rather than connecting directly to Yahoo's or Hotmail gateways, use your ISP's mail relay. That's what it's there for.
Problems:
1. Many ISPs won't relay mail unless the "From:" is in the ISP's domain. This prevents forgeries by zombies that try to relay through with random from addresses. More importantly for the ISP, they get to use viruses and spam as an excuse to force you to use their e-mail address, making it harder for you to switch ISPs.
1a. Yes, I know about "Reply-To:." Many brain-dead mail servers, list servers,
and even e-mail clients apparently don't.
2. ISPs often have limitations on attachment size. If I want to e-mail a 9MB file to a client or family member that can't deal with passworded FTP, I don't need my ISP's mail server rejecting the e-mail.
3. ISPs often disallow attachments which are executable. Again, not a hassle when dealing with computer-savvy recipients, but not all recipients are that sharp.
4. If the ISP ends up on a blacklist, your e-mail doesn't go through to mail servers that use that blacklist. I have a much better ability to control spam going through my server than to control spam going out through my ISP's mail server.
5. You're at the mercy of the ISP. It their mail server goes down or experiences other problems, your outgoing e-mail is either lost or delayed.
6. If there are e-mail delivery problems, your server won't have useful logs (since the actual delivery was attempted by your ISP's mail server. You won't be able to tell how many times a message was retried, whether something timed out in the protocol, etc.
7. I'd rather not have my ISP retaining copies of my e-mail, auto-scanning it with who-knows-what software, passing it on to the FBI for warrantless PATRIOT Act fishing expeditions, etc. While I know that they could do that with a port 25 snoop, chances are that they wouldn't routinely do that.
Yes, I know that there are inconvenient workarounds for some of the problems listed above, but, all in all, it's far preferable to use your own server.
chrisbartle.com resolves to 216.17.137.189
but
216.17.137.189 resolves to bartle189.dsl.frii.net
So it doesn't resolve correctly. You might think you have a "static" IP address, but it appears the same as any other, dynamic, home DSL address.
Unless you can get frii.net to change their DNS entry for you, you'll continue to have your mail rejected.
Guys, he's got a TXT record for SPF. If he starts relaying through his ISP's mail servers, he's either got to give up SPF, or add the appropriate include: modifier. Neither is particularly thrilling.
:)
Chris, I'm assuming you want to send out only from mail.~, since you have no PTR in the SPF, right? Still might be useful to add "A" to the SPF, since mail.~ is a CNAME. Or maybe not.
And if it helps any, you seem to be clear off all the RBLs I was able to check.