Slashdot Mirror
Protecting Your Small Domain from Spam Hijacking?
"My domain hosting service, CubeSoft, has been a good host for my domain for the past three years, and they have been very helpful in re-enabling most of my account, but at the moment they don't want to re-enable my e-mail because of the flood of returned spam coming in (30,000 messages per day). Since the return addresses are all invalid (e.g. 'nonexistent_address@gelhaus.net'), I would think it would be simple to filter out all messages that aren't specific ones I've set up (e.g. 'valid_address@gelhaus.net'). I can't believe my domain is the first to have experienced this problem. It would be a tragedy to have to just shut down my domain because of this. CubeSoft says there isn't any way to prevent it because there is nothing that stops a spammer from using a fake return e-mail address. What have others with small domains done to protect themselves?"
All the above is conjecture, of course. But it may be something for your ISP to think about. It may be possible to re-enable the MX for your domain in a short while without having to do anything.
BTW, this is generally known as a Joe Job.
I can't say that I don't give a fuck. I've just run out of fuck to give.
We have had the same issue, unfortunately. I asked on the debian-isp mailing list about it and the only real suggestion was to report the spammer in question to their ISP, which I believe to be in Russia.
The long and short of it is that we couldn't do much about it, other than try to minimize the resource waste. In our exim configuration we turned on "receiver_verify" in our exim configuration, which means before the incoming message enters the delivery phase, it's verified that there is a valid receiver. (Before doing this, the incoming message would run through spamassassin and then generate a bounce, using CPU time, memory, etc.) I know it's not much; I hope someone comes up with more suggestions.
See http://spf.pobox.com You can publish your DNS now, indicating which legitimate IPs are in use for mail from your domain.
by a secure protocol, I doubt very much anything can be done to protect against what is essentially a DDoS attack (which is, of course, a mere side effect of spam). But nobody seems interested in a modern-day email alternative. Whenever something bad happens, it's always the bad guys' fault, right? Remember, we don't need security, just a world with no bad people.
You need to change your domain name. Obligatory "Office Space" quote:
Samir: You know, there's nothing wrong with that name.
Michael Bolton: There WAS nothing wrong with it. Until I was about 12 years old, and that no-talent-ass-clown because famous and started winning Grammys.
Samir: Why don't you just go by Mike, instead of Michael?
Michael Bolton: No way! Why should I change it? He's the one who sucks.
Well, why not kill the MX for your normal domain and simply use a subdomain for a while (maybe, me.mydomain.com vs mydomain.com. At least then, all bounces won't resolve, and you can have your domain back.
-
ping -f 255.255.255.255 # if only
At the moment it looks like I may never be able to have any @gelhaus.net e-mail again.
Since the return addresses are all invalid (e.g. 'nonexistent_address@gelhaus.net'), I would think it would be simple to filter out all messages that aren't specific ones I've set up (e.g. 'valid_address@gelhaus.net').
See that, you answered your own question. Just block invalid addresses.
I've had this happen before to my domain, and eventually it died down. If it doesn't die down for you maybe you could track the spammer down and sue her.
Any sane protocol would never suffer from this problem. Yet people still claim that email is not broken...
If you find that the jobber is indeed an American, though, if I recall correctly, you can sue for damages. Of course, you generally have to find the scumbag first.
This sig no verb.
So, what happens when the receiving e-mail server tries to verify account name too? The spammer has to use someone's real account name (which has happened to me more than once). Since the spammer is using his own mail server to send the messages, your account and domain names don't only get checked ageanst your mail server when the recipient server tries to verify that they exist and not when the spam is originally sent. Thus, it's almost impossible to prevent.
Your only hope is finding the spammer somehow and making them miserable in some way (getting their ISP to cut them off, legal action), but that usually leads to the spammers friends making an exaple out of you (yet more unfortunate personal experience). I would just wait it out. Your ISP is doing the only thing they can by disabling your domain's e-mail. Soon, the "from" lookups will start failing for the spammer and he/she'll have to pick someone else to impersonate. I hope that your ISP will let you re-enable your domain's e-mail when it blows over. Good luck!
US Democracy:The best person for the job (among These pre-selected choices...)
My host is set up so that all emails recieved that have no account (invalid email address) are forwarded to an account with a quota of 1K. Of course the quota is full, so it is an instant bounce. Problem solved. Hope this may help you.
It's simple really!o gogogoch.com?
All you need to do is get a *really* long domainname.
For instance, would you expect any spam to originate from llanfairpwllgwyngyllgogerychwyrndrobwllllantysili
I think not!
Yet I'm sure there's at least a postmaster account running there (and surely a real account or two, even if just for fun's sake).
I wish I could offer some helpful advice but I can't, so instead I'll relate similar experiences I've had.
I have two domain names, one personal, one business.
The personal one was 'hijacked' in a very bizarre way a few years ago. I annoyed the owner of a popular site (by publishing an article about him swindling his visitors) so he posted my address dozens of times, all over the front page of his site. Obviously he wanted anyone who still believed his side of the story to send me hate mail, and that's exactly what happened. That was mailbombing though. The 'hijacking' was secondary, because of course my e-mail address is now in the address book of hundreds, if not thousands of people who are, let's say, not spectacularly bright. You can imagine how many e-mail viruses I get as a result of being in those address books.
The problem with my other domain is someone sending out viruses with my business address as the return address. This results in lots of auto-rejections from ISP spam filters. It's an inconvenience but it is NOTHING like as bad as the 30,000 you're getting, so you have my sincere sympathy. It must be very depressing to have something like this happen on such a large scale, and I do hope you figure out a way to prevent it.
Then how will legitimate mail arrive?
That still exists?
Hilary Rosen's speech was about her love of money and her desire to roll around naked in a pile of money.
http://spf.pobox.com/
Sure, not many MTA/MUAs check SPF records yet, but the fact that you are working to keep people from 'joe-jobbing' you should make your isp happy.
Awesome furniture, accessories and cabinetry in Santa Rosa, CA: http://humanity-home.com/
I am nearly in the same situation like you, except that I have complete control of my domain name (slett.net). I run my own DNS, my own SMTP server (Exim with SpamAssassin at SMTP Time), etc.. A nice side benefit is the ability to teergrube spammer hosts.
If you are technically inclined, and you have a broadband connection, this is definitely the best way at present to take control of spam.
Incidentally, I believe the ultimate solution to spam must involve banks and financial institutions - basically, an international mandate for these to not honor payment requests (e.g. credit card payments) to spammers. In the mean time, a mandatory upgrade or replacement to the SMTP protocol, to provide foolproof sender validation (by way of private/public keys or similar), will certainly go a long way towards solving the problem.
-tor
I work for a hosting company, and yes we've had this problem, although not on such a massive scale. We found that by removing any catch-all type setup, and bouncing the email address, the end users are much happier. This of course doesn't change the loading on the server much. IF however you know which IP's the emails are being sent from, your ISP can block those IP's with iptables, or, even in their router.
You shouldn't be so SOL, in my opinion.
Nobodies Prefect
Tidbits for Techs Technology Blog
US Democracy:The best person for the job (among These pre-selected choices...)
A brief investigation of a few of the bounces revealed that the spammer was using a variety of email addresses and domains in the message as their contact point. Many of the domains shared the same mail server, which was obviously a co-lo box, so she simply pointed all of the MX records for her domain towards the spammers primary email server. Unfortunately it wasn't misconfigured to actually accept the bounces, but each bounce was tying up resources and bandwidth belonging to the spammer. When she reset the MX records back a month or so later it was all over.
This is only applicable if you have your own domain like in this instance of course, I doubt an ISP would even consider this course of action with one of their subdomains as it's a dubious course of action to say the least. You also lose all use of your domain while the MX records as repointed, so you better be *damn* sure nothing sensitive is going to be received in legit email because the spammer could, if they wanted, accept and read your email.
Interesting and apparently effective strategy though.
UNIX? They're not even circumcised! Savages!
I have a question. Since we have certificates from Trust Authorities to do secure http, why can't we use those same certificates to do Secure SMTP? Since it would be a new protocol, it wouldn't need to be backwards compatible with SMTP except that the MTA might fall back to that as a last resort. Being able to verify that a message is actually being sent by acmewidgetcorp.com would certainly make it easier to separate junk from business communications. It would be much more difficult to abuse since a certificate could be revoked by the CA and there is a cost associated with obtaining them as well as the time involved.
Set rules in yer MDA. Alias work for this. Legitimate addressies get delivered to the appropriate box. Yer last alias is *. This one has a mailbox /dev/null.
Any mail not intended for a named recipient /will/ use bandwidth - then go "poof"...
"Flyin' in just a sweet place,
Never been known to fail..."
Yep, would help the guy but _not_ his ISP. His ISP probably does not want to waste the bandwidth created by 30k messages and that is whey they disabled his email. Bouncing, forwarding to /dev/null etc do not help because he will already have accepted the email (and thus wasted the bandwidth).
So either you scan already while receiving the email (as several people mentioned before, scan the header for invalid sender ips and then discard the bounces immediately BEFORE the whole email is accepted) or just wait it out.
I feel sorry for everyone out there whose domain gets used that way... =( I hope it dies down soon as his ISP does not seem to want to try to filter.
~Squisher
It is a long shot, but if you can track these people down, you have plenty of grounds for a lawsuit against them. Just prove they used your idenity without your permission. Even if they are in one of the few countries that won't help you out, there is a good chance that they have backers in a country, and you can sue the backers. Or if you can find who they are, and who the customers are, you can get the goverment to watch money transfers, and force all customers money inro your account (A very big maybe here). But you need a lawyer to 1) win the case for you, and 2) tell you how you can collect.
Good luck, but I urge you to do this. You should have plenty of grounds, and you might join the few guys who have actually shut down a spammer.
Thanks to everyone who's posted replies on my topic. I've worked with my hoster to change my default alias to route messages with an invalid address to oblivion. Until this happened I didn't even realize that I had a default alias set up, which shows how dangerous a little ignorance can be. We're now re-enabling my aliases one at a time and watching closely to make sure these valid addresses are not being overrun with this returned spam.
By the way, I should mention that my hosting service, CubeSoft, has been very good through all this. I've been in constant contact with them through e-mail (but not my domain e-mail, hah), and they have been very helpful in suggesting solutions and in trying to work with me rather than just blowing me off as not their problem. After this, I can strongly recommend them as a hosting provider.
(as I have done) instead of using your webhosting service's free email service, you just use a yahoo mail account? I don't think (tho I may be wrong) that yahoo would react the same way a normal webhost would because a) they should know that I at least didn't send all those spams from my account, and b) they probably filter a friggin' billion spams a day already.
:-)
My general opinion is that a division of labor should be kept between web page hosting and email hosting, even tho, of course, the server is designed to handle both services; perhaps the cost of setting up one of those 100M yahoo email accounts may be justified. The Chinese say "pay a lot, cry once" -- perhaps this is a side effect of the 'free pop emails' that hosting services always offer.
It's also *really* nice to be able to access your email from anywhere in the world (yeah, I know you can access your pop from anywhere, too, but it's definitely more of a pain). That, combined with need to never spam your contacts with a new email address (tho that's what the old Napster guy's working on now) every time I change ISPs.
Of course, none of the spam horseshi*t is going to change until email accounts only receive email from address on "the list". Come programmers! I've got enough shtuff to do already -- I don't have time to write that (relatively) simple application
May the Peace & Blessings of our Creator be with you all,
bmac
Your life will change within minutes of seeking to deliver your spirit back to our Creator within your lifetime -- www.mihr.com
Your host should be able to disable the catch-all account for your domain, which will result in any message not sent to a specific account being bounced.
You should also be able to set up filters in your accounts control panel. If your host does not support this, you need a new host.
I recently needed to respond to an e-mail from a small company. When I replied, my e-mail was bounced back to me because Comcast.net's SMTP server was blackholed. (This happened even though I have my own domain name and only use Comcast's SMTP server as a smarthost.)
To get around this, I changed Sendmail to start sending out mail directly inside of using a smarthost. Now I get bounces from people with AOL addresses because AOL somehow knows that I am using a dynamic IP address to send mail from.
The only reason I am having any of these problems at all is because of spam. Spam is ruining the Internet and what's worse, I can see no way of fixing it that doesn't destroy privacy.
Thanks for letting me vent.