Users can be found via google:
http://***.google.com/search?hl=en&q=site:drupal.org&q=porn&btnG=Google%20Search
Other similar searches lead to further spam:
http://***.google.com/search?hl=en&q=site%3Adrupal.org+celebrity+&btnG=Search
http://***.google.com/search?hl=en&q=site%3Adrupal.org+interest+models&btnG=Search
http://***.google.com/search?hl=en&q=site%3Adrupal.org+interest+nude&btnG=Search
...

So please someone delete the spam. Anyway, the problem needs to be addressed at a higher level as urls like "interested in xxx" will stay in the results.
I propose to add following line to robots.txt
Disallow: /profile/interest/

Not much value for people to search for +duplicate content +paradise to build spam links +http 200 where it should be 404. ( http://drupal.org/profile/interest/12349087234095365 )

Comments

s.daniel’s picture

vm’s picture

http://drupal.org/user/297666 - profile cleansed, user blocked
http://drupal.org/user/263691 - profile cleansed, user blocked
http://drupal.org/user/291487 - profile cleansed, user blocked
http://drupal.org/user/268899 - profile cleansed, user blocked
http://drupal.org/user/268893 - profile cleansed, user blocked
http://drupal.org/user/268896 - profile cleansed, user blocked

Leaving this open to allow other eyes to take a look.

s.daniel’s picture

More Spam to delete:
google.com/search?hl=en&q=site:drupal.org&q=Soma+&btnG=Google%20Search
google.com/search?hl=en&q=site%3Adrupal.org+viagra&btnG=Search
association.drupal.org/user/733
Some spam can be found on pages like
drupal.org/profile/interest/celebrity
drupal.org/user/322614
...

vm’s picture

profiles cleansed - users blocked

s.daniel’s picture

d.o/user/349048

gerhard killesreiter’s picture

Would anybody mind if I blocked the "interest" profile pages in robots.txt? Might make it less interesting to spam d.o.

I'd so like to disable gmail.com as an email domain on d.o. 97% of all spammers that I blocked last week had a gmail.com address. Unfortunately, this would also affect existing accounts.

beeradb’s picture

I don't think it was a serious suggestion, but I'm definitely -1 to blocking gmail.com. As an example, my account is registered to a gmail.com domain. Although I agree it would solve a lot of problems, it would also be blocking what is arguably the most popular email host for legitimate users. I think the thing to do is to disable being able to register using the email+morestuff@gmail.com format to register d.o accounts. Additionally, we should consider using mollom or some other blackbox spam filtering software. I think the combination of the google captcha system(when registering gmail accounts), the disabling of plus addressing, and better filtering on d.o would help us see a reduction in spam registrations.

gerhard killesreiter’s picture

The suggestion was half-serious at least, I am quite annoyed. If I hd a way to not let it affect existing accounts, I'd implement it right now, followed by an open letter to google.

Blocking mail addresses using + notation would have avoided only about 30% of the spammy accounts.

Using some captcha before registration seems like a good idea, however, the spammers have already defeated google's captchas in creating the mail accounts, so I am quite sure they'll also defy ours.

I had also suggested to use some blocklist to check the urls that the spammers enter, hoever, when I manually checked some of the URLs, I didn't find all of them in the blocklists I looked at. Maybe I'd need to look at other blocklists.

I am not sure that mollom or a similar service would be able to filter the domains accurately either.

beeradb’s picture

I know category spam is only one type of spam we get, but what about the idea of having 'honeypot' categories? The idea behind this is that probably 90%-95% of users who use categories such as "sex", "nude", "porn" etc. are spammers, so what about automatically blocking user accounts which use those categories? An email could be dispatched to the webmasters list (or we could use any other form of communication) to inform people of the account being blocked, and all the webmasters would have to do is review the list periodically to undo any 'misses'.

As far as helping prevent spam users sitewide, I do think more detailed information on a few things would be useful:

1). What percentage of/how many active d.o users are using gmail accounts. (active being unblocked accounts)
2). What percentage of/how many blocked d.o users registered using gmail accounts.
3). What percentage of/how many blocked d.o users registered using gmail accounts with + addressing.

Obviously compiling this information would use the time of our already overworked infrastructure team, but I do thing it would help us make informed decisions about the best path forward.

gerhard killesreiter’s picture

The categories idea is not helpful. Most spam accounts don't bother with this anymore. They use a spammy name and a URL only. I suspect that these profiles might be spamvertized later.

1) 83496 unblocked users use gmail accounts.

2) 1219 blocked users use gmail.

3) There are 346 valid users using + notation with gmail and 175 spammers using +notation with gmail. The spammers are all very recent accounts.

we have a total of 344716 unblocked users and a total of 3756 blocked accounts. So, almost 1/3 of all spammers use gmail.

s.daniel’s picture

Is the spam posted mostly automated?
On g.d.o there seems to be less spam, is that correct?

Blocking areas of d.o through robots.txt is an easy way to reduce the negative effect done by spam.
The way I found most of the users was when I searched google like "site:drupal.org amanda" looking for someone here using amanda the backup software. The first result is d.o/profile/companies/amanda+hayes+p*rn and even though the user is deleted now d.o. will stay in the index for that term google.de/search?hl=de&client=firefox-a&rls=org.mozilla%3Ade%3Aofficial&hs=vAn&q=amanda+hayes+porn&btnG=Suche&meta=

I mean d.o gets a lot of traffic for terms it would normaly not rank for but I don't think people searching for "asiansex" or "p*rn wallpapers" are our target audience. Another reason for using robots.txt to block some areas is that having d.o associated with p*rn will probably harm the other rankings we would pefer.
google.de/search?q=porn+wallpapers

You can take a look here how many people come here because they search for p*rn related keywords.
http://drupal.org/node/193416
http://groups.drupal.org/node%252F6636

So
Disallow: /profile/interest/
Disallow: /profile/industries/

gerhard killesreiter’s picture

Status: Active » Fixed

Has happened.

Anonymous’s picture

Status: Fixed » Closed (fixed)

Automatically closed -- issue fixed for two weeks with no activity.

s.daniel’s picture

Status: Closed (fixed) » Active

http://drupal.org/user/424542
Others can be found here: google.com/search?q=site%3Adrupal.org+viagra+-spam&btnG=Search&hl=en&sa=2

Take a look at the first result in particular.

gerhard killesreiter’s picture

Status: Active » Fixed

thanks, I blocked about 10.

Status: Fixed » Closed (fixed)

Automatically closed -- issue fixed for 2 weeks with no activity.