From time to time, we receive a message from Computing Services alerting us to user-generated spam on a site on the Contributed webserver.

Where we need to go to react to such cases depends on what we need to block:

User-level block

For pwnage that ISO is unhappy about we follow roughly the following script (stuff we notice doesn't have to be quite as drastic):

ssh contrib-cgi    # or other appropriate machine(s)
ksu
cd /var/apache/andrew-contrib
mv -i usr/<username> suspended/

...then send an email to the appropriate contact so they can fix it.

Also send a note to ISO along the lines of: "Thanks for the notice. We've suspended serving CGIs for this Andrew user on Contrib until they have resolved these issues."

Per-file blocking

If we can be more selective/strategic about what to block than a whole-user ban, edit /etc/apache2/main_vhosts/contrib.conf on both front-ends and add whatever Apache config directives are needed to have the desired effect. Make sure to block the file over both HTTP (port 80) and HTTPS (port 443); this requires editing two different sections of the conf file.

In particular, /etc/apache2/main_vhosts/contrib.conf has two sections (one for 80 and one for 443) under the comment:

    # SPECIAL BLACK-HOLED FILES
    # Leave a comment with your name, the date, and any INC references.  Prefer
    # to use HTTP 410 ("Gone") when it's more correct.  Don't forget to
    # black-hole the file for both port 80 and port 443 (two different sections
    # of this conf file).  Comment out old rules rather than deleting, to
    # preserve the examples.

…which, if people are following those instructions, should have some examples you can work from.


CategoryContributedServices

Services/Contrib Spam (last edited 2018-11-26 17:01:21 by tparenti@CLUB.CC.CMU.EDU)