Thwarting Spammers
Charly's Column – grepcidr
Often it is the very simple tools that, when used appropriately, lead to the greatest success. This time, sys admin columnist Charly employs an IP address filter to count the devices in his home and trip up spammers to boot.
Although Linux has many grep
variants, you can always find a new one. I only discovered grepcidr
[1] a few months ago. As the name suggests, the tool filters input by IP addresses and networks. It works equally well with IPv4 and IPv6. To show grepcidr
's capabilities, I will use it to compile a list of all IPv4 addresses on my home network. I got this from the Syslog on the firewall, which is also the DHCP server:
cd /var/log grepcidr 10.0.0.0/24 syslog|grep DHCPACK|tail -n 1500|cut -f9 -d" "|sort|uniq > 1stlist
The 1stlist
file now contains 46 IP addresses:
10.0.0.135 10.0.0.15 10.0.0.150 10.0.0.16 10.0.0.166 [...41 more...]
I automated this discovery process with the following command:
grepcidr 10.0.0.0/24 -c < ./1stlist
In this simple case, it would have been faster with wc -l
, but grepcidr
ultimately shows its strengths when you have to filter out different networks from such a file.
After a while, I repeat the game and write the list of IP addresses written to the 2ndlist
file. For my statistics,
grepcidr 10.0.0.0/24 -c < .2ndlist
shows that the file is one line shorter, so there is one less device on the network. I can easily find out which one is missing with diff
:
diff 1stlist 2ndlist 2d1 < 10.0.0.15
To see how many devices in my house are run 24/7, I send the data to a round-robin database and have a history graph drawn. The interruption at about 6am in Figure 1 is a routine reboot of the DHCP server.
Fighting Off Mail Pests
I also used grepcidr
to create IP blacklists on my mail server. I have set up some mail addresses that I don't use but that spammers will typically test for their existence: sales@...
, for example. Twice a day, grepcidr
plows through the log and extracts all the IPs from the log lines that relate to these mailboxes and writes them to the harvest
file.
Because these lines also contain my own servers' IPs, I have to remove them with a whitelist, which is also a file with IP addresses.
The ready-to-use blacklist is now built by grepcidr
in a single step:
cat harvest|grepcidr -vf whitelist > blacklist
This trick has helped me keep many a spammer out of my life.
Infos
- grepcidr: http://www.pc-tools.net/unix/grepcidr/
Buy this article as PDF
(incl. VAT)