[asterisk-users] Asterisk SIP attacks and sshguard

[Joe Greco]

Hello,

We had been seeing SIP-guessing attacks on our Asterisk server here.

While it wasn't that hard to write a once-a-minute cron job to spank
the lusers, that runs once a minute and creates little spikes in the
usage and I/O graphs, and is slower to respond than I'd really prefer.
I felt that it'd be much cooler to get something more comprehensive 
put together.  We don't use fail2ban because I don't like having to 
install python.

sshguard is a high-performance compiled C application that can run
off a log file or a pipe from syslogd to sshguard, meaning that it
can respond a lot more quickly than once a minute, and works with
very modest overhead on the host system.  It also has features such
as touchiness, so that it can get tougher on a miscreant as time goes
on; my own shell script is naive in that once it passes a threshold,
there's just a permanent rule generated.  This worries me if I ever
have a situation where a legitimate remote client gets messed up and
tries the wrong password or something like that; sshguard does a much
nicer job in this regard.

In any case, my initial attempts to create rules for sshguard didn't
work right, quite possibly because I don't often work in LEX/YACC.
I submitted a request to the sshguard guys suggesting new rules.

http://www.sshguard.net/support/submission/detail/49ce7182028d8b6f3e3d/

and on their mailing list, a little more:

http://sourceforge.net/mailarchive/forum.php?thread_name=F4E10075-5D93-43B4-B73A-1FD217BE130D%40sshguard.net&forum_name=sshguard-users

In particular, they're looking for log examples of some of those 
messages, but I have no idea how to generate the conditions that would
cause these messages.  I'm also not sure if there's a way to disable
color codes in the Asterisk log files; we log indirectly via BSD's
"logger"

# asterisk -vvv 2>&1 | logger -t asterisk

so it may be thinking that the console is color-capable.  We use this
method because this forces them through the syslog mechanism; we need 
that for centralized logging, and it's handy for things like sshguard
too.

Specifically looking for examples of (or how to generate)

1)	.*No registration for peer '.*' (from <HOST>)
2)	.*Host <HOST> failed MD5 authentication for '.*' (.*)
3)	.*Failed to authenticate user .*@<HOST>.*

If anyone who is more familiar with the attacks or how to generate
these messages would give me some assistance, or chime in on the
sshguard-users list, that'd be most appreciated.

Thanks.

... JG
-- 
Joe Greco - sol.net Network Services - Milwaukee, WI - http://www.sol.net
"We call it the 'one bite at the apple' rule. Give me one chance [and] then I
won't contact you again." - Direct Marketing Ass'n position on e-mail spam(CNN)
With 24 million small businesses in the US alone, that's way too many apples.