If you expose a WordPress or a Windows server in their default settings to the 'net and don't keep on top of every single patch offered that's relevant to what you expose, you're asking to be descended upon.
Sophos is claiming that White Hat researchers are keeping the Script Kiddies in the game: http://nakedsecurity.sophos.com/2012/12/05/web-exploit-kits-whitehat/
On Mon, Mar 18, 2013 at 3:07 PM, Mark Hutchings mark.hutchings@gmail.comwrote:
Speaking of which, if you're running any kind of script on your site (WordPress, etc) make sure it is up to date. If there is a 0-day exploit out for it, the script kiddies are going down the list on Google searching for "Powered by WordPress" and seeing if your server is open to exploits. Most of the time this isnt caused by a single IP address, it's usually a botnet from around the world, but sometimes it can be.
Any way you could post some of the logs? Like show what kind of http request they were making?
On 3/18/2013 3:01 PM, Andrew Beals wrote:
It his pipe is full, then he has bigger problems than that which J. Random Unix Jock can explain over a mass e-mail. Especially when yum knows not of fail2ban.
Serving 404 pages to script kiddies shouldn't Bork a server. It shouldn't even put an appreciable load on it. Script kiddies are here to stay, thanks to the fringe members of the "information should be free" crowd. (Just as an example, there appear to be about 1.6k copies of The Anarchist Cookbook out there, ready for downloading.)
Andy Ps. There are too many kittens - please spay/neuter your pets.
Any typos are the direct result of Swiftkey X's autocorrect function. On Mar 18, 2013 2:40 PM, "Billy Crook" billycrook@gmail.com wrote:
Every time you use a route table as a firewall, God kills a kitten.
If you want a firewall, use..... a firewall. iptables is the command.
If you want something that scales, and won't require your time to maintain a shitlist of IPs; use fail2ban, and it will manage the list per your specifications.
Most likely, your DoS is apache-local. i.e. they aren't actually flooding your entire pipe. If you use fail2ban/iptables, this should fix you right up.
If they are flooding your actual pipe, you need to apply the filter on the far end of your pipe. i.e. Get your ISP (or a new isp) that will let you administer an ACL on the router on THEIR side of your line. Or get a DDoS prevention service. Blocking on the sonic wall will have NO affect on a flood if the sonic wall is at the same site as the targeted server.
Fail2ban can integrate with this remote filtering too. You simply modify fail2ban's 'action' to call a script that adds the IP upstream.
On Mon, Mar 18, 2013 at 2:27 PM, Andrew Beals andrew.beals@gmail.com wrote:
If they're coming from just the single IP, then black-hole'ing their IP
is
easier. If the address they're coming from is 128.115.1.1, then simply paste this at a shell prompt and give it your password when sudo asks
for
it:
sudo route add 128.115.1.1 gw 127.0.0.1 lo
This will cause all packets destined to go back to them to get dropped
on
the floor and should be sufficient. You'd really prefer to do this (or
just
add them to the naughty list which is something that I believe the SW
can
do, even with ancient builds of their SW) on your SonicWall box, but
you can
get away with doing it on your server.
Adding an IP tables (again, if you can't convince your SW to just drop packets from them) is more efficient, of course, but it's hairier to
set up.
On Mon, Mar 18, 2013 at 2:19 PM, J. Wade Michaelis jwade@userfriendlytech.net wrote:
I have a CentOS web server that has recently been brought to a halt on
two
separate occasions. Checking the access.log, it appears that it was a Denial of Service (DOS) attack (hundreds of HTTP requests in a very
short
time, all from a single IP address).
I want to prevent these types of attacks from bringing the server to
its
knees. We have a hardware firewall (SonicWall) in place, but it isn't
quite
new enough to run the firmware that allows rate-limiting.
I have found a number of tutorials that show how to do this type of
thing
with IPTABLES. Is there a better solution?
Supposing I go with IPTABLES, do I need to include rules to allow FTP
and
SSH (the only other services on the server)?
Would any of you be willing to assist me with this?
Thanks, ~ j. jwade@userfriendlytech.net
KCLUG mailing list KCLUG@kclug.org http://kclug.org/mailman/listinfo/kclug
KCLUG mailing list KCLUG@kclug.org http://kclug.org/mailman/listinfo/kclug
KCLUG mailing listKCLUG@kclug.orghttp://kclug.org/mailman/listinfo/kclug
KCLUG mailing list KCLUG@kclug.org http://kclug.org/mailman/listinfo/kclug