2009/10/29 Athmane Madjoudj <athmanem@xxxxxxxxx>:
> On Thu, Oct 29, 2009 at 12:52 PM, jdow <jdow@xxxxxxxxxxxxx> wrote:
>> From: "Michael Cronenworth" <mike@xxxxxxxxxx>
>> Sent: Wednesday, 2009/October/28 16:03
>>
>>
>>> It seems in the past month brute force attacks are on the rise. They are
>>> targeting anyone listening on port 22 and go after root. If you do not
>>> have a hardened box, you will see thousands upon thousands of
>>> connections in your logs. Once logged in they will set your system up in
>>> their botnet.
>>>
>>> Google: dt_ssh5
>>> This little baby will get placed in /tmp and will be running. Looks to
>>> be a SSH gateway for the attackers for easy access/control.
>>>
>>> -Make sure your root password is not a dictionary word.
>>> -Add iptables rules to limit multiple connections on SSH to 4 within a
>>> minute.[1] Perhaps this needs to become a Fedora default.
>>
>> Once within 3 minutes is entirely practical and effective. In the last
>> two days a pair of dolts kept trying 6621 times and 2185 times after the
>> door slammed shut in their faces. Their ISPs have been notified.
>>
>>> -Update your system.
>>> -Use SELinux.
>>>
>>> Why am I sending this message? Is it SPAM? No. I've seen this hit a
>>> customer and cause an explosion in their network traffic. The backdoor
>>> was installed on Sept. 30th and was not detected until recently. Google
>>> results seem to indicate this past month with higher than normal brute
>>> force activity.
>>>
>>> [1]
>>> -A INPUT -i eth0 -p tcp -m tcp --dport 22 -m state --state NEW -m recent
>>> --update --seconds 60 --hitcount 4 --name DEFAULT --rsource -j DROP
>>> -A INPUT -i eth0 -p tcp -m tcp --dport 22 -m state --state NEW -m recent
>>> --set --name DEFAULT --rsource
>>
>> I love those rules and have been spreading them around for quite some
>> time now. I am glad to see somebody else has either adopted or discovered
>> the rule trick. It is devastatingly effective. Guessing a password as
>> simple as "mE3" would take decades of attempts. (Now I want to configure
>> sshd so that it logs the attempted password along with the attempted user
>> name.)
>>
>> {^_-}
>>
>> --
>> fedora-list mailing list
>> fedora-list@xxxxxxxxxx
>> To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
>> Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
>>
>
> You can install fail2ban
> #yum install fail2ban
>
> Links:
> http://www.fail2ban.org/
We've implemented a similar strategy here at the Genomics Center using
denyhosts.
When taking this to production, there are a few issues and problems
that can arise, that you could DOS your own servers if you're not
careful. We came up with a solution that i've documented in a blog
post recently, and so far it's been humming along nicely. I went away
for a week on vacation, came back, and there were no hiccups.
http://loupgaroublond.blogspot.com/2009/10/doing-denyhosts-bit-better.html
If people are interested, we have a puppet policy you can use for your
systems that implements this. The only trouble is the setup of the
policy is a bit complicated and custom. Namely, it's dependent on some
of our service accounts, and our firewall management tools.
-Yaakov
--
fedora-list mailing list
fedora-list@xxxxxxxxxx
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
