| 1 |
Hi, |
| 2 |
|
| 3 |
Le mercredi 25 octobre 2006 16:29, Longman, Bill a écrit : |
| 4 |
> > I've seen many recommendations for port knocking, but I feel that's |
| 5 |
> > unnecessarily complex when compared to simply changing the port |
| 6 |
> > sshd listens on. While the use of port knocking no doubt further |
| 7 |
> > decreases your exposure over an alternate sshd port, the difference |
| 8 |
> > is only a small percentage of the benefit you receive from moving |
| 9 |
> > away from port 22 in the first place. |
| 10 |
> |
| 11 |
> I've moved most of my public SSH ports off 22 because it reduces by |
| 12 |
> thousandfolds the script kiddies playing with their toys and filling |
| 13 |
> my logs and pipes. There is no more efficient means that will give |
| 14 |
> you such returns with such little effort. |
| 15 |
> |
| 16 |
> I only wish I'd done it sooner. Yeah, you'll still get port scanned |
| 17 |
> and someone will snoop around, but that's not in the face of the |
| 18 |
> storm on port 22. |
| 19 |
|
| 20 |
This is REALLY not a solution against bruteforce attacks... You say the |
| 21 |
reason : a port scan and the bruteforce can begin... |
| 22 |
|
| 23 |
Solutions could be : |
| 24 |
|
| 25 |
- using a daemon that analyse the logs and blacklist the guy (with |
| 26 |
iptables or host.deny) after X login tries, such as fail2ban (in |
| 27 |
portage), or |
| 28 |
|
| 29 |
- using ipt_recent like this (for example) : |
| 30 |
|
| 31 |
iptables -A INPUT -p tcp --dport ssh -m state --state NEW -m |
| 32 |
recent --set --name SSH |
| 33 |
iptables -A INPUT -p tcp --dport ssh -m state --state NEW -m |
| 34 |
recent --update --seconds 120 --hitcount 5 --rttl --name SSH -j DROP |
| 35 |
|
| 36 |
my 2 cents, |
| 37 |
Guillaume |
| 38 |
|
| 39 |
-- |
| 40 |
Guillaume Castagnino |
| 41 |
guilc@×××××××.net / casta@×××××.info |
| 42 |
|
| 43 |
-- |
| 44 |
gentoo-hardened@g.o mailing list |
Archives