Linux VPN Guide will
help you to create a Linux-based VPN server which can be accessed from
any Windows client (primarily Windows XP, but can be extended to work
with older Windows operating systems such as Windows 2000 or 95/98/NT).
I wrote this guide because of the pain I went through to get VPN up and
running at my work. After a couple of days of struggling and heavy
debugging I was able to set up a VPN server on Fedora Core 2 and 3 with
latest updates. I decided that VPN installation should not be such a
nightmare and tried to make everything as simple as possible in this
guide.
Linux-Windows VPN Guide with OpenSWAN, l2tpd and pppd
1) Who should use this guide?
If you have a
private network that you want to be able to access remotely in VPN
(virtual private networking) environment (for example, if you have
servers at your workplace that are firewalled from outside with
internal IP addresses such as 192.168.x.x and you want to be able to
access them from the internet), this guide is definitely for you.
2) Can this guide be used on production servers?
To
be honest, I'm not sure yet. I have been using this VPN solution for a
while now and haven't had any problems so far. But if you notice any
performance/compatibility/stability related issues, please let me know.
3) What about security?
VPN
is a secure virtual tunnel, which means security is built into it.
However, just like any "bullet-proof" solution out there, it has its
own security-related issues. First of all, this guide only shows you
how to use PSK (pre-shared security key) method of authenticating a VPN
tunnel. Your VPN tunnel is encrypted with this security key and in case
the key is stolen (or broken with a brute-force attack), your security
is compromised. You can use other methods of encryption based on SSL
(secure socket layer), but the process of setting up a VPN server and
connecting the clients to it is rather complicated. You can refer to Jacco's VPN tutorial
to set up a VPN solution based on SSL. As long as your PSK is very
long, unguessable and not prone to a simple brute-force, you should not
worry about security.
4) On what operating systems has this guide been tested on?
Installation
has been successfully tested on Redhat Linux 9.0 and Fedora Core
1/2/3/4, but it should also work on any other linux flavor. For BSD
systems, you would have to obtain source files for OpenSWAN and change
some configuration options on l2tpd to be able to compile everything
successfully.
5) Stuff we'll be needing (sources)
Download the following sources to a src directory i.e. /usr/local/src
- OpenSWAN 2.4.0 from http://www.openswan.org | Mirror
- l2tpd 0.69 from http://www.l2tpd.org | Mirror
- l2tpd Legacy PTY patch from this site
- l2tpd SysV PTY patch from this site
- l2tpd startup file from this site
6) Assumptions
It
is assumed that ppp is already installed on the system and you are
running the latest version of linux kernel 2.4 or 2.6. You can check if
you have ppp installed on the system by running a query in your package
manager ("rpm -q -a | grep ppp" in Redhat or Fedora). You can check the
kernel version by running "cat /proc/version".
7) OpenSWAN installation and configuration
I
recommend downloading an RPM from openswan.org instead of the source
tarball, since the RPM has the patches integrated to fit your system.
In case you want to compile OpenSWAN yourself, you will have to find
out which patches you need for your system. Compilation and
installation from source is pretty straightforward - just read the
README file and follow the instructions. If you want to install from
RPM, first make sure that you don't have any previous versions of
OpenSWAN in your system. Most probably you don't, but Fedora Core 3 was
shipped with an old version of OpenSWAN and you need to remove it
before installing OpenSWAN from source or RPM. Run the following
command to see if you have the package: "rpm -q -a | grep openswan". If
the query returns anything, remove it by typing "rpm -e openswan". If
openswan is not installed, proceed with the installation.
# cd /usr/local/src
# rpm -i openswan-2.4.0-1.i386.rpm
# rpm -i openswan-doc-2.4.0-1.i386.rpm
The openswan-doc rpm file includes all the documentation which you might need in the future.
The OpenSWAN installation has a sample ipsec.conf file included. Let's go ahead and rename it:
# mv /etc/ipsec.conf /etc/ipsec.conf.old
Now copy-paste the following into /etc/ipsec.conf
version 2.0
config setup
interfaces=%defaultroute
klipsdebug=none
plutodebug=none
overridemtu=1410
nat_traversal=yes
virtual_private=%v4:10.0.0.0/8,%v4:172.16.0.0/12,%v4:192.168.0.0/16
conn %default
keyingtries=3
compress=yes
disablearrivalcheck=no
authby=secret
type=tunnel
keyexchange=ike
ikelifetime=240m
keylife=60m
conn roadwarrior-net
leftsubnet=192.168.0.0/16
also=roadwarrior
conn roadwarrior-all
leftsubnet=0.0.0.0/0
also=roadwarrior
conn roadwarrior-l2tp
leftprotoport=17/0
rightprotoport=17/1701
also=roadwarrior
conn roadwarrior-l2tp-updatedwin
leftprotoport=17/1701
rightprotoport=17/1701
also=roadwarrior
conn roadwarrior
pfs=no
left=150.150.150.150
leftnexthop=150.150.150.1
right=%any
rightsubnet=vhost:%no,%priv
auto=add
#Disable Opportunistic Encryption
include /etc/ipsec.d/examples/no_oe.conf
Don't forget to replace "150.150.150.150" with your external IP
address and "150.150.150.1" with your default gateway. I'm not going to
explain what each line above does. If you want to find out for
yourself, please refer to OpenSWAN documentation. The above configuration should work for most people, though.
Now edit the file "/etc/ipsec.secrets" and put the following:
150.150.150.150 %any: PSK "a_very_long_string_up_to_256_characters"
The format of the file is: "external_network_address
connecting_from: PSK password". Change "150.150.150.150" to your
external IP address. If you want to be able to access the network from
anywhere on the Internet, leave "%any" intact. I recommend specifying
the IP address of the machine that will be used to initiate VPN
connections for security reasons though. PSK stands for "Pre-Shared
Key" - it's the key that will be shared by both the server and the
client for authentication. Make sure that the key is long (up to 256
characters) and very random. Every client trying to establish a VPN
connection will have to provide this pre-shared key.
8) l2tpd installation and configuration
First
of all, let me give you some information on l2tpd. The project started
out really well and a big community was involved in its development.
But seems like the project has been suspended for some reason and there
haven't been any new releases of l2tpd since version 0.69 which is more
than a year old. Latest versions of the linux distributions such as
Fedora and Mandrake are compiled with a new SysV pty system. The
current version of l2tpd only supports the old pty system and therefore
will not work on these distros unless its patched. Jacco de Leeuw talks about this problem in his detailed VPN guide here. An alternative is to switch to rp-l2tp,
which is another l2tpd daemon. But it is not a good way out, since it
does not support automatic IP allocations to new connections. Another
solution is to recompile the kernel with "Legacy (BSD) PTY support",
but most people simply don't know how to do that or are too scared to
do it. Recompiling the kernel might be a big problem for those who
don't have physical access to the machine. After many hours of
googling, I was able to find some patches that fix current l2tpd issues
and add some features such as binding l2tpd to a specific IP address.
Now here is the interesting part - the rp-l2tp package contains pty.c
code that is known to work with the new pty system. I tried to simply
copy-paste the script into l2tpd pty.c file, modified l2tpd.c to call
"pty_get" instead of the old "getPtyMaster", removed all other pty
calls and it worked! I was able to test l2tpd setup successfully on a
Fedora Core 3 machine without recompiling the kernel
I have
included two patches on this page. If you have a recent version of a
linux distribution with 2.6 kernel, go ahead and download the file
"l2tpd-0.69.sysv.patch". If you have an older version of linux with the
old pty system, go ahead and download the file "l2tpd-0.69.lpty.patch".
Now, let's install l2tpd and apply the patch:
# cd /usr/local/src
# tar zxf l2tpd-0.69.tar.gz
# mv l2tpd-0.69.sysv.patch l2tpd-0.69/
# mv l2tpd /etc/rc.d/init.d/
# cd l2tpd-0.69
# patch < l2tpd-0.69.sysv.patch
# make
# cp l2tpd /usr/sbin
# chmod 755 /usr/sbin/l2tpd
Running "make" should compile an l2tpd executable without errors.
You might get a couple of warnings, but no fatal errors. If you try to
compile the source without patching it first, you will most probably
get compilation errors (especially on systems with a newer gcc). Now
configure the startup environment:
# chmod 755 /etc/rc.d/init.d/l2tpd
# chkconfig --add l2tpd
# chkconfig l2tpd on
Let's move on to configuring l2tpd. The configuration files for
l2tpd should be placed in "/etc/l2tpd" folder. Go ahead and create the
folder, then copy paste the following into l2tpd.conf:
[global]
port = 1701
[lns default]
ip range = 192.168.1.101-192.168.1.254
local ip = 192.168.1.100
require chap = yes
refuse pap = yes
require authentication = yes
name = LinuxVPN
ppp debug = yes
pppoptfile = /etc/ppp/options.l2tpd
length bit = yes
The "ip range" line is the start and the end pool of the IP
addresses the clients will be given when they establish a VPN
connection (on their end). The "local ip" is server IP address - it's
used only when at least one connection is established. If you have a
different internal network, go ahead and change both lines. Everything
else should work for most people. Again, if you want to find out what
each line does, please feel free to refer to l2tpd documentation.
9) PPP configuration
L2TP
tunnels through PPP, which is why we need to configure it to work with
l2tpd. The l2tpd configuration above specifies the
"/etc/ppp/options.l2tpd" file as "pppoptfile" (PPP options file). Go
ahead and create this file and copy-paste the following:
ipcp-accept-local
ipcp-accept-remote
ms-dns 192.168.1.2
ms-wins 192.168.1.3
noccp
auth
crtscts
idle 1800
mtu 1410
mru 1410
nodefaultroute
debug
lock
proxyarp
connect-delay 5000
silent
Change the line "ms-dns" to your real DNS server and "ms-wins" to your WINS server (if you have any).
Now
let's create the authentication files. We use CHAP for PPP
authentication. Go ahead and edit the file "/etc/ppp/chap-secrets". The
format of this file should be "client server secret IP addresses". Here
is a sample file:
# Secrets for authentication using CHAP
# client server secret IP addresses
username * "password" 192.168.1.0/24
* username "password" 192.168.1.0/24
You need two lines for each username because it's a two-sided
authentication - one from client to server and one from server to
client. Both the password and the IP address pool should be the same
for both lines. The IP address network that we specify here
(192.168.1.0/24) means "allow this username to get an IP address in the
range 192.168.1.0 - 192.168.1.255". The l2tpd configuration file above
specifies the range to be 192.168.1.101-192.168.1.254, so only the IPs
in this range are given to the client on successful authentication. You
can change the networks and subnets in both files as needed. If you
need more than one account to be able to access the server, go ahead
and add it in the same format.
10) Starting and testing VPN
Everything is installed and configured. Let's go ahead and start the daemons:
Starting l2tpd:
/etc/rc.d/init.d/l2tpd start
Starting l2tpd: [ OK ]
Starting OpenSWAN:
/etc/rc.d/init.d/ipsec start
ipsec_setup: Starting Openswan IPsec 2.4.0...
ipsec_setup: insmod /lib/modules/2.6.10-1.741_FC3/kernel/net/key/af_key.ko
ipsec_setup: insmod /lib/modules/2.6.10-1.741_FC3/kernel/net/ipv4/xfrm4_tunnel.ko
Take a look at /var/log/messages and /var/log/secure and make sure
there are no errors. You should see notification messages such as
"l2tpd: l2tpd startup succeeded" and "ipsec_setup: Starting Openswan
IPsec 2.4.0...". If there are no errors, everything is up and running -
proceed to next step.
11) Firewall configuration
To
be able to route packets from your external to internal interfaces,
packet forwarding must be turned on. Edit /etc/sysctl.conf and change
"net.ipv4.ip_forward = 0" to "net.ipv4.ip_forward = 1". If you don't
have that file, just type "echo 1 > /proc/sys/net/ipv4/ip_forward"
and add this line to your /etc/rc.d/rc.local file. Restart your
networking by typing "/etc/rc.d/init.d/network restart" or whichever
way you restart your network interfaces.
Make sure that the
following ports are open: UDP 500 & 4500, TCP 4500. If any of these
ports are blocked, VPN will not work. If you have Redhat/Fedora
distributions with iptables, insert these lines into your
/etc/sysconfig/iptables file, before REJECT line at the end of the file:
-A RH-Firewall-1-INPUT -i ppp+ -j ACCEPT
-A RH-Firewall-1-INPUT -i eth1 -j ACCEPT
-A RH-Firewall-1-INPUT -m state --state NEW -m udp -p udp --dport 500 -j ACCEPT
-A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp --dport 4500 -j ACCEPT
-A RH-Firewall-1-INPUT -m state --state NEW -m udp -p udp --dport 4500 -j ACCEPT
One more thing you should keep in mind while changing your firewall
configuration - your internal and VPN network should be trusted,
otherwise your firewall will keep rejecting packets from 192.168.x.x
network into your LAN. I solve this by simply adding the first two
lines above into my iptables file. The interface eth0 is external and
eth1 is physically inside my internal network, of course
출처 :
희망의 나래님 블로그 보안 해킹에서
