Written by Martin Maartensson
attachments and a more readable markdown on github << this is a link
We have set up 3 images on Nebula, one "baseVM" (Server 18) and then 2 individual VM's, then we have created a whole environment of VM's inside the baseVM where we are managing the network from one of the internal VM's (Server 29).
First we install the Suse 13 distobution on Nebula and assign 2 nics
After the installation its important to do a few configurations
systemctl stop SuSefirewall2
systemctl disable SuSefirewall2
systemctl stop wicked
systemctl disable wicked
systemctl start sshd
systemctl enable sshd
zypper update
zypper install zsh # for nicer shell
zypper install vim-data # for syntax in vim
useradd {USERNAME} -m -s /bin/zsh -g wheel
passwd {USERNAME}
echo "server18" > etc/hostname
- Power off the host and remove the susecd from hosts
- Go to edit
/boot/grub2/grub.cfg - Find all the timeouts and set them to 0
set timeout=0Go to /etc/ssh/sshd_config and set these values
PermitRootLogin no
PasswordAuthentication no
We will setup libvirt and install a more modern version of suse so that we can do more stuff easier.
libvirt can be installed directly from zypper
zypper install libvirt
then to make installation easier we will install virt-install and accept all
zypper install virt-install
We need to get the latest version of suse leap
this is the link to download Suse Leap
we can save the file in the /var/lib/libvirt/boot/ folder as suse.iso
then we will create the installation script in /root/ and call it virtinstall.sh and give it execution rights chmod +x virtinstall.sh
#!/bin/bash
cd=suse
name=suse
virt-install \
--virt-type=kvm \
--name $name \
--ram 2048 \
--vcpus=2 \
--os-variant=generic \
--cdrom="/var/lib/libvirt/boot/$cd.iso" \
--network=bridge=br0,model=virtio \
--network=bridge=br1,model=virtio \
--network=bridge=br2,model=virtio \
--graphics vnc \
--disk path=/var/lib/libvirt/images/$name.qcow2,size=10,bus=virtio,format=qcow2notice that we have defined the 3 bridges br1, br2 and br3. We need to create them before we sucessfully can run virtinstall.sh
The full script can also be found in the scripts folder
ip link add link eth0 name eth0.100 type vlan id 100
ip link add name br0 type bridge
ip link add name br1 type bridge
ip link add name br2 type bridge
ip link set up dev br0
ip link set up dev br1
ip link set up dev br2
ip link set dev eth0 master br0
ip link set dev eth1 master br1
ip link set dev eth0.100 master br2
ip link set up dev eth0
ip link set up dev eth1
ip link set up dev eth0.100
ip addr add dev br0 "192.168.150.18/24" brd +
ip route add default dev br0 via "192.168.150.1" scope global
echo 1 > /proc/sys/net/ipv4/ip_forwardNotice we are not assigning any ip address for either eth0 or eth2, this is not necesarry because the br0 is now representing eth0 and we can access the host machine with the ip address for br0. br1 is connected to eth2 on the link layer and we can assign the ip address for that one later inside the VM. so lets create the VM.
We can now run the script virtinstall.sh that we created ealier.
./virtinstall.shThen we will wait a few second until the scripts is done.
We now need to connect to the vm via VNC so we will forward the port of the VM's vnc to our local workstation
ssh -J \
[email protected] \
[email protected] \
-L 5900:localhost:5900The first VM's vnc port will be on 5900, if we make more then the port will increment.
When we are in vnc we just follow the installation through all the steps and restart the VM.
We now will open virsh on the main server18 and start the VM and enable autostart.
virsh start suse
virsh autostart suseThen we can open vnc again and check the ip of the VM so that we can ssh into the vm and set it up, but lets first setup the other 3 VM's
the only difference here compared to the Suse Leap installation is that we download the image of ubuntu server 20.04 and the latest version of archlinux, then we change the name and the cd in the virtinstall.sh script and also we change the bridges they receive.
arch linux will be in the same subnet as server18 and server29, so we give him br0
#!/bin/bash
cd=arch
name=arch
virt-install \
--virt-type=kvm \
--name $name \
--ram 2048 \
--vcpus=2 \
--os-variant=generic \
--cdrom="/var/lib/libvirt/boot/$cd.iso" \
--network=bridge=br0,model=virtio \
--graphics vnc \
--disk path=/var/lib/libvirt/images/$name.qcow2,size=10,bus=virtio,format=qcow2Ubu1 will be in the internal network so he will get br1
#!/bin/bash
cd=ubuntu
name=ubu1
virt-install \
--virt-type=kvm \
--name $name \
--ram 2048 \
--vcpus=2 \
--os-variant=generic \
--cdrom="/var/lib/libvirt/boot/$cd.iso" \
--network=bridge=br1,model=virtio \
--graphics vnc \
--disk path=/var/lib/libvirt/images/$name.qcow2,size=10,bus=virtio,format=qcow2Ubu2 will be in the VLAN so he will get br3
#!/bin/bash
cd=ubuntu
name=ubu2
virt-install \
--virt-type=kvm \
--name $name \
--ram 2048 \
--vcpus=2 \
--os-variant=generic \
--cdrom="/var/lib/libvirt/boot/$cd.iso" \
--network=bridge=br2,model=virtio \
--graphics vnc \
--disk path=/var/lib/libvirt/images/$name.qcow2,size=10,bus=virtio,format=qcow2That was it then we just need to autostart them all with virsh
virsh autostart arch
virsh autostart ubu1
virsh autostart ubu2 At this point there is no internet for any of the hosts and we will need to setup Server29 who has access to all the network interfaces, and is gonna be the main router of the system.
At this point we have not accessed server29 before and we need to do the same initial process as we did with server18 in the post installation.
- Disable default firewall
- Disable wicked
- Enable SSH
- Update the system
- install zsh
- install ranger (visual terminal file manager)
- Create users and make them sudo
- Change hostname
- Optimize reboot speed
- Seccure SSH
- Enable ip forwarding
and because we are gonna use this one alot we will also install oh-my-zsh and setup the ~/.zshrc with the follwing settings.
export ZSH="/root/.oh-my-zsh"
ZSH_THEME="kardan"
export UPDATE_ZSH_DAYS=30
plugins=(git extract)
source $ZSH/oh-my-zsh.sh
export EDITOR='vim'
alias r="ranger"
unsetopt beep
bindkey -v
zstyle :compinstall filename '/root/.zshrc'
autoload -Uz compinit
compinit
[ -f ~/.fzf.zsh ] && source ~/.fzf.zsh
This will make the terminal look, and feel very nice. We will also be able to edit our commands with very good vim support.
and then we will setup the network script for the VM
This network script is similar but still differen from the base server, since we will assign ip addresses for each interface. Also we will setup DNS since it is not done automatically for us without wicked
#/root/bin/network.sh
...
echo "search omicron2.eitlab.diplom.dtu.dk" > /etc/resolv.conf
echo "nameserver 192.168.150.1" >> /etc/resolv.conf
ip addr add "192.168.150.29/25" dev eth0 brd +
ip addr add "10.100.0.1/24" dev eth1 brd +
ip addr add "192.168.169.1/24" dev eth2 brd +
ip link set dev eth0 up
ip link set dev eth1 up
ip link set dev eth2 up
ip route add default dev eth0 via 192.168.150.1 scope global
...Then we need to add a service file for the script and enable it in systemd
all the full scripts and service files for each VM is attached
Now we can setup DHCP.
So we have 2 interfaces now that we need to manage DHCP for.
- br1 (eth1) - 192.168.169.0/24
- br2 (eth2) - 10.100.0.0/24 :: VLAN
First lets install dhcp-server
zypper install dhcp-serverthen we will edit the file /ect/sysconfig/dhcpd and add eth1 and eth2 as our dhcpd interfaces
DHCPD_INTERFACE="eth1 eth2"
then we will setup the subnet for the dhcp server in /eth/dhcpd.conf
ddns-update-style none;
default-lease-time 21600;
max-lease-time 43200;
subnet 192.168.169.0 netmask 255.255.255.0 {
option routers 192.168.169.1;
option domain-name "omicron2.eitlab.diplom.dtu.dk";
option domain-name-servers 192.168.150.1;
range 192.168.169.101 192.168.169.130;
}
subnet 10.100.0.0 netmask 255.255.255.0 {
option routers 10.100.0.1;
option domain-name "omicron2.eitlab.diplom.dtu.dk";
option domain-name-servers 192.168.150.1;
range 10.100.0.101 10.100.0.130;
}
then we need to start and enable dhdpd
systemctl enable dhcpd
systemctl start dhcpdthats it now clients connected to the same interface as eth1 will get an ip address from our dhcp server.
At this point it might be a good idea to restart server18 with everything inside it.
On the host machine server18 we have set up a vlan eth0.100 which is controled by a bridge br2, the brigde is then passed into the router vm server29
server29 is functioning as a firewall and DHCP server for the VLAN,
in libvirt on server18 we created another vm ubu2 which only has one nic, which is assigned to the VLAN eth0.100 through br2 and is therfore assigned an ip address by server29 in the range 10.100.0.101 to 10.100.0.130
If we try to access the internet from ubu2 we will not get any response since there is no route from the vlan to the internet. However we have a route to the host server29 on the link layer and can access all its ip addesses 10.100.0.1, 192.168.169.1 and 192.168.150.29. This means we can connect to it via ssh.
So on server29 we create a new user called jumper and then we go to our sshd config in /etc/ssh/sshd_config and add this to the bottom
Match User jumper
AllowTcpForwarding yes
PasswordAuthentication no
ForceCommand echo 'forward to backend'
PermitOpen 130.225.170.70:22022
Now we want to login to the server on 130.225.170.70 and create a similar user as jumper and set similar rules in /etc/ssh/sshd_config
so first we add the user:
useradd -m linux
then we add the config to the end of sshd_config:
Match user linux
ForceCommand echo 'hello fello'
PasswordAuthentication no
AllowTcpForwarding yes
PermitOpen localhost:443
We will allow localhost on port 443 since this is also the port we expose from that server to the internet, but the idea here is just to restrict access to anything on the server, so that it can only be used for remote forwarding.
Now we want to automate the action of forwarding so we need to authenticate ourselves with an ssh key instead of writing password
Now we create a ssh key pair for the user on ubu2 and put them in etc/server/ssh_keys/
ssh-keygen -t rsa -b 2048
and manually copy the key to the jumper accout on server29 and the linux account on 130.225.170.70. (we cannot copy it with ssh-copy-id because of the ForceCommand we made in the sshd_config's)
Then we will edit ssh client config file on ubu2 in /etc/ssh/ssh_config and add the following content to the end:
Host parent
User jumper
Hostname 10.100.0.1
IdentityFile /etc/server/ssh_keys/id_rsa
Host backend
User linux
Hostname 130.225.170.70
port 22022
ProxyJump parent
RemoteForward 8080 localhost:8080
RemoteForward 8081 localhost:8081
RemoteForward 8082 localhost:8082
RemoteForward 8083 localhost:8083
RemoteForward 8084 localhost:8084
IdentityFile /etc/server/ssh_keys/id_rsa
This means that we have an alias parent which is linking to server29 and and alias backend which is using parent as a proxy and forwarding a range of ports from ubu2 to the host on 130.225.170.70
to forwad the ports we simply need to write
ssh -N backend
-N means we will not ask to get a shell
So now we can add this to a service file so we can have persistent port forwarding.
Lets create a file /etc/systemd/system/server.service with the following content:
[Unit]
Description=Ssh forwarding service
After=network.target
[Service]
Type=simple
ExecStart=/usr/bin/ssh -N backend
RestartSec=5
Restart=always
[Install]
WantedBy=multi-user.target
then we can restart systemd
systemctl daemon-reload
and because we are using known hosts to avoid dns spoofing attacks, then we need to login as root one time and write yes to accept the host. Or else the systemd service will never start forwading ports.
as root
ssh backend
and then enable and start the service
systemctl enable server
systemctl start server
we can now check if it works by either checking journalctl for a big pinguin
journalctl -xeu server
it should look like this
May 13 10:06:40 ubu2 ssh[1444]: *******************************************
May 13 10:06:40 ubu2 ssh[1444]: WELCOME TO Group4 BACKEND SERVER
May 13 10:06:40 ubu2 ssh[1444]: *******************************************
May 13 10:06:40 ubu2 ssh[1444]: _nnnn_
May 13 10:06:40 ubu2 ssh[1444]: dGGGGMMb ,"""""""""""""".
May 13 10:06:40 ubu2 ssh[1444]: @p~qp~~qMb | Linux Rules! |
May 13 10:06:40 ubu2 ssh[1444]: M|@||@) M| _;..............'
May 13 10:06:40 ubu2 ssh[1444]: @,----.JM| -'
May 13 10:06:40 ubu2 ssh[1444]: JS^\__/ qKL
May 13 10:06:40 ubu2 ssh[1444]: dZP qKRb
May 13 10:06:40 ubu2 ssh[1444]: dZP qKKb
May 13 10:06:40 ubu2 ssh[1444]: fZP SMMb
May 13 10:06:40 ubu2 ssh[1444]: HZM MMMM
May 13 10:06:40 ubu2 ssh[1444]: FqM MMMM
May 13 10:06:40 ubu2 ssh[1444]: __| ". |\dS"qML
May 13 10:06:40 ubu2 ssh[1444]: | `. | `' \Zq
May 13 10:06:40 ubu2 ssh[1444]: _) \.___.,| .'
May 13 10:06:40 ubu2 ssh[1444]: \____ )MMMMMM| .'
May 13 10:06:40 ubu2 ssh[1444]: `-' `--' h__
or we can also login to the backend server and check if the ports are open with
ssh -tlpn
At this point we can set up a reverse proxy like NGINX on the remote server to share our services from the docker containers we will setup on the ubu2 before the examn on the 28th.
this will be:
- Dendrite beta (a matrix server written in go)
- A front end for the matrix server
- SQLlite
Ok, now lets make som security:
We have looked at many different options of security and we have even created our own small tools in Go, to help us with managing logging and sending mails.
We were looking at alot and we ended up using these:
- MITMProxy (a simple http and https proxy with a web interface and python API)
- Tripwire (a Host intrusion detection system)
- Our own logwatcher which is listening to journalctl, and sending mails in a user defined interval.
- Firewall with
iptables
We started by using squid but we got some problems when we needed to work with https, so we started looking at MITMProxy since we could get a direct binary and execute it on the old Suse 13, then we could set up simple scripts in python so modify the content of a request. a collection of scripts for MITMProxy
#~/script.py
from mitmproxy import http
def request(flow: http.HTTPFlow) -> None:
if flow.request.url.contains("google"):
flow.response = http.HTTPResponse.make(
200,
b"Hello from group 10.1\n",
{"Content-Type": "text/html"}
)then we can just start our proxy like this
mitmweb -s ~/script.py
MITMProxy is also providing their own docker image so we wanted to make our arch machine a docker station for proxying.
First we update the machine, install docker, start and enable it.
sudo pacman -Syu
sudo pacman -S docker
sudo systemctl start docker
sudo systemctl enable dockerThen we created a docker file so that we can get our script inside the container as well
FROM mitmproxy/mitmproxy:latest
COPY ./conf.py .
EXPOSE 8081
EXPOSE 8080
CMD ["mitmweb", "--mode", "transparent", "--web-host", "0.0.0.0", "-s", "conf.py", "--showhost"]
Then we build the image
sudo docker build -t mitm .Finally we start the server.
docker run --rm -it -v ~/.mitmproxy:/home/mitmproxy/.mitmproxy -p 8080:8080 -p 8081:8081 mitmWe share the location .mitmproxy since our autogenerated ca-certificates are saved here, and then we can reuse them next time we start the server. (we could also create our own very easily but now we just use the autogenerated ones)
On the client machines we can then take the pem certificate and copy them into /usr/local/share/ca-certificates/mitmproxy.crt and then we need to run
update-ca-certificatesThen we can use the proxy but it has to be transparent since the proxy is set to transparent mode and we will get into how we do that when we talk about the Firewall.
On the router machine (server29) we will set up a firewall script that we will launch on every boot.
it looks like this
...
# SET POLICIES
iptables -P INPUT DROP
iptables -P OUTPUT DROP
iptables -P FORWARD DROP
# ALLOW ALL TRAFIC ON LOCALHOST INTERFACE
iptables -A INPUT -i lo -j ACCEPT
# ALLOW ALL TRAFIC ON DMZ INTERFACE
iptables -A INPUT -i eth1 -j ACCEPT
iptables -A OUTPUT -o eth1 -j ACCEPT
iptables -A FORWARD -i eth1 -j ACCEPT
# SET INPUT AND OUTUT STATE ORIENTED ACCEPT FOR INPUT OUTPUT AND FORWARD
iptables -A INPUT -m conntrack --ctstate ESTABLISHED,RELATED -j ACCEPT
iptables -A OUTPUT -m conntrack --ctstate ESTABLISHED,RELATED -j ACCEPT
iptables -A FORWARD -i eth0 -o eth2 -m conntrack --ctstate ESTABLISHED,RELATED -j ACCEPT
iptables -A FORWARD -i eth2 -o eth0 -m conntrack --ctstate ESTABLISHED,RELATED -j ACCEPT
# ALLOW HTTP AND HTTPS
iptables -A OUTPUT -p tcp --dport 80 -j ACCEPT
iptables -A OUTPUT -p tcp --dport 443 -j ACCEPT
#Allow Ping
iptables -A INPUT -p icmp -m icmp --icmp-type 8 -j ACCEPT
# ALLOW DNS
iptables -A OUTPUT -p udp --sport 53 -j ACCEPT
iptables -A OUTPUT -p udp --dport 53 -j ACCEPT
# ALLOW SSH
iptables -A INPUT -p tcp --dport 22 -j ACCEPT
iptables -A OUTPUT -p tcp --dport 22 -j ACCEPT
# https://www.digitalocean.com/community/tutorials/how-to-forward-ports-through-a-linux-gateway-with-iptables
iptables -A FORWARD -i eth0 -o eth2 -p tcp --dport 22 -j ACCEPT
iptables -A FORWARD -i eth2 -o eth0 -p tcp --dport 80 -j ACCEPT
iptables -A FORWARD -i eth2 -o eth0 -p tcp --dport 443 -j ACCEPT
iptables -A FORWARD -i eth2 -o eth0 -p udp --dport 53 -j ACCEPT
# PROXY HTTP AND HTTPS THROUGH SQUID ()
# https://www.youtube.com/watch?v=mCmn3bb26xc
iptables -A FORWARD -i eth2 -o eth0 -p tcp --dport 8080 -j ACCEPT
iptables -t nat -A PREROUTING -i eth2 -p tcp --dport 80 -j DNAT --to 192.168.150.30:8080
iptables -t nat -A PREROUTING -i eth0 -p tcp --dport 80 -j REDIRECT --to-port 8080
iptables -t nat -A PREROUTING -i eth2 -p tcp --dport 443 -j DNAT --to 192.168.150.30:8080
iptables -t nat -A PREROUTING -i eth0 -p tcp --dport 443 -j REDIRECT --to-port 8080
iptables -t nat -A POSTROUTING -j MASQUERADE
...
We had some problems with the transparent proxy when it was on an external machine or in a docker image, and we didnt have time to figure out how to make it work. Transparent proxying work working fine when we were just using the router as the proxy as well.
Becayse we are using suse Leap we can just use zypper to install.
zypper install tripwirethen create the local and the site keyfiles.
twadmin --generate-keys --local-keyfile /etc/tripwire/server29-local.key
twadmin --generate-keys --site-keyfile /etc/tripwire/site.keythen create the config file
twadmin --create-cfgfile -S /etc/tripwire/site.key /etc/tripwire/twcfg.txt
create a file in /etc/tripwire/ called twpol.txt and add this content:
(
rulename = "Tripwire Data Files",
severity = 100
)
{
/var/lib/tripwire -> $(Dynamic) -i ;
/var/lib/tripwire/report -> $(Dynamic) (recurse=0) ;
}
(
rulename = "Root & Home",
severity = 100
)
{
/ -> $(IgnoreAll) (recurse=1) ;
/home -> $(IgnoreAll) (recurse=1) ;
}
(
rulename = "System Directories",
severity = 100
)
{
/bin -> $(IgnoreNone)-SHa ;
/boot -> $(IgnoreNone)-SHa ;
/etc -> $(IgnoreNone)-SHa ;
/lib -> $(IgnoreNone)-SHa ;
/opt -> $(IgnoreNone)-SHa ;
/root -> $(IgnoreNone)-SHa ;
/sbin -> $(IgnoreNone)-SHa ;
/usr -> $(IgnoreNone)-SHa ;
}
then copy config and rules:
cp twcfg.txt tw.cfg
cp twpol.txt tw.poland create the polfile
twadmin --create-polfile -S site.key /etc/tripwire/twpol.txtnow we can initiate the database
tripwire --initand check if it all works
tripwire --checkWe wrote a simple program in Go to read the journalctl file as a json and parse it, the source code is included on github.