Linux Multiple Network Interfaces (NICs) – One Interface with Static Public IP and One Interface with Private DHCP LAN IP Address – Routes and Routing

Friday, July 24th, 2020

Linux KVM:  Using Multiple NICs and Routing Traffic Properly Between Them

When setting up a KVM guest to use multiple network interface controllers (NICs), additional ip routes may be needed in order for the additional interfaces to work properly.  For example, if you configure a NIC with a public static IP address and a NIC with an internal private DHCP LAN IP address, you must create a route for any traffic that comes through the DHCP LAN IP address to respond via the interface from which the request originated.  Otherwise, forwarded NAT traffic from the main KVM host to the DHCP internal LAN IP will reach its destination, but no response will be sent back (because it will attempt to send the response via the configured static IP address interface which may NOT be the original destination of the senders request).

The Solution:

https://unix.stackexchange.com/questions/4420/reply-on-same-interface-as-incoming/23345#answer-23345

From the above link, the solution for me was to do the following in the KVM guest virtual machine:

Only needs to be done once:

sudo -i
echo 200 isp1 >> /etc/iproute2/rt_tables

Setting up the route (adjust variables as necessary):

sudo -i
ip rule add from <interface_IP> table isp1 priority 900
ip rule add from <interface_IP> dev <interface> table isp1
ip route add default via <gateway_IP> dev <interface> table isp

The command I used for my specific setup:

sudo -i
ip rule add from 192.168.122.10 table isp1 priority 900 
ip rule add from 192.168.122.10 dev ens9 table isp1 
ip route add default via 192.168.122.1 dev ens9 table isp1

Making it permanent (apply on system start up):

sudo -i
nano /etc/network/interfaces

I added the below post-up rules (adjust variables as necessary):

auto ens9
iface ens9 inet dhcp
        post-up ip rule add from <interface_IP> table isp1 priority 900
        post-up ip rule add from <interface_IP> dev <interface> table isp1
        post-up ip route add default via <gateway_IP> dev <interface> table isp1

The route is created whenever the dhcp interface is brought up.

CentOS – Using NAT with KVM Guests

Wednesday, July 17th, 2019

CentOS – Using NAT with KVM Guests

Please note that all commands in this guide must be run on the main HOST machine (the physical machine).  They should not be run on KVM guests (virtual machines).

If your server has a limited number of IPv4 addresses, it might be best to setup and run some virtual machines that are configured to use the default NAT network interface that KVM provides.  This will allow you to run multiple virtual machines that share the same IP address.  Think of it as setting up a home network with multiple devices with certain ports forwarded to specific devices for incoming connections.

First, create the virtual machines that are going to use NAT using virt-manager as you normally would.  In the virtual machine configuration wizard, assign the default NAT network interface named "virbr0".  After the virtual machines have been created, shut down the virtual machines.  Now, we'll assign these virtual machines static LAN IP addresses so that we can port forward certain ports and always have them reach the proper virtual machines. 

The first thing we need to do is get the MAC address of each virtual machine.  Write down the name of the virtual machine and its MAC address, as we'll need this information later on when we edit the NAT interface and assign static LAN IP addresses to our virtual machines.  Run this command to retrieve the MAC address for a specific virtual machine.

virsh dumpxml VM_NAME | grep -i '<mac'

Using the MAC address information from the VMs we want to use NAT with, edit the default NAT interface by running the below command. 

virsh net-edit default

If for some reason the NAT interface is not named default, you can find it by running the below command:

virsh net-list

After the <range /> entry, assign the static LAN IP addresses similar to the following:

<dhcp>
      <range start='192.168.122.2' end='192.168.122.254'/>
      <host mac='52:54:00:ff:4a:2a' name='vm1' ip='192.168.122.13'/>
      <host mac='52:54:00:bb:35:67' name='vm2' ip='192.168.122.14'/>
      <host mac='52:54:00:aa:d9:f2' name='vm3' ip='192.168.122.15'/>
</dhcp>

Save the file with your desired values and quit the editor.  Restart the NAT interface by running the below commands:

virsh net-destroy default
virsh net-start default

Now, you'll need to setup your iptables port forwarding rules.  Adjust the below rules as necessary (changing the port numbers to the ones you want to use) and then save them so that they persist:

iptables -I FORWARD -o virbr0 -d 192.168.122.13 -j ACCEPT
iptables -t nat -I PREROUTING -p tcp --dport 39989 -j DNAT --to 192.168.122.13:39989
iptables -I FORWARD -o virbr0 -d 192.168.122.14 -j ACCEPT
iptables -t nat -I PREROUTING -p tcp --dport 39990 -j DNAT --to 192.168.122.14:39990
iptables -I FORWARD -o virbr0 -d 192.168.122.15 -j ACCEPT
iptables -t nat -I PREROUTING -p tcp --dport 39991 -j DNAT --to 192.168.122.15:39991
iptables -t nat -A POSTROUTING -s 192.168.122.0/24 -j MASQUERADE
iptables -A FORWARD -o virbr0 -m state --state RELATED,ESTABLISHED -j ACCEPT
iptables -A FORWARD -i virbr0 -o br0 -j ACCEPT
iptables -A FORWARD -i virbr0 -o lo -j ACCEPT
iptables-save
service iptables save

Congrats, your virtual machines are now using NAT, have been assigned static LAN IP addresses, and iptables rules on the host server have been configured to port forward specific ports to each NAT VM.


Persistently Saving NAT Port Forward Rules

The only solution I found that would persistently save my NAT forwarding rules is to create a libvirt hook bash script as mentioned here

service iptables stop
iptables -F
service iptables save
service iptables start
mkdir -p  /etc/libvirt/hooks
nano  /etc/libvirt/hooks/qemu

The contents of the "/etc/libvirt/hooks/qemu" file should look similar to the following:

#!/bin/bash
# IMPORTANT: Change the "VM NAME" string to match your actual VM Name.
# In order to create rules to other VMs, just duplicate the below block and configure
# it accordingly.
if [ "${1}" = "vm1" ]; then   # Update the following variables to fit your setup
   GUEST_IP=192.168.122.13
   GUEST_PORT=39989
   HOST_PORT=39989   if [ "${2}" = "stopped" ] || [ "${2}" = "reconnect" ]; then
    /sbin/iptables -D FORWARD -o virbr0 -d  $GUEST_IP -j ACCEPT
    /sbin/iptables -t nat -D PREROUTING -p tcp --dport $HOST_PORT -j DNAT --to $GUEST_IP:$GUEST_PORT
   fi
   if [ "${2}" = "start" ] || [ "${2}" = "reconnect" ]; then
    /sbin/iptables -I FORWARD -o virbr0 -d  $GUEST_IP -j ACCEPT
    /sbin/iptables -t nat -I PREROUTING -p tcp --dport $HOST_PORT -j DNAT --to $GUEST_IP:$GUEST_PORT
   fi
fi
if [ "${1}" = "vm2" ]; then   # Update the following variables to fit your setup
   GUEST_IP=192.168.122.14
   GUEST_PORT=39990
   HOST_PORT=39990   
   if [ "${2}" = "stopped" ] || [ "${2}" = "reconnect" ]; then
        /sbin/iptables -D FORWARD -o virbr0 -d  $GUEST_IP -j ACCEPT
        /sbin/iptables -t nat -D PREROUTING -p tcp --dport $HOST_PORT -j DNAT --to $GUEST_IP:$GUEST_PORT
   fi
   if [ "${2}" = "start" ] || [ "${2}" = "reconnect" ]; then
        /sbin/iptables -I FORWARD -o virbr0 -d  $GUEST_IP -j ACCEPT
        /sbin/iptables -t nat -I PREROUTING -p tcp --dport $HOST_PORT -j DNAT --to $GUEST_IP:$GUEST_PORT
   fi
fi
if [ "${1}" = "vm3" ]; then   # Update the following variables to fit your setup
   GUEST_IP=192.168.122.15
   GUEST_PORT=39991
   HOST_PORT=39991
   if [ "${2}" = "stopped" ] || [ "${2}" = "reconnect" ]; then
        /sbin/iptables -D FORWARD -o virbr0 -d  $GUEST_IP -j ACCEPT
        /sbin/iptables -t nat -D PREROUTING -p tcp --dport $HOST_PORT -j DNAT --to $GUEST_IP:$GUEST_PORT
   fi
   if [ "${2}" = "start" ] || [ "${2}" = "reconnect" ]; then
        /sbin/iptables -I FORWARD -o virbr0 -d  $GUEST_IP -j ACCEPT
        /sbin/iptables -t nat -I PREROUTING -p tcp --dport $HOST_PORT -j DNAT --to $GUEST_IP:$GUEST_PORT
   fi
fi

Save and exit.  Make the script executable.

chmod +x /etc/libvirt/hooks/qemu

Reboot the host server.


Old Instructions for Persistent Saving (Non-Working)

If your iptables forwarding rules are not persisted after the host machine is rebooted or shutdown, run the following commands:

sudo -i
yum install -y iptables-services
systemctl stop firewalld
systemctl disable firewalld
systemctl enable iptables
nano /etc/sysconfig/iptables-config 

Change the below values to "yes":

IPTABLES_SAVE_ON_RESTART="yes"
IPTABLES_SAVE_ON_STOP="yes"

Save and exit.  Reboot the server.

If you're still having issues, try this (will clear your existing iptables rules):

iptables-save > iptables_bk
service iptables stop
iptables -F
<run iptables NAT rules here>
<run any other iptables rules you want>
service iptables save
service iptables start

More Detailed Guide