Proxmox with BGP+EVPN+VXLAN

Proxmox by default does not support BGP+EVPN+VXLAN but there is a small piece of documentation on the Wiki of Proxmox.

Using the routing daemon Frrouting a Proxmox cluster can also be configured to use BGP with EVPN+VXLAN for it’s routing allowing for very flexible networks.

I won’t go into all the details of Frr, BGP, EVPN and VXLAN as the internet already more then enough resources about this. I’ll get right to it.

Frrouting

After installing Proxmox (7.1) on the node I temporarily connected the node via a ad-hoc connection to the internet to be able to install Frr.

curl -s https://deb.frrouting.org/frr/keys.asc | sudo apt-key add -
echo deb https://deb.frrouting.org/frr bullseye frr-7 | sudo tee -a /etc/apt/sources.list.d/frr.list
apt install frr frr-pythontools

After installing Frr I configured the /etc/frr/frr.conf config file:

frr version 7.5.1
frr defaults traditional
hostname infra-72-45-36
log syslog informational
no ip forwarding
no ipv6 forwarding
service integrated-vtysh-config
!
interface enp101s0f0np0
  no ipv6 nd suppress-ra
!
interface enp101s0f1np1
  no ipv6 nd suppress-ra
!
interface lo
  ip address 10.255.254.5/32
  ipv6 address 2a05:1500:xxx:xx::5/128
!
router bgp 4200400036
  bgp router-id 10.255.254.5
  no bgp ebgp-requires-policy
  no bgp default ipv4-unicast
  no bgp network import-check
  neighbor core peer-group
  neighbor core remote-as external
  neighbor core ebgp-multihop 255
  neighbor enp101s0f0np0 interface peer-group core
  neighbor enp101s0f1np1 interface peer-group core
  !
  address-family ipv4 unicast
    redistribute connected
    neighbor core activate
    exit-address-family
  !
  address-family ipv6 unicast
    redistribute connected
    neighbor core activate
    exit-address-family
  !
  address-family l2vpn evpn
    neighbor core activate
    advertise-all-vni
    exit-address-family
  !
line vty
!

In this case the host will be connecting to two Cumulus Linux routers using BGP Unnumbered.

The interfaces enp101s0f0np0 and enp101s0f1np1 are the uplinks of this node the the two routers.

/etc/network/interfaces

Now we need to make sure the /etc/network/interface file is populated with the proper information.

auto lo
iface lo inet loopback

auto enp101s0f1np1
iface enp101s0f1np1 inet manual
    mtu 9216

auto enp101s0f0np0
iface enp101s0f0np0 inet manual
    mtu 9216

This makes sure the interfaces (Mellanox ConnectX-5 2x25Gb SFP28) interfaces are online and running with an MTUof 9216.

The MTU of 9216 is needed so I can transport traffic with an MTU of 9000 within my VXLAN packets. VXLAN has an overhead of 50 bytes. So to transport an Ethernet packet of 1500 bytes you need to make sure you have at least an MTU of 1550 on your VXLAN underlay network.

VXLAN bridges

I now created a bunch of devices in the interfaces file:

auto vxlan201
iface vxlan201 inet static
    mtu 1500
    pre-up ip link add vxlan201 type vxlan id 201 dstport 4789 local 10.255.254.5 nolearning
    up ip link set vxlan201 up
    down ip link set vxlan201 down
    post-down ip link del vxlan201

auto vmbr201
iface vmbr201 inet manual
    bridge_ports vxlan201
    bridge-stp off
    bridge-fd 0

auto vxlan202
iface vxlan202 inet static
    mtu 1500
    pre-up ip link add vxlan202 type vxlan id 202 dstport 4789 local 10.255.254.5 nolearning
    up ip link set vxlan202 up
    down ip link set vxlan202 down
    post-down ip link del vxlan202

auto vmbr202
iface vmbr202 inet manual
    address 192.168.202.36/24
    bridge_ports vxlan202
    bridge-stp off
    bridge-fd 0

I created vmbr201 and vmbr202 which correspond to VNI 201 and 202 on the network. These can now be used with Proxmox to connect VMs to.

The IP-Address (10.255.254.5) set at the local argument of the ip command is the IP-address connected to the loopback interface and advertised using BGP.

This will be the address used for the VTEP in EVPN/VXLAN communication.

In reality however I have much more bridges

  • vmbr200
  • vmbr201
  • vmbr202
  • vmbr601
  • vmbr602
  • vmbr603
  • vmbr604
Overview of interfaces in Proxmox

Proxmox cluster network

To be able to create a cluster with Proxmox you need a Layer2 network between the hosts where corosync can be used for cluster communication.

In this case I’m using vmbr202 which has IP-address 192.168.202.36/24 on this node. Other nodes in the cluster have a IPv4 address in the same network and allows them to communicate with the others.

Using fail2ban to block unauthorized calls to CloudStack API

Apache CloudStack does not have a build-in mechanism to rate-limit failed authentication attemps on the API. This potentially allows an attacker to brute-force credentials and gain access.

The api.allowed.source.cidr.list configuration option in CloudStack can be used to globally or on an account level limit the source IPs where the API allows requests from. This is always good to do (if possible), but it does not cover every use-case.

Sometimes you just want to keep malicious traffic outside the door and fail2ban can help there.

Nginx proxy in front of CloudStack

A common use-case is that the Management server of Apache CloudStack is not directly connected to the network, but placed behind a reverse proxy like Nginx or something similar.

This proxy can then also handle SSL termination.

In this example we’re using Nginx as a proxy.

fail2ban

Using fail2ban we can scan the access logs of Nginx and block IP addresses who are abusing our API. In this case we filter on two HTTP status codes:

  • 401
  • 531

This results in that we create the following files:

  • /etc/fail2ban/jail.d/nginx-401.conf
  • /etc/fail2ban/jail.d/nginx-531.conf
  • /etc/fail2ban/filter.d/nginx-401.conf
  • /etc/fail2ban/filter.d/nginx-531.conf

jail.d/nginx-401.conf

[nginx-401]
enabled = true
port = http,https
filter = nginx-401
action = iptables-allports
logpath = %(nginx_access_log)s
bantime = 3600
findtime = 600
maxretry = 25
ignoreip = 127.0.0.1/8

filter.d/nginx-401.conf

[Definition]
failregex = ^ -."(GET|POST|HEAD).HTTP.*" 401
ignoreregex =

Change 401 to 531 where needed to also block HTTP codes 531.

iptables

The action taken by fail2bain is iptables-allports which causes iptables to block all traffic from the particular source IP when it is being banned.

Combined Charging System (CCS) and IPv6

When we talk about IP protocols and IPv6 in this particular case we think about routing over the internet. But what many do not know is that IPv6 is already playing a major role in all kinds of systems.

In this post I’m not going to explain IPv6 in detail, for many things I’ll link to external pages.

Ad-Hoc networking using Link-Local

IPv6 has a great feature where each host on a network will generate a Link Local Address (LL) which is mandatory according to the protocol.

The great thing about LL is that it allows hosts to establish communication without the need of DHCP or anything else. Just plug in a host on a Layer 2 Ethernet network and they can start communicating right away.

fe80::5054:ff:fe98:aaee is an example of a IPv6 LL address and using this address it can search for (via multicast) and communicate with other hosts in the network.

Examples of systems using IPv6 LL are Matter and Combined Charging System (CCS).

Combined Charging System (CCS)

CCS is a standard for (fast charging) electric vehicles. Some years ago I stumbled upon multiple documents showing that CCS would be using a combination of PLC, IPv6, TCP/IP, UDP, HTTP, TLS and many other very common protocols for communication.

Using PowerLine Communication (PLC) the vehicle and charger will establish a Layer 2 Ethernet network over which they will communicate using IPv6 LL.

I was not able to find the exact details of the IPv6 part of the CCS standard, but it is very interesting to see that IPv6 is being used for something like charging electric vehicles.

Kudos to the people behind CCS and for using IPv6 for this!

In my opinion it’s a no-brainer to use IPv6 for such functionality as the LL addressing allow you to create networks very quickly and in a reliable manner.

You can also use a very well documented protocol and the many tools for IPv6.

Now I would like to run Wireshark on such a link!

If you have more information about this, please contact me!