% VXLAN lab based on OpenVSwitch and LXD containers
The very first idea when I started writing this lab was to illustrate the Virtual Extensible LAN (VXLAN) technology. Now that OpenVSwitch configuration is smoothely integrated in the Debian networking configuration files, this should have resulted in somewhat easy-to-read gist.
It was while advancing in the writing of the gist that things started to go wrong.
Although it's nice to have a VLAN (broadcast domain) sitting on top of an interconnection of different IP networks, what if the containers within this VLAN need failover between the two tenants where they stand ? That's when trouble comes. We need automatic IPv4 and IPv6 addressing for the containers with resiliency !
This is the way a short written gist that was supposed to stay short became a way too long document.
- Traffic flows are marked coming back and forth of each tenant and are sent to a dedicated routing table
- The gateway IPv4 and IPv6 addresses in the shared VLAN are handled by keepalived
- IPv4 dynamic addressing is handled by the ISC DHCP server with its failover feature
- IPv6 SLAAC (StateLess Automatic Address Configuration) is provided by radvd
Sorry for the inconvenience, but the result is worth it even if the reading is long and painful.
We use three virtual machines :
- redRouter is there to route the traffic between the two tenants. It also has to route the traffic coming from the containers to the Internet.
- blueTenant is the left side tenant which hosts some LXD containers on the VXLAN shared VLAN
- greenTenant is the right side tenant which hosts some other LXD containers on the same VXLAN shared VLAN
IMPORTANT ! All commands are run as normal user as this user account must belong to the mandatory system groups.
- kvm group for virtualisation
- lxd group for containerisation
Otherwise, atlmost all the command examples below should be preceded by sudo.
In order to run the three virtual machines we have a few preparation steps.
ionice -c3 cp $HOME/masters/debian-testing-amd64.qcow2 blueTenant.qcow2
ionice -c3 cp $HOME/masters/debian-testing-amd64.qcow2 greenTenant.qcow2
ionice -c3 cp $HOME/masters/debian-testing-amd64.qcow2 redRouter.qcow2
wget https://raw.githubusercontent.com/platu/inetdoc/master/guides/vm/files/ovs-startup.sh
chmod +x ovs-startup.sh
This shell script starts a KVM hypervisor based virtual machine. It runs with three parameters :
- virtual machine image file
- initial amount of guest RAM memory
- network tap interface number
Notice that the MAC address of the virtual machine is built from the ba:ad:ca:fe prefix followed by the tap interface number converted to hexadecimal on the two bytes on the right. If the MAC address prefix is already used in your infrastructure, the script has to be edited to change it.
- openvswitch-switch
- qemu-system-x86
aptitude versions openvswitch-switch
i 3.1.0~git20230108.006e1c6-1+b1 testing 500
aptitude versions qemu-system-x86
i 1:7.2+dfsg-2 testing 500
If your host system has a GUI, you may want to access to virtual machines screen through the SPICE protocol. Therefore, you have to install the spice-client-gtk package and then use the spicy tool.
Setup the three tap interfaces in order to plug the virtual machines on Host distribution switch dsw-host
As we use the Debian GNU/Linux network interfaces configuration is under ifupdown control.
Here is an excerpt from Host /etc/network/interfaces file:
# Host main distribution switch
# The Host physical port is named eno1. Change it to fit your context.
auto dsw-host
iface dsw-host inet manual
ovs_type OVSBridge
ovs_ports eno1 vlan28 tap20 tap220 tap221
# Host physical port
# Configuration mode may be changed from manual to static or dhcp depending on context
allow-dsw-host eno1
iface eno1 inet manual
ovs_bridge dsw-host
ovs_type OVSPort
up ip link set dev $IFACE up
down ip link set dev $IFACE down
# Host SVI interface
# This network layer Switch Virtual Interface is in access mode and belongs to
# VLAN 28.
allow-dsw-host vlan28
iface vlan28 inet static
ovs_bridge dsw-host
ovs_type OVSIntPort
ovs_options tag=28 vlan_mode=access
address 198.18.28.20/24
iface vlan28 inet6 static
ovs_bridge dsw-host
ovs_type OVSIntPort
ovs_options tag=28 vlan_mode=access
address 2001:678:3fc:1c::14/64
# redRouter -> trunk mode
allow-dsw-host tap20
iface tap20 inet manual
ovs_bridge dsw-host
ovs_type OVSPort
ovs_options vlan_mode=trunk
pre-up ip tuntap add mode tap dev $IFACE group kvm multi_queue
up ip link set dev $IFACE up
down ip link set dev $IFACE down
post-down ip tuntap del mode tap dev $IFACE multi_queue
# blueTenant left side tenant -> access mode on VLAN 220
allow-dsw-host tap220
iface tap220 inet manual
ovs_bridge dsw-host
ovs_type OVSPort
ovs_options tag=220 vlan_mode=access
pre-up ip tuntap add mode tap dev $IFACE group kvm multi_queue
up ip link set dev $IFACE up
down ip link set dev $IFACE down
post-down ip tuntap del mode tap dev $IFACE multi_queue
# greenTenant right side tenant -> access mode on VLAN 221
allow-dsw-host tap221
iface tap221 inet manual
ovs_bridge dsw-host
ovs_type OVSPort
ovs_options tag=221 vlan_mode=access
pre-up ip tuntap add mode tap dev $IFACE group kvm multi_queue
up ip link set dev $IFACE up
down ip link set dev $IFACE down
post-down ip tuntap del mode tap dev $IFACE multi_queue
Don't forget to run ifdown and/or ifup on the above interfaces before to go further !
For instance:
sudo ifdown dsw-host && sudo ifup dsw-host
Here is a way to check interfaces status :
ip a ls | egrep 'state (UP|UNKNOWN)'
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
2: eno1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc mq master ovs-system state UP group default qlen 1000
7: dsw-host: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc noqueue state UNKNOWN group default qlen 1000
10: vlan28: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc noqueue state UNKNOWN group default qlen 1000
36: tap20: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc mq master ovs-system state UP group default qlen 1000
37: tap220: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc mq master ovs-system state UP group default qlen 1000
38: tap221: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc mq master ovs-system state UP group default qlen 1000
Edit /etc/sysctl.conf in order to get the following result
egrep -v '(^#|^$)' /etc/sysctl.conf
net.ipv4.conf.default.rp_filter=2
net.ipv4.conf.all.rp_filter=2
net.ipv4.ip_forward=1
net.ipv6.conf.all.forwarding=1
net.ipv4.conf.all.log_martians = 1
Don't forget to turn routing on after editing the /etc/sysctl.conf file
sudo sysctl --systemRemember that the normal user account has to be part of the kvm system group.
id | grep -o kvm
kvmNow we are ready to launch the virtual machines.
#!/bin/bash
$HOME/masters/scripts/ovs-startup.sh redRouter.qcow2 1024 20
$HOME/masters/scripts/ovs-startup.sh blueTenant.qcow2 1024 220
$HOME/masters/scripts/ovs-startup.sh greenTenant.qcow2 1024 221
exit 0Take note of the MAC addresses below. They will be useful in the following steps of the lab.
bash startup.sh
~> Virtual machine filename : redRouter.qcow2
~> RAM size : 1024MB
~> SPICE VDI port number : 5920
~> telnet console port number : 2320
~> MAC address : b8:ad:ca:fe:00:14
~> Switch port interface : tap20, access mode
~> IPv6 LL address : fe80::baad:caff:fefe:14%vlan28
~> Virtual machine filename : blueTenant.qcow2
~> RAM size : 1024MB
~> SPICE VDI port number : 6120
~> telnet console port number : 2520
~> MAC address : b8:ad:ca:fe:00:dc
~> Switch port interface : tap220, trunk mode
~> IPv6 LL address : fe80::baad:caff:fefe:dc%dsw-host
~> Virtual machine filename : greenTenant.qcow2
~> RAM size : 1024MB
~> SPICE VDI port number : 6121
~> telnet console port number : 2521
~> MAC address : b8:ad:ca:fe:00:dd
~> Switch port interface : tap221, trunk mode
~> IPv6 LL address : fe80::baad:caff:fefe:dd%dsw-host
\newpage
As it is the main purpose of this virtual machine, it would be stupid to forget it !
Edit /etc/sysctl.conf in order to get the following result
egrep -v '(^#|^$)' /etc/sysctl.conf
net.ipv4.conf.default.rp_filter=2
net.ipv4.conf.all.rp_filter=2
net.ipv4.ip_forward=1
net.ipv6.conf.all.forwarding=1
net.ipv4.conf.all.log_martians = 1
Turn routing on after editing the /etc/sysctl.conf file
sudo sysctl --system
Traffic coming back and forth from the containers to the Internet may follow two different paths. One through the link between the Blue tenant and another one through the Green tenant. With IPv4, this could lead to an asymetric routing which is not compatible with connection tracking introduced by firewall rules.
This is why we choose to setup two separate routing tables : one for each path.
We have to add two new lines at the end of the /etc/iproute2/rt_tables file:
#
# reserved values
#
255 local
254 main
253 default
0 unspec
#
# local
#
#1 inr.ruhep
220 blue
221 green
- The new routing tables entries will be set in the network interfaces configuration file
- The
ruleto send traffic to a specific routing table is also set in the network interfaces configuration file - The traffic will be marked by rules in the
mangletable of netfilter/iptables configuration file
The router's main interface named enp0s1 is a trunk port connected to
tap20 on Host system. We have to set one subinterface per VLAN.
We can check the tap20 switch port configuration from the Host system.
The main parameter Here is vlan_mode at the end of the screen copy below.
sudo ovs-vsctl list port tap20
_uuid : d1cbfd72-d44f-429f-aee1-fe6e54c5430b
bond_active_slave : []
bond_downdelay : 0
bond_fake_iface : false
bond_mode : []
bond_updelay : 0
cvlans : []
external_ids : {}
fake_bridge : false
interfaces : [44e1f871-4dae-4110-9d69-4a83e0e5ac8d]
lacp : []
mac : []
name : tap20
other_config : {}
protected : false
qos : []
rstp_statistics : {}
rstp_status : {}
statistics : {}
status : {}
tag : 28
trunks : []
vlan_mode : trunk
Here is an excerpt of redRouter networking configuration file :
/etc/network/interfaces
# The main network interface
# Interface name enp0s1 may have to be changed !
auto enp0s1
iface enp0s1 inet manual
up ip link set dev $IFACE up
down ip link set dev $IFACE down
# The Host system link
auto enp0s1.28
iface enp0s1.28 inet static
address 198.18.28.200/24
gateway 198.18.28.1
## Use quad9.net as DNS resolver
dns-nameservers 9.9.9.9
iface enp0s1.28 inet6 static
address 2001:678:3fc:1c::c8/64
gateway fe80:1c::1
pre-up echo 0 > /proc/sys/net/ipv6/conf/$IFACE/accept_dad
# The Blue tenant link
auto enp0s1.220
iface enp0s1.220 inet static
address 172.16.220.1/24
## IPv4 Blue routing table
post-up ip route add 172.16.220.0/24 dev $IFACE table blue
post-up ip route add 192.0.2.0/24 via 172.16.220.2 dev $IFACE table blue
post-up ip route add 198.18.28.0/24 dev enp0s1.28 table blue
post-up ip route add default via 198.18.28.1 dev enp0s1.28 table blue
## Traffic marked with 220 goes to the IPv4 Blue routing table
post-up ip rule add fwmark 220 table blue
iface enp0s1.220 inet6 static
address 2001:db8:dc::1/64
up ip -6 addr add fe80:dc::1/64 dev $IFACE
## IPv6 Blue routing table
post-up ip -6 route add 2001:db8:dc::/64 dev $IFACE table blue
post-up ip -6 route add 2001:db8:7::/64 via 2001:db8:dc::2 dev $IFACE table blue
post-up ip -6 route add 2001:678:3fc:1c::/64 dev enp0s1.28 table blue
post-up ip -6 route add default via fe80:1c::1 dev enp0s1.28 table blue
## Traffic marked with 220 goes to the IPv6 Blue routing table
post-up ip -6 rule add fwmark 220 table blue
pre-up echo 0 > /proc/sys/net/ipv6/conf/$IFACE/accept_dad
# The Green tenant link
auto enp0s1.221
iface enp0s1.221 inet static
address 172.16.221.1/24
## IPv4 Green routing table
post-up ip route add 172.16.221.0/24 dev $IFACE table green
post-up ip route add 192.0.2.0/24 via 172.16.221.2 dev $IFACE table green
post-up ip route add 198.18.28.0/24 dev enp0s1.28 table green
post-up ip route add default via 198.18.28.1 dev enp0s1.28 table green
## Traffic marked with 221 goes to the IPv4 Green routing table
post-up ip rule add fwmark 221 table green
iface enp0s1.221 inet6 static
address 2001:db8:dd::1/64
up ip -6 addr add fe80:dd::1/64 dev $IFACE
## IPv6 Green routing table
post-up ip -6 route add 2001:db8:dd::/64 dev $IFACE table green
post-up ip -6 route add 2001:db8:7::/64 via 2001:db8:dd::2 dev $IFACE table green
post-up ip -6 route add 2001:678:3fc:1c::/64 dev enp0s1.28 table green
post-up ip -6 route add default via fe80:1c::1 dev enp0s1.28 table green
## Traffic marked with 221 goes to the IPv6 Green routing table
post-up ip -6 rule add fwmark 221 table green
pre-up echo 0 > /proc/sys/net/ipv6/conf/$IFACE/accept_dad
In order to check the results of the network interfaces configurations use commands like these :
-
List rules for marked traffic :
ip rule ls 0: from all lookup local 32764: from all fwmark 0xdd lookup green <-- firewall mark 221 leads to green IPv4 and IPv6 routing tables 32765: from all fwmark 0xdc lookup blue <-- firewall mark 220 leads to blue IPv4 and IPv6 routing tables 32766: from all lookup main 32767: from all lookup default -
List IPv4 Green tenant routing table entries
ip route ls table green default via 198.18.28.1 dev enp0s1.28 172.16.221.0/24 dev enp0s1.221 scope link 192.0.2.0/24 via 172.16.221.2 dev enp0s1.221 198.18.28.0/23 dev enp0s1.28 scope link -
List IPv6 Blue tenant routing table entries
ip -6 route ls table blue 2001:678:3fc:1c::/64 dev enp0s1.28 metric 1024 pref medium 2001:db8:7::/64 via 2001:db8:dc::2 dev enp0s1.220 metric 1024 pref medium 2001:db8:dc::/64 dev enp0s1.220 metric 1024 pref medium default via fe80:1c::1 dev enp0s1.28 metric 1024 pref medium
The link between the redRouter and the Host system is the lab boarder to the real world. This is where connection tracking stands.
Check that the iptables-persistent package is installed on redRouter, so we
have to save the rules for IPv4 and IPv6 in the /etc/iptbales directory.
- First, the
nattable holds the source address translations for both IPv4 and IPv6. - Second, the
mangletable is used to set marks to the IPv4 and IPv6 traffic coming back and forth from the Blue or Green tenants
In the /etc/iptables/rules.v4 file :
## ~~~~~~~~~~~~~~~~ NAT table ~~~~~~
*nat
:PREROUTING ACCEPT [0:0]
:INPUT ACCEPT [0:0]
:POSTROUTING ACCEPT [0:0]
:OUTPUT ACCEPT [0:0]
-A POSTROUTING -o enp0s1.28 -p tcp --syn -j TCPMSS --clamp-mss-to-pmtu
-A POSTROUTING -o enp0s1.28 -j MASQUERADE
COMMIT
## ~~~~~~~~~~~~~~~~ MANGLE table ~~~~~~
*mangle
:PREROUTING ACCEPT [0:0]
-A PREROUTING -j CONNMARK --restore-mark
-A PREROUTING -i enp0s1.220 -s 192.0.2.0/24 -j MARK --set-mark 220
-A PREROUTING -i enp0s1.221 -s 192.0.2.0/24 -j MARK --set-mark 221
-A PREROUTING -j CONNMARK --save-mark
:INPUT ACCEPT [0:0]
:FORWARD ACCEPT [0:0]
:OUTPUT ACCEPT [0:0]
:POSTROUTING ACCEPT [0:0]
COMMIT
In the /etc/iptables/rules.v6 file :
## ~~~~~~~~~~~~~~~~ NAT table ~~~~~~
*nat
:PREROUTING ACCEPT [0:0]
:INPUT ACCEPT [0:0]
:POSTROUTING ACCEPT [0:0]
:OUTPUT ACCEPT [0:0]
-A POSTROUTING -o enp0s1.28 -j MASQUERADE
COMMIT
## ~~~~~~~~~~~~~~~~ MANGLE table ~~~~~~
*mangle
:PREROUTING ACCEPT [0:0]
-A PREROUTING -j CONNMARK --restore-mark
-A PREROUTING -i enp0s1.220 -s 2001:db8:7::/64 -j MARK --set-mark 220
-A PREROUTING -i enp0s1.221 -s 2001:db8:7::/64 -j MARK --set-mark 221
-A PREROUTING -j CONNMARK --save-mark
:INPUT ACCEPT [0:0]
:FORWARD ACCEPT [0:0]
:OUTPUT ACCEPT [0:0]
:POSTROUTING ACCEPT [0:0]
COMMIT
One way to apply the rules written in the two files is :
sudo sh -c "iptables-restore </etc/iptables/rules.v4"
sudo sh -c "ip6tables-restore </etc/iptables/rules.v6"
In order to check the results of these firewalling rules, use the following commands:
- List IPv4 NAT rules
sudo iptables -vnL -t nat
Chain PREROUTING (policy ACCEPT 27 packets, 5876 bytes)
pkts bytes target prot opt in out source destination
Chain INPUT (policy ACCEPT 15 packets, 2087 bytes)
pkts bytes target prot opt in out source destination
Chain POSTROUTING (policy ACCEPT 43 packets, 5589 bytes)
pkts bytes target prot opt in out source destination
0 0 TCPMSS tcp -- * enp0s1.28 0.0.0.0/0 0.0.0.0/0 tcp flags:0x17/0x02 TCPMSS clamp to PMTU
16 947 MASQUERADE all -- * enp0s1.28 0.0.0.0/0 0.0.0.0/0
Chain OUTPUT (policy ACCEPT 47 packets, 2747 bytes)
pkts bytes target prot opt in out source destination
- List IPv6 NAT rules
sudo ip6tables -vnL -t nat
Chain PREROUTING (policy ACCEPT 4 packets, 379 bytes)
pkts bytes target prot opt in out source destination
Chain INPUT (policy ACCEPT 1 packets, 138 bytes)
pkts bytes target prot opt in out source destination
Chain POSTROUTING (policy ACCEPT 11 packets, 968 bytes)
pkts bytes target prot opt in out source destination
10 1005 MASQUERADE all * enp0s1.28 ::/0 ::/0
Chain OUTPUT (policy ACCEPT 18 packets, 1732 bytes)
pkts bytes target prot opt in out source destination
- List IPv4 mangle rules
sudo iptables -vnL -t mangle
Chain PREROUTING (policy ACCEPT 240K packets, 31M bytes)
pkts bytes target prot opt in out source destination
240K 31M CONNMARK all -- * * 0.0.0.0/0 0.0.0.0/0 CONNMARK restore
18 1290 MARK all -- enp0s1.220 * 192.0.2.0/24 0.0.0.0/0 MARK set 0xdc
36 2580 MARK all -- enp0s1.221 * 192.0.2.0/24 0.0.0.0/0 MARK set 0xdd
240K 31M CONNMARK all -- * * 0.0.0.0/0 0.0.0.0/0 CONNMARK save
Chain INPUT (policy ACCEPT 25319 packets, 6506K bytes)
pkts bytes target prot opt in out source destination
Chain FORWARD (policy ACCEPT 214K packets, 24M bytes)
pkts bytes target prot opt in out source destination
Chain OUTPUT (policy ACCEPT 5495 packets, 1014K bytes)
pkts bytes target prot opt in out source destination
Chain POSTROUTING (policy ACCEPT 220K packets, 25M bytes)
pkts bytes target prot opt in out source destination
- List IPv6 mangle rules
sudo ip6tables -vnL -t mangle
Chain PREROUTING (policy ACCEPT 28476 packets, 45M bytes)
pkts bytes target prot opt in out source destination
28476 45M CONNMARK all * * ::/0 ::/0 CONNMARK restore
272 23796 MARK all enp0s1.220 * 2001:db8:7::/64 ::/0 MARK set 0xdc
1586 119K MARK all enp0s1.221 * 2001:db8:7::/64 ::/0 MARK set 0xdd
28476 45M CONNMARK all * * ::/0 ::/0 CONNMARK save
Chain INPUT (policy ACCEPT 2576 packets, 287K bytes)
pkts bytes target prot opt in out source destination
Chain FORWARD (policy ACCEPT 24514 packets, 45M bytes)
pkts bytes target prot opt in out source destination
Chain OUTPUT (policy ACCEPT 1031 packets, 74320 bytes)
pkts bytes target prot opt in out source destination
Chain POSTROUTING (policy ACCEPT 25597 packets, 45M bytes)
pkts bytes target prot opt in out source destination
\newpage
Here is an excerpt of blueTenant networking configuration file :
/etc/network/interfaces
# The primary network interface
auto enp0s1
iface enp0s1 inet static
address 172.16.220.2/24
gateway 172.16.220.1
dns-nameservers 9.9.9.9
iface enp0s1 inet6 static
address 2001:db8:dc::2/64
gateway fe80:dc::1
dns-nameservers 2620:fe::fe
auto C-3PO-BLUE
iface C-3PO-BLUE inet manual
ovs_type OVSBridge
ovs_ports sw-vlan7 vxlanBlueGreen
ovs_mtu 9000
up ip link set dev $IFACE up
down ip link set dev $IFACE down
allow-C-3PO-BLUE sw-vlan7
iface sw-vlan7 inet static
ovs_type OVSBridge
ovs_bridge C-3PO-BLUE
ovs_options C-3PO-BLUE 7
ovs_mtu 9000
address 192.0.2.220/24
iface sw-vlan7 inet6 static
ovs_type OVSBridge
ovs_bridge C-3PO-BLUE
ovs_options C-3PO-BLUE 7
ovs_mtu 9000
address 2001:db8:7::dc/64
up ip -6 addr add fe80:7::dc/64 dev $IFACE
allow-C-3PO-BLUE vxlanBlueGreen
iface vxlanBlueGreen inet manual
ovs_bridge C-3PO-BLUE
ovs_type OVSTunnel
ovs_tunnel_type vxlan
ovs_tunnel_options options:remote_ip=172.16.221.2 options:key=WhereIsMyTCAM
This blue virtual machine is, just as the others, a router. We want the traffic coming back and forth from the containers to be routed.
Edit /etc/sysctl.conf in order to get the following result.
$ egrep -v '(^#|^$)' /etc/sysctl.conf
net.ipv4.conf.default.rp_filter=2
net.ipv4.conf.all.rp_filter=2
net.ipv4.ip_forward=1
net.ipv6.conf.all.forwarding=1
net.ipv4.conf.all.log_martians = 1
Turn routing on after editing the above /etc/sysctl.conf file.
$ sudo sysctl --system
$ sudo apt install snapd
$ sudo snap install lxd
$ sudo adduser etu lxd
Normal user account is the UNprivileged user and must belong to lxd group.
Log out and log back in to make it effective.
$ id | grep -o lxd
lxd
List the installed snaps
$ snap list
Name Version Rev Tracking Publisher Notes
core 16-2.54.3 12725 latest/stable canonical✓ core
core18 20211215 2284 latest/stable canonical✓ base
core20 20220304 1376 latest/stable canonical✓ base
lxd 4.23 22525 latest/stable canonical✓ -
$ lxd init
Would you like to use LXD clustering? (yes/no) [default=no]: no
Do you want to configure a new storage pool? (yes/no) [default=yes]:
Name of the new storage pool [default=default]:
Name of the storage backend to use (btrfs, ceph, dir, lvm) [default=btrfs]:
Create a new BTRFS pool? (yes/no) [default=yes]:
Would you like to use an existing block device? (yes/no) [default=no]:
Size in GB of the new loop device (1GB minimum) [default=15GB]:
Would you like to connect to a MAAS server? (yes/no) [default=no]:
Would you like to create a new local network bridge? (yes/no) [default=yes]: no <-- CHANGE TO NO
Would you like to configure LXD to use an existing bridge or host interface? (yes/no) [default=no]: yes <-- CHANGE TO YES
Name of the existing bridge or host interface: sw-vlan7 <-- HERE WE USE OUR OWN SWITCH
Would you like LXD to be available over the network? (yes/no) [default=no]:
Would you like stale cached images to be updated automatically? (yes/no) [default=yes]
Would you like a YAML "lxd init" preseed to be printed? (yes/no) [default=no]:no
We have to change the nictype: from macvlan to bridged and we are done with the default profile.
$ lxc profile device set default eth0 nictype bridged
$ lxc profile device get default eth0 nictype
bridged
$ lxc launch images:debian/12 blueDHCP
Creating blueDHCP
Starting blueDHCP
$ lxc launch images:debian/12 blueC0
Creating blueC0
Starting blueC0
$ lxc launch images:debian/12 blueC1
Creating blueC1
Starting blueC1
We choose to use radvd on blueTenant as we want resilient and failover addressing on both Blue and Green tenants.
This daemon is installed on the virtual machine as the sw-vlan7 Switched Virtual Interface (SVI) stands on this machine.
$ sudo apt install radvd
$ sudo systemctl enable radvd
Here is a copy of the /etc/radvd.conf file for VLAN 7. Do not forget to
restart service after editing this configuration file.
interface sw-vlan7
{
AdvSendAdvert on;
AdvRASrcAddress {
fe80:7::de;
};
prefix 2001:db8:7::/64
{
AdvOnLink on;
AdvAutonomous on;
AdvRouterAddr on;
};
RDNSS 2620:fe::fe
{
};
};
The most important part of the above file is the router or gateway address
advertised : fe80:7::de. This address is under control of the
keepalived daemon.
We choose to use the ISC DHCP server as failover for dynamic IPv4 addressing is available.
We also choose to run this server into a container as it is an application layer service.
. Here is a copy of the /etc/network/interfaces of the blueDHCP container.
This container uses a static IPv4 address.
$ lxc exec blueDHCP -- cat /etc/network/interfaces
# The loopback network interface
auto lo
iface lo inet loopback
auto eth0
iface eth0 inet static
address 192.0.2.250/24
# Gateway address is 192.0.2.220 before keepalived setup
gateway 192.0.2.222
dns-nameservers 9.9.9.9
. Install ISC DHCP server package.
$ lxc exec blueDHCP -- apt install isc-dhcp-server
. Edit the /etc/default/isc-dhcp-server file to designate the network
interface eth0.
$ lxc exec blueDHCP -- sed -i 's/INTERFACESv4=""/INTERFACESv4="eth0"/g' /etc/default/isc-dhcp-server
$ lxc exec blueDHCP -- egrep -v '(^#|^$)' /etc/default/isc-dhcp-server
INTERFACESv4="eth0"
INTERFACESv6=""
. Edit the /etc/dhcp/dhcpd.conf to setup failover and address range for
dynamic IPv4 addressing.
$ lxc exec blueDHCP -- grep -v ^# /etc/dhcp/dhcpd.conf | cat -s
default-lease-time 600;
max-lease-time 7200;
ddns-update-style none;
failover peer "failover-partner" {
primary;
address 192.0.2.250;
port 519;
peer address 192.0.2.251;
peer port 520;
max-response-delay 60;
max-unacked-updates 10;
mclt 3600;
split 128;
load balance max seconds 3;
}
subnet 192.0.2.0 netmask 255.255.255.0 {
option domain-name-servers 9.9.9.9;
option routers 192.0.2.222;
pool {
failover peer "failover-partner";
range 192.0.2.10 192.0.2.90;
}
}
As in the case of IPv6 addressing, the most important part of the above file is
the router or gateway address advertised : 192.0.2.222. This address is under
control of the keepalived daemon.
Let's place three new scripts to be used with keepalived configuration in the
/usr/local/bin directory.
- Check IPv4 connectivity to the Internet from the keepalived daemon with the
/usr/local/bin/keepalived_check_ipv4.shscript.
#!/bin/bash
/usr/bin/ping -c 1 -W 1 9.9.9.9 > /dev/null 2>&1
- Check IPv6 connectivity to the Internet from the keepalived daemon with the
/usr/local/bin/keepalived_check_ipv6.shscript.
#!/bin/bash
/usr/bin/ping -c 1 -W 1 2620:fe::fe > /dev/null 2>&1
- Keepalived daemon status notifications with the
/usr/local/bin/keepalived_notify.shscript.
#!/bin/bash
echo "$1 $2 has transitioned to the $3 state with a priority of $4" > /var/run/keepalived_status
All these scripts should be executable.
$ sudo chmod +x /usr/local/bin/keepalived_*
After package installation, we edit the main configuration file.
$ sudo apt install keepalived
Here is a copy of the /etc/keepalived/keepalived.conf configuration file.
As usual, do not forget to restart service after editing.
! Configuration File for keepalived
global_defs {
notification_email {
root@localhost
}
notification_email_from etu@localhost
smtp_server localhost
smtp_connect_timeout 30
enable_script_security
script_user etu
vrrp_version 3
}
vrrp_script keepalived_check_ipv4 {
script "/usr/local/bin/keepalived_check_ipv4.sh"
interval 1
timeout 5
rise 3
fall 3
}
vrrp_script keepalived_check_ipv6 {
script "/usr/local/bin/keepalived_check_ipv6.sh"
interval 1
timeout 5
rise 3
fall 3
}
vrrp_sync_group vrrp_group {
group {
VXLAN_7_IPv4
VXLAN_7_IPv6
}
track_script {
keepalived_check_ipv4
}
notify "/usr/local/bin/keepalived_notify.sh"
}
vrrp_instance VXLAN_7_IPv4 {
state MASTER
interface sw-vlan7
virtual_router_id 74
priority 220
advert_int 1
virtual_ipaddress {
192.0.2.222/24
}
}
vrrp_instance VXLAN_7_IPv6 {
state MASTER
interface sw-vlan7
virtual_router_id 76
priority 220
advert_int 1
virtual_ipaddress {
fe80:7::de/64
}
}
We must not forget to restart the daemon
$ sudo systemctl restart keepalived.service
\newpage
The configuration files of the Blue and Green tenant are very similar. IP addresses and names are switched from blue to green.
Here is an excerpt of blueTenant networking configuration file : /etc/network/interfaces
# The primary network interface
auto enp0s1
iface enp0s1 inet static
address 172.16.221.2/24
gateway 172.16.221.1
dns-nameservers 9.9.9.9
iface enp0s1 inet6 static
address 2001:db8:dd::2/64
gateway fe80:dd::1
dns-nameservers 2620:fe::fe
auto C-3PO-GREEN
iface C-3PO-GREEN inet manual
ovs_type OVSBridge
ovs_ports sw-vlan7 vxlanGreenBlue
ovs_mtu 9000
up ip link set dev $IFACE up
down ip link set dev $IFACE down
allow-C-3PO-GREEN sw-vlan7
iface sw-vlan7 inet static
ovs_type OVSBridge
ovs_bridge C-3PO-GREEN
ovs_options C-3PO-GREEN 7
ovs_mtu 9000
address 192.0.2.221/24
iface sw-vlan7 inet6 static
ovs_type OVSBridge
ovs_bridge C-3PO-GREEN
ovs_options C-3PO-GREEN 7
ovs_mtu 9000
address 2001:db8:7::dd/64
up ip -6 addr add fe80:7::dd/64 dev $IFACE
allow-C-3PO-GREEN vxlanGreenBlue
iface vxlanGreenBlue inet manual
ovs_bridge C-3PO-GREEN
ovs_type OVSTunnel
ovs_tunnel_type vxlan
ovs_tunnel_options options:remote_ip=172.16.220.2 options:key=WhereIsMyTCAM
This green virtual machine is, just as the others, a router. We want the traffic coming back and forth from the containers to be routed.
Edit /etc/sysctl.conf in order to get the following result
$ egrep -v '(^#|^$)' /etc/sysctl.conf
net.ipv4.conf.default.rp_filter=2
net.ipv4.conf.all.rp_filter=2
net.ipv4.ip_forward=1
net.ipv6.conf.all.forwarding=1
net.ipv4.conf.all.log_martians = 1
Turn routing on after editing the above /etc/sysctl.conf file
$ sudo sysctl --system
$ sudo apt install snapd
$ sudo snap install lxd
$ sudo adduser etu lxd
Normal user account is the UNprivileged user and must belong to lxd group. Log out and log back in to make it effective.
$ id | grep -o lxd
lxd
List the installed snaps
$ snap list
Name Version Rev Tracking Publisher Notes
core 16-2.49 10859 latest/stable canonical core
core18 20210128 1988 latest/stable canonical base
lxd 4.12 19766 latest/stable canonical -
$ lxd init
Would you like to use LXD clustering? (yes/no) [default=no]: no
Do you want to configure a new storage pool? (yes/no) [default=yes]:
Name of the new storage pool [default=default]:
Name of the storage backend to use (btrfs, ceph, dir, lvm) [default=btrfs]:
Create a new BTRFS pool? (yes/no) [default=yes]:
Would you like to use an existing block device? (yes/no) [default=no]:
Size in GB of the new loop device (1GB minimum) [default=15GB]:
Would you like to connect to a MAAS server? (yes/no) [default=no]:
Would you like to create a new local network bridge? (yes/no) [default=yes]: no <-- CHANGE TO NO
Would you like to configure LXD to use an existing bridge or host interface? (yes/no) [default=no]: yes <-- CHANGE TO YES
Name of the existing bridge or host interface: sw-vlan7 <-- HERE WE USE OUR OWN SWITCH
Would you like LXD to be available over the network? (yes/no) [default=no]:
Would you like stale cached images to be updated automatically? (yes/no) [default=yes]
Would you like a YAML "lxd init" preseed to be printed? (yes/no) [default=no]:no
We have to change the nictype: from macvlan to bridged and we are done with the default profile.
$ lxc profile device set default eth0 nictype bridged
$ lxc profile device get default eth0 nictype
bridged
$ lxc launch images:debian/12 greenDHCP
Creating greenDHCP
Starting greenDHCP
$ lxc launch images:debian/12 greenC0
Creating greenC0
Starting greenC0
$ lxc launch images:debian/12 greenC1
Creating greenC1
Starting greenC1
We choose to use radvd on greenTenant as we want resilient and failover
addressing on both Blue and Green tenants.
This daemon is installed on the virtual machine as the sw-vlan7 Switched Virtual Interface (SVI) stands on this machine.
$ sudo apt install radvd
$ sudo systemctl enable radvd
Here is a copy of the /etc/radvd.conf file for VLAN 7. Do not forget to
restart service after editing this configuration file.
interface sw-vlan7
{
AdvSendAdvert on;
AdvRASrcAddress {
fe80:7::de;
};
prefix 2001:db8:7::de/64
{
AdvOnLink on;
AdvAutonomous on;
AdvRouterAddr on;
};
RDNSS 2620:fe::fe
{
};
};
The most important part of the above file is the router or gateway address
advertised : fe80:7::de. This address is under control of the keepalived
daemon.
We choose to use the ISC DHCP server as failover for dynamic IPv4 addressing is available.
We also choose to run this server into a container as it is an application layer service.
. Here is a copy of the /etc/network/interfaces of the greenDHCP container.
This container uses a static IPv4 address.
$ lxc exec greenDHCP -- cat /etc/network/interfaces
# The loopback network interface
auto lo
iface lo inet loopback
auto eth0
iface eth0 inet static
address 192.0.2.251/24
# Gateway address is 192.0.2.221 before keepalived setup
gateway 192.0.2.222
dns-nameservers 9.9.9.9
. Install ISC DHCP server package
$ lxc exec greenDHCP -- apt install isc-dhcp-server
. Edit the /etc/default/isc-dhcp-server file to designate the network
interface eth0
$ lxc exec greenDHCP -- sed -i 's/INTERFACESv4=""/INTERFACESv4="eth0"/g' /etc/default/isc-dhcp-server
$ lxc exec greenDHCP -- egrep -v '(^#|^$)' /etc/default/isc-dhcp-server
INTERFACESv4="eth0"
INTERFACESv6=""
. Edit the /etc/dhcp/dhcpd.conf to setup failover and address range for
dynamic IPv4 addressing
$ lxc exec greenDHCP -- grep -v ^# /etc/dhcp/dhcpd.conf | cat -s
default-lease-time 600;
max-lease-time 7200;
ddns-update-style none;
failover peer "failover-partner" {
secondary;
address 192.0.2.251;
port 520;
peer address 192.0.2.250;
peer port 519;
max-response-delay 60;
max-unacked-updates 10;
load balance max seconds 3;
}
subnet 192.0.2.0 netmask 255.255.255.0 {
option domain-name-servers 9.9.9.9;
option routers 192.0.2.222;
pool {
failover peer "failover-partner";
range 192.0.2.10 192.0.2.90;
}
}
As in the case of IPv6 addressing, the most important part of the above file is the router or gateway address advertised : 192.0.2.222. This address is under control of the keepalived daemon.
Just as it is done on the Blue tenant, we first have to place three new scripts to be used by keepalived configuration.
- Check IPv4 connectivity to the Internet from the keepalived daemon with the
/usr/local/bin/keepalived_check_ipv4.shscript.
#!/bin/bash
/usr/bin/ping -c 1 -W 1 9.9.9.9 > /dev/null 2>&1
- Check IPv6 connectivity to the Internet from the keepalived daemon with the
/usr/local/bin/keepalived_check_ipv6.shscript.
#!/bin/bash
/usr/bin/ping -c 1 -W 1 2620:fe::fe > /dev/null 2>&1
- Keepalived daemon status notifications with the
/usr/local/bin/keepalived_notify.shscript.
#!/bin/bash
echo "$1 $2 has transitioned to the $3 state with a priority of $4" > /var/run/keepalived_status
All these scripts should be executable.
$ sudo chmod +x /usr/local/bin/keepalived_*
When the above scripts are there, we carry on with package installation and configuration.
$ sudo apt install keepalived
Here is a copy of the /etc/keepalived/keepalived.conf configuration file.
As usual, do not forget to restart service after editing.
! Configuration File for keepalived
global_defs {
notification_email {
root@localhost
}
notification_email_from greenTenant@localhost
smtp_server localhost
smtp_connect_timeout 30
enable_script_security
script_user etu
vrrp_version 3
}
vrrp_script keepalived_check_ipv4 {
script "/usr/local/bin/keepalived_check_ipv4.sh"
interval 1
timeout 5
rise 3
fall 3
}
vrrp_script keepalived_check_ipv6 {
script "/usr/local/bin/keepalived_check_ipv6.sh"
interval 1
timeout 5
rise 3
fall 3
}
vrrp_sync_group vrrp_group {
group {
VXLAN_7_IPv4
VXLAN_7_IPv6
}
track_script {
keepalived_check_ipv4
}
notify "/usr/local/bin/keepalived_notify.sh"
}
vrrp_instance VXLAN_7_IPv4 {
state MASTER
interface sw-vlan7
virtual_router_id 74
priority 221
advert_int 1
virtual_ipaddress {
192.0.2.222/24
}
}
vrrp_instance VXLAN_7_IPv6 {
state MASTER
interface sw-vlan7
virtual_router_id 76
priority 221
advert_int 1
virtual_ipaddress {
fe80:7::de/64
}
}
We must not forget to restart the daemon
sudo systemctl restart keepalived.service
\newpage
. First, VXLAN ends know each other. Here are IPv4 and IPv6 tests from Green to Blue.
$ ping -q -c2 192.0.2.220
PING 192.0.2.220 (192.0.2.220) 56(84) bytes of data.
--- 192.0.2.220 ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1000ms
rtt min/avg/max/mdev = 1.099/1.189/1.280/0.090 ms
$ ping -q -c2 fe80:7::dc%sw-vlan7
PING fe80:7::dc%sw-vlan7(fe80:7::dc%sw-vlan7) 56 data bytes
--- fe80:7::dc%sw-vlan7 ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1002ms
rtt min/avg/max/mdev = 1.142/1.745/2.349/0.603 ms
. Second, we take a look at the switch TCAM (Ternary Content Addressable Memory) table for VLAN 7.
sudo ovs-appctl fdb/show C-3PO-GREEN
port VLAN MAC Age
1 7 3a:98:95:a9:6b:44 1
5 7 00:16:3e:b2:71:28 1
2 7 00:16:3e:1e:f5:70 1
2 7 4e:e8:f5:75:ec:4f 1
2 7 00:16:3e:2d:4f:16 1
2 7 00:16:3e:0b:7b:bc 1
3 7 00:16:3e:04:a9:e8 1
4 7 00:16:3e:70:16:53 1
. On the Blue tenant side.
$ lxc ls -c ns46
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
+----------+---------+--------------------+----------------------------------------+
| NAME | STATE | IPV4 | IPV6 |
+----------+---------+--------------------+----------------------------------------+
| blueC0 | RUNNING | 192.0.2.60 (eth0) | 2001:db8:7:0:216:3eff:fe2d:4f16 (eth0) |
+----------+---------+--------------------+----------------------------------------+
| blueC1 | RUNNING | 192.0.2.53 (eth0) | 2001:db8:7:0:216:3eff:fe0b:7bbc (eth0) |
+----------+---------+--------------------+----------------------------------------+
| blueC2 | RUNNING | 192.0.2.59 (eth0) | 2001:db8:7:0:216:3eff:fee2:4c76 (eth0) |
+----------+---------+--------------------+----------------------------------------+
| blueDHCP | RUNNING | 192.0.2.250 (eth0) | 2001:db8:7:0:216:3eff:fe1e:f570 (eth0) |
+----------+---------+--------------------+----------------------------------------+
. On the Green tenant side.
lxc ls -c ns46
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
+-----------+---------+--------------------+-----------------------------------------+
| NAME | STATE | IPV4 | IPV6 |
+-----------+---------+--------------------+-----------------------------------------+
| greenC0 | RUNNING | 192.0.2.19 (eth0) | 2001:db8:7:0:70be:e8ff:fe7a:a9e0 (eth0) |
+-----------+---------+--------------------+-----------------------------------------+
| greenC1 | RUNNING | 192.0.2.20 (eth0) | 2001:db8:7:0:216:3eff:fe70:1653 (eth0) |
+-----------+---------+--------------------+-----------------------------------------+
| greenDHCP | RUNNING | 192.0.2.251 (eth0) | 2001:db8:7:0:216:3eff:feb2:7128 (eth0) |
+-----------+---------+--------------------+-----------------------------------------+
. On the Blue tenant side.
$ lxc exec blueC1 -- ip route ls
default via 192.0.2.222 dev eth0
192.0.2.0/24 dev eth0 proto kernel scope link src 192.0.2.53
$ lxc exec blueC1 -- ip -6 route ls
2001:db8:7::/64 dev eth0 proto kernel metric 256 expires 86289sec pref medium
fe80::/64 dev eth0 proto kernel metric 256 pref medium
default via fe80:7::de dev eth0 proto ra metric 1024 expires 1689sec hoplimit 64 pref medium
. On the Green tenant side.
$ lxc exec greenC1 -- ip route ls
default via 192.0.2.222 dev eth0
192.0.2.0/24 dev eth0 proto kernel scope link src 192.0.2.55
$ lxc exec greenC1 -- ip -6 route ls
2001:db8:7::/64 dev eth0 proto kernel metric 256 expires 86175sec pref medium
fe80::/64 dev eth0 proto kernel metric 256 pref medium
default via fe80:7::de dev eth0 proto ra metric 1024 expires 1575sec hoplimit 64 pref medium
. ICMP tests on the Blue tenant side: 0% packet loss.
$ for c in blueC0 blueC1 blueDHCP; do echo ------------// $c && lxc exec $c -- ping -q -c3 9.9.9.9; done
------------// blueC0
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
PING 9.9.9.9 (9.9.9.9) 56(84) bytes of data.
--- 9.9.9.9 ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2004ms
rtt min/avg/max/mdev = 14.588/16.602/20.451/2.722 ms
------------// blueC1
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
PING 9.9.9.9 (9.9.9.9) 56(84) bytes of data.
--- 9.9.9.9 ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2003ms
rtt min/avg/max/mdev = 14.378/15.808/16.893/1.055 ms
------------// blueDHCP
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
PING 9.9.9.9 (9.9.9.9) 56(84) bytes of data.
--- 9.9.9.9 ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2004ms
rtt min/avg/max/mdev = 14.776/15.699/16.276/0.659 ms
$ for c in blueC0 blueC1 blueDHCP; do echo ------------// $c && lxc exec $c -- ping -q -c3 2620:fe::fe; done
------------// blueC0
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
PING 2620:fe::fe(2620:fe::fe) 56 data bytes
--- 2620:fe::fe ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2003ms
rtt min/avg/max/mdev = 40.765/41.883/44.098/1.566 ms
------------// blueC1
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
PING 2620:fe::fe(2620:fe::fe) 56 data bytes
--- 2620:fe::fe ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2004ms
rtt min/avg/max/mdev = 40.502/42.337/44.911/1.874 ms
------------// blueDHCP
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
PING 2620:fe::fe(2620:fe::fe) 56 data bytes
--- 2620:fe::fe ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2004ms
rtt min/avg/max/mdev = 41.009/43.974/45.594/2.099 ms
. ICMP tests on the Green tenant side: 0% packet loss.
$ for c in greenC0 greenC1 greenDHCP; do echo ------------// $c && lxc exec $c -- ping -q -c3 9.9.9.9; done
------------// greenC0
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
PING 9.9.9.9 (9.9.9.9) 56(84) bytes of data.
--- 9.9.9.9 ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2004ms
rtt min/avg/max/mdev = 13.106/13.386/13.805/0.301 ms
------------// greenC1
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
PING 9.9.9.9 (9.9.9.9) 56(84) bytes of data.
--- 9.9.9.9 ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2001ms
rtt min/avg/max/mdev = 12.997/15.535/19.150/2.624 ms
------------// greenDHCP
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
PING 9.9.9.9 (9.9.9.9) 56(84) bytes of data.
--- 9.9.9.9 ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2004ms
rtt min/avg/max/mdev = 13.140/13.546/13.751/0.287 ms
$ for c in greenC0 greenC1 greenDHCP; do echo ------------// $c && lxc exec $c -- ping -q -c3 2620:fe::fe; done
------------// greenC0
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
PING 2620:fe::fe(2620:fe::fe) 56 data bytes
--- 2620:fe::fe ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2003ms
rtt min/avg/max/mdev = 44.135/44.822/46.163/0.948 ms
------------// greenC1
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
PING 2620:fe::fe(2620:fe::fe) 56 data bytes
--- 2620:fe::fe ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2003ms
rtt min/avg/max/mdev = 39.431/42.982/45.335/2.554 ms
------------// greenDHCP
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
PING 2620:fe::fe(2620:fe::fe) 56 data bytes
--- 2620:fe::fe ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2003ms
rtt min/avg/max/mdev = 43.803/44.232/44.768/0.400 ms
. HTTP(s) on the Blue tenant side.
$ for c in blueC0 blueC1 blueDHCP; do echo ------------// $c && lxc exec $c -- apt update; done
------------// blueC0
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
Hit:1 http://deb.debian.org/debian bullseye InRelease
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
All packages are up to date.
------------// blueC1
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
Hit:1 http://deb.debian.org/debian bullseye InRelease
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
All packages are up to date.
------------// blueDHCP
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
Hit:1 http://deb.debian.org/debian bullseye InRelease
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
All packages are up to date.
. HTTP(s) on the Green tenant side.
$ for c in greenC0 greenC1 greenDHCP; do echo ------------// $c && lxc exec $c -- apt update; done
------------// greenC0
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
Hit:1 http://deb.debian.org/debian bullseye InRelease
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
All packages are up to date.
------------// greenC1
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
Hit:1 http://deb.debian.org/debian bullseye InRelease
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
All packages are up to date.
------------// greenDHCP
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
Hit:1 http://deb.debian.org/debian bullseye InRelease
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
All packages are up to date.
The two following commands give the list of the stateful entries stored in netfilter tables.
$ sudo conntrack -f ipv4 -L | grep mark=221
udp 17 3 src=192.0.2.251 dst=9.9.9.9 sport=59798 dport=53 src=9.9.9.9 dst=198.18.28.200 sport=53 dport=59798 [ASSURED] mark=221 use=1
udp 17 0 src=192.0.2.20 dst=9.9.9.9 sport=34668 dport=53 src=9.9.9.9 dst=198.18.28.200 sport=53 dport=34668 [ASSURED] mark=221 use=1
udp 17 0 src=192.0.2.20 dst=9.9.9.9 sport=57195 dport=53 src=9.9.9.9 dst=198.18.28.200 sport=53 dport=57195 mark=221 use=1
conntrack v1.4.6 (conntrack-tools): 86 flow entries have been shown.
udp 17 3 src=192.0.2.251 dst=9.9.9.9 sport=43648 dport=53 src=9.9.9.9 dst=198.18.28.200 sport=53 dport=43648 mark=221 use=1
$ sudo conntrack -f ipv6 -L | grep mark=221
conntrack v1.4.6 (conntrack-tools): 6 flow entries have been shown.
tcp 6 105 TIME_WAIT src=2001:db8:7:0:216:3eff:feb2:7128 dst=2a04:4e42:2d::644 sport=52746 dport=80 src=2a04:4e42:2d::644 dst=2001:678:3fc:1c::c8 sport=80 dport=52746 [ASSURED] mark=221 use=1
tcp 6 99 TIME_WAIT src=2001:db8:7:0:70be:e8ff:fe7a:a9e0 dst=2a04:4e42:2d::644 sport=51062 dport=80 src=2a04:4e42:2d::644 dst=2001:678:3fc:1c::c8 sport=80 dport=51062 [ASSURED] mark=221 use=1
tcp 6 102 TIME_WAIT src=2001:db8:7:0:216:3eff:fe70:1653 dst=2a04:4e42:2d::644 sport=38244 dport=80 src=2a04:4e42:2d::644 dst=2001:678:3fc:1c::c8 sport=80 dport=38244 [ASSURED] mark=221 use=1
. First, we have to locate the IPv4 and IPv6 gateway addresses. Stated that the
Green tenant has the highest VRRP priority, the sw-vlan7 interface on the
Green side hold the gateway adresses.
$ ip addr ls dev sw-vlan7
6: sw-vlan7: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc noqueue state UNKNOWN group default qlen 1000
link/ether c2:8e:8d:87:0c:4c brd ff:ff:ff:ff:ff:ff
inet 192.0.2.221/24 brd 192.0.2.255 scope global sw-vlan7
valid_lft forever preferred_lft forever
inet 192.0.2.222/24 scope global secondary sw-vlan7
valid_lft forever preferred_lft forever
inet6 2001:db8:7::dd/64 scope global
valid_lft forever preferred_lft forever
inet6 fe80:7::de/64 scope link nodad
valid_lft forever preferred_lft forever
inet6 fe80:7::dd/64 scope link
valid_lft forever preferred_lft forever
inet6 fe80::c08e:8dff:fe87:c4c/64 scope link
valid_lft forever preferred_lft forever
We also have the ability to check the logs for VRRP state.
$ grep 'vrrp.*state' /var/log/syslog | tail -1
Mar 14 09:48:04 greenTenant Keepalived_vrrp[1339]: VRRP_Group(vrrp_group) Syncing instances to MASTER state
. Second, we take a look at the default gateway IPv4 and IPv6 addresses on the Blue tenant containers.
$ for c in blueC0 blueC1 blueDHCP; do echo ------------// $c && lxc exec $c -- ip route ls default; done
------------// blueC0
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
default via 192.0.2.222 dev eth0
------------// blueC1
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
default via 192.0.2.222 dev eth0
------------// blueDHCP
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
default via 192.0.2.222 dev eth0 onlink
$ for c in blueC0 blueC1 blueDHCP; do echo ------------// $c && lxc exec $c -- ip -6 route ls default; done
------------// blueC0
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
default via fe80:7::de dev eth0 proto ra metric 1024 expires 1328sec hoplimit 64 pref medium
------------// blueC1
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
default via fe80:7::de dev eth0 proto ra metric 1024 expires 1328sec hoplimit 64 pref medium
------------// blueDHCP
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
default via fe80:7::de dev eth0 proto ra metric 1024 expires 1328sec hoplimit 64 pref medium
. Third, we simulate Internet connectivity loss on the Green side with an
iptables rule which drops traffic.
$ sudo iptables -A OUTPUT -d 9.9.9.9/32 -j DROP
Then we check the logs.
$ tail -n 7 /var/log/syslog
Mar 14 19:03:06 greenTenant Keepalived_vrrp[1339]: VRRP_Script(keepalived_check_ipv4) failed (exited with status 1)
Mar 14 19:03:06 greenTenant Keepalived_vrrp[1339]: (VXLAN_7_IPv4) Entering FAULT STATE
Mar 14 19:03:06 greenTenant Keepalived_vrrp[1339]: (VXLAN_7_IPv4) sent 0 priority
Mar 14 19:03:06 greenTenant Keepalived_vrrp[1339]: VRRP_Group(vrrp_group) Syncing instances to FAULT state
Mar 14 19:03:06 greenTenant Keepalived_vrrp[1339]: (VXLAN_7_IPv6) Entering FAULT STATE
Mar 14 19:03:06 greenTenant Keepalived_vrrp[1339]: (VXLAN_7_IPv6) Entering FAULT STATE
Mar 14 19:03:06 greenTenant avahi-daemon[385]: Withdrawing address record for 192.0.2.222 on sw-vlan7.
. Fourth, The Green has lost the gateway addresses.
$ ip addr ls dev sw-vlan7
6: sw-vlan7: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc noqueue state UNKNOWN group default qlen 1000
link/ether c2:8e:8d:87:0c:4c brd ff:ff:ff:ff:ff:ff
inet 192.0.2.221/24 brd 192.0.2.255 scope global sw-vlan7
valid_lft forever preferred_lft forever
inet6 2001:db8:7::dd/64 scope global
valid_lft forever preferred_lft forever
inet6 fe80:7::dd/64 scope link
valid_lft forever preferred_lft forever
inet6 fe80::c08e:8dff:fe87:c4c/64 scope link
valid_lft forever preferred_lft forever
The Blue is now holding the gateway addresses.
$ ip addr ls dev sw-vlan7
6: sw-vlan7: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc noqueue state UNKNOWN group default qlen 1000
link/ether 6e:aa:78:68:4d:49 brd ff:ff:ff:ff:ff:ff
inet 192.0.2.220/24 brd 192.0.2.255 scope global sw-vlan7
valid_lft forever preferred_lft forever
inet 192.0.2.222/24 scope global secondary sw-vlan7
valid_lft forever preferred_lft forever
inet6 2001:db8:7::dc/64 scope global
valid_lft forever preferred_lft forever
inet6 fe80:7::de/64 scope link nodad
valid_lft forever preferred_lft forever
inet6 fe80:7::dc/64 scope link
valid_lft forever preferred_lft forever
inet6 fe80::6caa:78ff:fe68:4d49/64 scope link
valid_lft forever preferred_lft forever
. Fifth, we run connectivity tests from the containers of both tenants.
$ for c in blueC0 blueC1 blueDHCP; do echo ------------// $c && lxc exec $c -- apt update; done
------------// blueC0
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
Hit:1 http://deb.debian.org/debian bullseye InRelease
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
All packages are up to date.
------------// blueC1
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
Hit:1 http://deb.debian.org/debian bullseye InRelease
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
All packages are up to date.
------------// blueDHCP
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
Hit:1 http://deb.debian.org/debian bullseye InRelease
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
All packages are up to date.
$ for c in greenC0 greenC1 greenDHCP; do echo ------------// $c && lxc exec $c -- apt update; done
------------// greenC0
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
Hit:1 http://deb.debian.org/debian bullseye InRelease
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
All packages are up to date.
------------// greenC1
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
Hit:1 http://deb.debian.org/debian bullseye InRelease
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
All packages are up to date.
------------// greenDHCP
WARNING: cgroup v2 is not fully supported yet, proceeding with partial confinement
Hit:1 http://deb.debian.org/debian bullseye InRelease
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
All packages are up to date.
. Sixth, we check the logs to see the Blue tenant is now the VRRP master.
$ grep vrrp.*state /var/log/syslog | tail -1
Mar 14 19:03:06 blueTenant Keepalived_vrrp[1323]: VRRP_Group(vrrp_group) Syncing instances to MASTER state
. Finally, things go back to initial state when the iptables rule is deleted
on the Green side.
$ sudo iptables -D OUTPUT -d 9.9.9.9/32 -j DROP
$ tail -n 15 /var/log/syslog
Mar 14 19:12:42 greenTenant Keepalived_vrrp[1339]: (VXLAN_7_IPv6) Entering BACKUP STATE
Mar 14 19:12:42 greenTenant Keepalived_vrrp[1339]: VRRP_Group(vrrp_group) Syncing instances to BACKUP state
Mar 14 19:12:42 greenTenant Keepalived_vrrp[1339]: (VXLAN_7_IPv4) Entering BACKUP STATE
Mar 14 19:12:43 greenTenant Keepalived_vrrp[1339]: (VXLAN_7_IPv6) received lower priority (220) advert from fe80:7::dc - discarding
Mar 14 19:12:43 greenTenant Keepalived_vrrp[1339]: (VXLAN_7_IPv4) received lower priority (220) advert from 192.0.2.220 - discarding
Mar 14 19:12:44 greenTenant Keepalived_vrrp[1339]: (VXLAN_7_IPv4) received lower priority (220) advert from 192.0.2.220 - discarding
Mar 14 19:12:44 greenTenant Keepalived_vrrp[1339]: (VXLAN_7_IPv6) received lower priority (220) advert from fe80:7::dc - discarding
Mar 14 19:12:45 greenTenant Keepalived_vrrp[1339]: (VXLAN_7_IPv4) received lower priority (220) advert from 192.0.2.220 - discarding
Mar 14 19:12:45 greenTenant Keepalived_vrrp[1339]: (VXLAN_7_IPv6) received lower priority (220) advert from fe80:7::dc - discarding
Mar 14 19:12:46 greenTenant Keepalived_vrrp[1339]: (VXLAN_7_IPv4) Entering MASTER STATE
Mar 14 19:12:46 greenTenant Keepalived_vrrp[1339]: (VXLAN_7_IPv4) using locally configured advertisement interval (1000 milli-sec)
Mar 14 19:12:46 greenTenant Keepalived_vrrp[1339]: VRRP_Group(vrrp_group) Syncing instances to MASTER state
Mar 14 19:12:46 greenTenant Keepalived_vrrp[1339]: (VXLAN_7_IPv6) Entering MASTER STATE
Mar 14 19:12:46 greenTenant Keepalived_vrrp[1339]: (VXLAN_7_IPv6) using locally configured advertisement interval (1000 milli-sec)
Mar 14 19:12:46 greenTenant avahi-daemon[385]: Registering new address record for 192.0.2.222 on sw-vlan7.IPv4.
This lab illustrates network gateway resiliency for containers hosted among two different tenants.
The VXLAN technology allows to share a single LAN or broadcast domain over an IP interconnection between these two tenants.
Hope this will give you ideas to experiment and investigate further ;).