ifupdown2 is the network interface manager for Cumulus Linux. It is an updated version of the original ifupdown in both Cumulus Linux and Debian. This article demonstrates the differences between the two in how you configure interfaces.

Viewing Every ifupdown2 Configuration Keyword/Option

To determine which options you can configure for the bridge settings or bond settings, run:

ifquery --syntax-help

Loopback Interfaces

ifupdown2 lets you configure IP addresses on loopback interfaces:

ifupdown

ifupdown2

auto lo
iface lo inet loopback
auto lo:1
iface lo:1 inet static
    address 10.1.1.1/32
auto lo:2
iface lo:2 inet static
   address 10.2.2.2/32

auto lo
iface lo inet loopback
   address 10.1.1.1/32
   address 10.2.2.2/32

Layer 2 Interfaces - Physical

In the original ifupdown, when the method is manual, you had to specify up link set $IFACE up and down link set $IFACE down in the configuration. ifupdown2 does this for you automatically.

ifupdown	ifupdown2
auto swp19 iface swp19 inet manual up link set $IFACE up down link set $IFACE down pre-up /sbin/ethtool -s $IFACE speed 1000	auto swp19 iface swp19 link-speed 1000

The following is some debug code:

sudo ifup -d swp19
DEBUG: args = Namespace(CLASS=None, all=False, debug=True, excludepats=None, force=False, iflist=['swp19'], jobs=-1, noact=False, nocache=False, perfmode=False, printdependency=None, quiet=False, verbose=False, withdepends=False)
DEBUG: creating ifupdown object ..
INFO: loading builtin modules from /usr/share/ifupdownaddons
INFO: looking for user scripts under /etc/network
INFO: loading scripts under /etc/network/if-pre-up.d ...
INFO: loading scripts under /etc/network/if-up.d ...
INFO: loading scripts under /etc/network/if-post-up.d ...
INFO: loading scripts under /etc/network/if-pre-down.d ...
INFO: loading scripts under /etc/network/if-down.d ...
INFO: loading scripts under /etc/network/if-post-down.d ...
DEBUG: reading interfaces file /etc/network/interfaces
WARNING: template engine mako not found. skip template parsing ..
DEBUG: populating dependency info for ['swp19']
DEBUG: run_without_dependents for ops ['pre-up', 'up', 'post-up'] for ['swp19']
DEBUG: swp19: pre-up : running module bridge
DEBUG: swp19: pre-up : running module mstpctl
DEBUG: swp19: pre-up : running module vlan
DEBUG: swp19: pre-up : running module ifenslave
DEBUG: swp19: pre-up : running module address
DEBUG: swp19: pre-up : running module usercmds
DEBUG: running cmd 'ethtool -s swp19 speed 1000'
INFO: Executing ethtool -s swp19 speed 1000
DEBUG: swp19: pre-up : running script /etc/network/if-pre-up.d/ethtool
DEBUG: Executing /etc/network/if-pre-up.d/ethtool
DEBUG: swp19: up : running module dhcp
DEBUG: swp19: up : running module link
INFO: Executing ip link set dev swp19 up <===== Run for you now!
DEBUG: swp19: up : running script /etc/network/if-up.d/ethtool
DEBUG: Executing /etc/network/if-up.d/ethtool
DEBUG: swp19: up : running script /etc/network/if-up.d/ip
DEBUG: Executing /etc/network/if-up.d/ip
DEBUG: swp19: up : running script /etc/network/if-up.d/mountnfs
DEBUG: Executing /etc/network/if-up.d/mountnfs
DEBUG: swp19: up : running script /etc/network/if-up.d/openssh-server
DEBUG: Executing /etc/network/if-up.d/openssh-server
DEBUG: swp19: post-up : running module usercmds

Layer 2 Interfaces - Trunk

With ifupdown2, you don't need to define sub-interfaces to add to the trunk. Just add it under the bridge interface.

Notice that bridge-ports is used instead of mstpctl-ports.

By default, bridge-stp turns on Rapid Spanning Tree.

ifupdown

ifupdown2

auto swp19
iface swp19 inet manual
  up link set $IFACE up
  down link set $IFACE down
  pre-up /sbin/ethtool -s $IFACE speed 1000

auto swp19.100
iface swp19.100 inet manual
  up link set $IFACE up
  down link set $IFACE down

auto swp19.200
iface swp19.200 inet manual
  up link set $IFACE up
  down link set $IFACE down

auto vlan100
iface vlan100 inet manual
bridge_ports swp19.100
mstpctl_stp on

auto vlan200
iface vlan200 inet manual
mstpctl_ports swp19.200
mstpctl_stp on

auto swp19
iface swp19
  link-speed 1000

auto vlan100
iface vlan100
  bridge-ports swp19.100
  bridge-stp on

auto vlan200
iface vlan200
  bridge-ports swp19.200
  bridge-stp on

Layer 2 Bond Interface - with Trunking

The following shows you how to configure bonds with trunking under ifupdown2.

ifupdown

ifupdown2

auto swp3
iface swp3 inet manual
  up link set $IFACE up
  down link set $IFACE down
  
auto swp4
iface swp4 inet manual
  up link set $IFACE up
  down link set $IFACE down
  
auto bond0
iface bond0 inet manual
  up link set $IFACE up
  down link set $IFACE down
  bond-slaves swp3 swp4
  bond-miimon 100
  bond-min-links 1
  bond-mode 802.3ad
  bond-xmit-hash-policy layer3+4
  bond-lacp-rate

auto bond0.100
iface bond0.100 inet manual
  up link set $IFACE up
  down link set $IFACE down
  
auto bond0.200
iface bond0.200 inet manual
  up link set $IFACE up
  down link set $IFACE down
  
auto vlan100
iface vlan100 inet manual
  up link set $IFACE up
  down link set $IFACE down
  mstpctl_ports bond0.100
  mstpctl_stp on
  
auto vlan200
iface vlan200 inet manual
  up link set $IFACE up
  down link set $IFACE down
  mstpctl_ports bond0.200
  mstpctl_stp on

auto swp3
iface swp3

auto swp4
iface swp4

auto bond0
iface bond0
  bond-slaves swp3 swp4
  bond-miimon 100
  bond-min-links 1
  bond-mode 802.3ad
  bond-xmit-hash-policy layer3+4
  bond-lacp-rate 1

auto vlan100
iface vlan100 
  bridge-ports bond0.100
  bridge-stp on

auto vlan200
iface vlan200
  bridge-ports bond0.200
  bridge-stp on

Layer 2 Trunk - Port Ranges

The glob keyword replaces regular expressions for creating port ranges, because it doesn't require the interfaces to be mentioned in /etc/network/interfaces. Multiple glob statements can be mentioned in the stanza so one can do a discontiguous port range.

ifupdown

ifupdown2

auto swp3
iface swp3 inet manual
  up link set $IFACE up
  down link set $IFACE down
  
auto swp3.100
iface swp3.100 inet manual
  up link set $IFACE up
  down link set $IFACE down
  
auto swp4
iface swp4 inet manual
  up link set $IFACE up
   down link set $IFACE down
   
auto swp4.100
iface swp4.100 inet manual
  up link set $IFACE up
  down link set $IFACE down

auto swp12
iface swp12 inet manual
  up link set $IFACE up
  down link set $IFACE down
  
auto swp12.100
iface swp12.100 inet manual
  up link set $IFACE up
  down link set $IFACE down
  
auto swp21
iface swp21 inet manual
  up link set $IFACE up
  down link set $IFACE down
  
auto swp21.100
iface swp21.100 inet manual
  up link set $IFACE up
  down link set $IFACE down
  
auto swp22
iface swp22 inet manual
  up link set $IFACE up
  down link set $IFACE down
  
auto swp22
iface swp22.100 inet manual
  up link set $IFACE up
  down link set $IFACE down
  
auto vlan100
iface vlan100 inet manual
  up link set $IFACE up
  down link set $IFACE down
  mstpctl_ports swp3.100 swp4.100 swp12.100 swp21.100 swp22.100
  mstpctl_stp on

auto swp3
iface swp3

auto swp4
iface swp4

auto swp12
iface swp12

auto swp21
iface swp21

auto swp22
iface swp22

auto vlan100
iface vlan100
  bridge-ports glob swp3-4.100 swp12.100 swp21-22.100
  bridge-stp on

IPv6 Address Assignment

ifupdown2 does not require IPv6 addresses to be configured under an independent inet6 section. All interface configuration is done under the same section. The same is true for the loopback configuration as well.

Single IPv6 Address

ifupdown	ifupdown2
auto swp1 iface swp1 inet6 manual address 2001:db8::1/64	auto swp1 iface swp1 address 2001:db8::1/64

Multiple IPv6 Addresses

ifupdown	ifupdown2
auto swp1:1 iface swp1:1 inet6 manual address 2001:db8::1:1/64 auto swp1:2 iface swp1:2 inet6 manual address 2001:db8::2:2/64	auto swp1 iface swp1 address 2001:db8::1:1/64 address 2001:db8::2:2/64

Setting Speed, Duplex and Auto-negotiation on a Port

ifupdown2 now supports keywords to set the speed, duplex and auto-negotiation.

ifupdown

ifupdown2

auto swp1
iface swp1 inet static
  address 10.1.1.1/24
  mtu 9000
  pre-up /sbin/ethtool -s $IFACE speed 1000 duplex half autoneg off
  up ip link set $IFACE up
  down ip link set $IFACE down

auto swp1
iface swp1
  address 10.1.1.1/24
  mtu 9000
  link-speed 1000
  link-duplex half
  link-autoneg off

Setting the Port Description

ifupdown2 lets you configure the IP address on loopback interfaces.

ifupdown

ifupdown2

auto swp1
iface swp1 inet static
  address 10.1.1.1/24
  post-up ip link set $IFACE alias customerA
  up ip link set $IFACE up
  down ip link set $IFACE down

auto swp1
iface swp1
  address 10.1.1.1/24
  alias customerA

# ip link show swp1
3: swp1: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast state DOWN mode DEFAULT qlen 500
    link/ether 08:9e:01:ce:dc:c2 brd ff:ff:ff:ff:ff:ff
    alias customerA

Using the "source" Keyword to Place Interface Configuration in Files Other than /etc/network/interfaces

This is useful for automation. This example shows how Ansible writes individual port configs into the /etc/network/ansible/ directory with ifupdown2.

To view all the configuration, run ifquery -a.

auto lo
iface lo inet loopback

auto eth0
iface eth0 inet dhcp

source /etc/network/ansible/*

# tree /etc/network/ansible
/etc/network/ansible
|-- swp1
|-- swp2
|-- swp3
`-- swp4
0 directories, 4 files
  
# ifquery -a
auto lo
iface lo inet loopback
    address 10.3.3.3/32
    address 10:3:3::3/128
auto eth0
iface eth0 inet dhcp
auto swp1
iface swp1
    link-speed 1000
auto swp2
iface swp2
    link-speed 1000
auto swp3
iface swp3
    link-speed 10000
auto swp4
iface swp4
    address 10.200.1.1/24

Creating the Default Configuration Using Mako

Mako is a templating engine that can be used to generate the /etc/network/interfaces configuration. It can be used to generate a default configuration. The example below shows how to set defaults for bonds. By default, ifupdown2 will read Mako files in the /etc/network/ifupdown2/templates directory. You can change this location in the /etc/network/ifupdown2/ifupdown2.conf. To view the expanded configuration, like a running config, run ifquery bond0 or ifquery -a.

ifupdown2 - /etc/network/interfaces

ifupdown2 - /etc/network/mako/bond_defaults

iface lo inet loopback

auto eth0
iface eth0 inet dhcp

 

### define name of default files and functions to use
<%namespace file='bond_defaults' import='bond_defaults'/>


auto swp1
iface swp1
   alias bond0-member

auto swp2
iface swp2
    alias bond0-member

auto bond0
iface bond0
     bond-slaves swp1 swp2
     ${bond_defaults()}

<%def name="bond_defaults()">
    bond-miimon 100 
    bond-min-links 1 
    bond-mode 802.3ad 
    bond-xmit-hash-policy layer3+4 
    bond-lacp-rate 1
</%def>

ifquery Output

#sudo ifquery -a
 
auto lo
iface lo inet loopback
 
auto eth0
iface eth0 inet dhcp
 
auto swp1
iface swp1
 
auto swp2
iface swp2
 
auto bond0
iface bond0
        bond-slaves swp1 swp2
        bond-miimon 100
        bond-min-links 1
        bond-mode 802.3ad
        bond-xmit-hash-policy layer3+4
        bond-lacp-rate 1

Comparing ifupdown2 Commands with ifupdown Commands

{{table_of_contents}}

Viewing Every ifupdown2 Configuration Keyword/Option

Loopback Interfaces

Layer 2 Interfaces - Physical

Layer 2 Interfaces - Trunk

Layer 2 Bond Interface - with Trunking

Layer 2 Trunk - Port Ranges

IPv6 Address Assignment

Multiple IPv6 Addresses

Setting Speed, Duplex and Auto-negotiation on a Port

Setting the Port Description

Using the "source" Keyword to Place Interface Configuration in Files Other than /etc/network/interfaces

Creating the Default Configuration Using Mako

ifquery Output

See Also

Comments

Products

Learn

About