Configuring Network Interfaces

Why This Matters

You have just been handed SSH credentials to a brand-new bare-metal server in your company's data center. The machine boots, but it has no IP address. Nobody can reach it, and it cannot reach the internet. Your job is to bring it onto the network -- assign an IP, set a default gateway, configure DNS, and make sure everything survives a reboot.

This is the bread and butter of Linux system administration. Every single server, VM, container host, or IoT device you will ever touch needs its network configured. Whether you are setting up a DHCP-based laptop or a multi-homed production server with VLANs and bonded interfaces, the tools in this chapter are what you will reach for.

Try This Right Now

Open a terminal on any Linux machine and run:

# Show all network interfaces and their addresses
ip addr show

# Show only interfaces that are UP
ip link show up

# Show the routing table
ip route show

You should see at least two interfaces: lo (the loopback) and something like eth0, ens33, enp0s3, or wlp2s0 (your real network interface). Take note of your interface name -- you will need it throughout this chapter.

The `ip` Command: Your Primary Tool

The ip command from the iproute2 package is the modern, standard way to configure networking on Linux. It replaced the older ifconfig, route, and arp commands. If you only learn one networking tool, make it ip.

The ip command is organized into objects:

ip <object> <command>

Objects:
  addr     - IP addresses
  link     - Network interfaces (layer 2)
  route    - Routing table
  neigh    - ARP / neighbor cache
  netns    - Network namespaces

Viewing Interface Information

# Full details of all interfaces
ip addr show

# Short form
ip a

# Just one interface
ip addr show dev eth0

# Only show IPv4 addresses
ip -4 addr show

# Only show IPv6 addresses
ip -6 addr show

# Machine-readable JSON output
ip -j addr show | python3 -m json.tool

Here is what typical ip addr show output looks like:

2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000
    link/ether 52:54:00:a1:b2:c3 brd ff:ff:ff:ff:ff:ff
    inet 192.168.1.100/24 brd 192.168.1.255 scope global eth0
       valid_lft forever preferred_lft forever
    inet6 fe80::5054:ff:fea1:b2c3/64 scope link
       valid_lft forever preferred_lft forever

Let's decode each piece:

+------------------------------------------------------------------+
| Field                        | Meaning                           |
|------------------------------|-----------------------------------|
| eth0                         | Interface name                    |
| <BROADCAST,MULTICAST,UP,     | Interface flags                   |
|  LOWER_UP>                   | UP=admin up, LOWER_UP=cable in    |
| mtu 1500                     | Maximum Transmission Unit         |
| state UP                     | Operational state                 |
| link/ether 52:54:00:a1:b2:c3| MAC address                       |
| inet 192.168.1.100/24        | IPv4 address with prefix length   |
| brd 192.168.1.255            | Broadcast address                 |
| scope global                 | Address is globally reachable     |
| inet6 fe80::...              | IPv6 link-local address           |
+------------------------------------------------------------------+

Managing Interfaces with `ip link`

# Bring an interface down
sudo ip link set eth0 down

# Bring it back up
sudo ip link set eth0 up

# Change the MTU
sudo ip link set eth0 mtu 9000

# Change the MAC address (interface must be down)
sudo ip link set eth0 down
sudo ip link set eth0 address 02:00:00:00:00:01
sudo ip link set eth0 up

# Show interface statistics
ip -s link show eth0

Safety Warning: Running ip link set eth0 down on a remote server over SSH through that same interface will immediately disconnect you. Always ensure you have out-of-band access (console, IPMI, or a second interface) before taking an interface down remotely.

Managing IP Addresses with `ip addr`

# Add an IP address to an interface
sudo ip addr add 192.168.1.100/24 dev eth0

# Add a second IP address (yes, one interface can have multiple)
sudo ip addr add 192.168.1.101/24 dev eth0

# Remove an IP address
sudo ip addr del 192.168.1.101/24 dev eth0

# Flush ALL addresses from an interface
sudo ip addr flush dev eth0

Safety Warning: ip addr flush will remove ALL IP addresses from the interface. On a remote machine, this is just as dangerous as taking the interface down.

Managing Routes with `ip route`

# View the routing table
ip route show

# Add a default gateway
sudo ip route add default via 192.168.1.1

# Add a specific route
sudo ip route add 10.0.0.0/8 via 192.168.1.254

# Delete a route
sudo ip route del 10.0.0.0/8

# Replace a route (add or update)
sudo ip route replace default via 192.168.1.1

Think About It: You have added an IP address and a default gateway with ip commands. You reboot the server. Are those settings still there? Why or why not?

The answer is no. Everything done with the ip command is temporary. It lives in the kernel's memory and vanishes on reboot. To make changes permanent, you need a configuration file or a network management daemon. That is what the rest of this chapter covers.

The Legacy: `ifconfig`

You will still see ifconfig in older scripts, blog posts, and Stack Overflow answers. It comes from the net-tools package and is considered deprecated.

# View interfaces (legacy)
ifconfig

# Assign an IP (legacy)
sudo ifconfig eth0 192.168.1.100 netmask 255.255.255.0 up

Why you should use ip instead:

ifconfig cannot show all addresses on an interface (only the primary)
ifconfig cannot handle advanced features like policy routing or network namespaces
ifconfig is not installed by default on many modern distributions
The iproute2 suite (ip) is actively maintained; net-tools is not

Distro Note: On minimal installs of RHEL/CentOS, Fedora, Debian, and Ubuntu Server, ifconfig may not even be present. The ip command is always available.

NetworkManager: The Desktop and Server Standard

NetworkManager is the most widely used network management daemon on Linux today. It is the default on Fedora, RHEL, CentOS, Ubuntu Desktop, and many others. It handles wired, wireless, VPN, and mobile broadband connections.

+-------------------------------------------------------+
|                    NetworkManager                      |
|                                                       |
|  nmcli   nmtui   GNOME Settings   nm-connection-editor|
|    |        |          |                  |            |
|    +--------+----------+------------------+            |
|                    |                                   |
|            NetworkManager daemon                       |
|                    |                                   |
|            Kernel networking stack                     |
+-------------------------------------------------------+

`nmcli`: The Command-Line Interface

nmcli is the most powerful way to interact with NetworkManager from the terminal.

# Show overall status
nmcli general status

# List all connections
nmcli connection show

# Show active connections
nmcli connection show --active

# Show details of a specific connection
nmcli connection show "Wired connection 1"

# List all devices
nmcli device status

Creating a Static IP Connection

# Create a new connection with a static IP
sudo nmcli connection add \
  con-name "static-eth0" \
  type ethernet \
  ifname eth0 \
  ipv4.addresses 192.168.1.100/24 \
  ipv4.gateway 192.168.1.1 \
  ipv4.dns "8.8.8.8 8.8.4.4" \
  ipv4.method manual

# Activate it
sudo nmcli connection up "static-eth0"

Switching to DHCP

# Create a DHCP connection
sudo nmcli connection add \
  con-name "dhcp-eth0" \
  type ethernet \
  ifname eth0 \
  ipv4.method auto

# Activate it
sudo nmcli connection up "dhcp-eth0"

Modifying an Existing Connection

# Change the DNS servers
sudo nmcli connection modify "static-eth0" ipv4.dns "1.1.1.1 9.9.9.9"

# Add a secondary IP address
sudo nmcli connection modify "static-eth0" +ipv4.addresses 192.168.1.101/24

# Apply changes without taking the connection down
sudo nmcli connection up "static-eth0"

Quick Cheat Sheet

# Bring a connection down
sudo nmcli connection down "static-eth0"

# Delete a connection
sudo nmcli connection delete "static-eth0"

# Set a connection to auto-connect on boot
sudo nmcli connection modify "static-eth0" connection.autoconnect yes

# Show the wifi networks (on a laptop)
nmcli device wifi list

# Connect to wifi
sudo nmcli device wifi connect "MyNetwork" password "secret123"

`nmtui`: The Text User Interface

If you prefer a visual, menu-driven approach in the terminal:

sudo nmtui

This launches a curses-based interface where you can:

Edit a connection
Activate a connection
Set the system hostname

It is perfect for quick configuration when you do not want to remember nmcli syntax.

systemd-networkd: Lightweight Network Configuration

On servers, containers, and embedded systems, systemd-networkd is a lighter alternative to NetworkManager. It is part of systemd and uses simple .network configuration files.

Enabling systemd-networkd

# If NetworkManager is running, disable it first to avoid conflicts
sudo systemctl stop NetworkManager
sudo systemctl disable NetworkManager

# Enable systemd-networkd and systemd-resolved
sudo systemctl enable --now systemd-networkd
sudo systemctl enable --now systemd-resolved

Configuration Files

Configuration lives in /etc/systemd/network/. Files are processed in alphabetical order and use the .network extension.

Static IP Configuration

Create /etc/systemd/network/20-wired.network:

[Match]
Name=eth0

[Network]
Address=192.168.1.100/24
Gateway=192.168.1.1
DNS=8.8.8.8
DNS=8.8.4.4

DHCP Configuration

Create /etc/systemd/network/20-wired.network:

[Match]
Name=eth0

[Network]
DHCP=yes

After creating or modifying files:

# Reload and apply
sudo networkctl reload

# Check status
networkctl status eth0
networkctl list

Distro Note: Ubuntu Server (since 17.10) uses Netplan, which is a YAML-based abstraction layer that can generate configuration for either NetworkManager or systemd-networkd. We cover Netplan shortly.

Debian-style: `/etc/network/interfaces`

On Debian and older Ubuntu systems (before Netplan), the classic configuration file is /etc/network/interfaces. It is managed by the ifupdown package.

Static IP

# /etc/network/interfaces

auto lo
iface lo inet loopback

auto eth0
iface eth0 inet static
    address 192.168.1.100
    netmask 255.255.255.0
    gateway 192.168.1.1
    dns-nameservers 8.8.8.8 8.8.4.4

DHCP

auto eth0
iface eth0 inet dhcp

Applying Changes

# Bring an interface down and back up
sudo ifdown eth0 && sudo ifup eth0

# Or restart the networking service
sudo systemctl restart networking

RHEL-style: Network Scripts (Legacy)

On older RHEL, CentOS (7 and earlier), and Fedora systems, network configuration lived in per-interface scripts under /etc/sysconfig/network-scripts/.

Static IP

# /etc/sysconfig/network-scripts/ifcfg-eth0

TYPE=Ethernet
DEVICE=eth0
BOOTPROTO=none
ONBOOT=yes
IPADDR=192.168.1.100
PREFIX=24
GATEWAY=192.168.1.1
DNS1=8.8.8.8
DNS2=8.8.4.4

DHCP

# /etc/sysconfig/network-scripts/ifcfg-eth0

TYPE=Ethernet
DEVICE=eth0
BOOTPROTO=dhcp
ONBOOT=yes

Applying Changes

sudo systemctl restart network
# or for a single interface:
sudo ifdown eth0 && sudo ifup eth0

Distro Note: RHEL 9 and later have removed legacy network scripts entirely. NetworkManager with nmcli is the only supported method. The ifcfg files may still work (NetworkManager reads them), but the new keyfile format in /etc/NetworkManager/system-connections/ is preferred.

Netplan (Ubuntu)

Ubuntu Server uses Netplan as an abstraction layer. You write YAML files, and Netplan generates the backend configuration.

Configuration lives in /etc/netplan/. The file is usually named something like 01-netcfg.yaml or 50-cloud-init.yaml.

Static IP

# /etc/netplan/01-netcfg.yaml
network:
  version: 2
  renderer: networkd       # or NetworkManager
  ethernets:
    eth0:
      addresses:
        - 192.168.1.100/24
      routes:
        - to: default
          via: 192.168.1.1
      nameservers:
        addresses:
          - 8.8.8.8
          - 8.8.4.4

DHCP

network:
  version: 2
  ethernets:
    eth0:
      dhcp4: true

Applying Changes

# Test the configuration (auto-reverts after 120 seconds if you don't confirm)
sudo netplan try

# Apply permanently
sudo netplan apply

# Generate backend config without applying
sudo netplan generate

The netplan try command is brilliant for remote servers -- if your configuration is broken, it will automatically revert, saving you from being locked out.

Static vs DHCP: When to Use Which

+--------------------------------------------------------------+
|              Static IP                |      DHCP             |
|---------------------------------------|---------------------- |
| Servers                               | Desktops/laptops      |
| Network infrastructure                | Guest networks        |
| DNS servers                           | Development VMs       |
| Load balancers                        | IoT devices (some)    |
| Database servers                      | Containers (often)    |
|                                       |                       |
| You control the exact address         | Address assigned       |
| Survives DHCP server outages          |   automatically       |
| Required for services others connect  | Less config to manage |
|   to by IP                            | Easy to move between  |
|                                       |   networks            |
+--------------------------------------------------------------+

Think About It: Your company has a DHCP server that hands out addresses in the range 192.168.1.100-200. You manually assign 192.168.1.150 as a static IP to your new server. What could go wrong?

The DHCP server might hand out 192.168.1.150 to another machine, creating an IP conflict. The fix is to either use a static IP outside the DHCP range or create a DHCP reservation for your server's MAC address.

Hands-On: Configure a Network Interface from Scratch

Let's walk through configuring a static IP using nmcli, which works across most modern distributions.

Step 1: Identify your interface

nmcli device status

Expected output:

DEVICE  TYPE      STATE         CONNECTION
eth0    ethernet  connected     Wired connection 1
lo      loopback  unmanaged     --

Step 2: Create a new connection profile

sudo nmcli connection add \
  con-name "my-static" \
  type ethernet \
  ifname eth0 \
  ipv4.addresses 10.0.0.50/24 \
  ipv4.gateway 10.0.0.1 \
  ipv4.dns "1.1.1.1" \
  ipv4.method manual \
  connection.autoconnect yes

Step 3: Activate the connection

sudo nmcli connection up "my-static"

Step 4: Verify

ip addr show eth0
ip route show
cat /etc/resolv.conf
ping -c 3 1.1.1.1

Step 5: Test DNS resolution

ping -c 3 google.com

If the ping to 1.1.1.1 works but google.com does not resolve, your DNS configuration needs fixing.

VLANs: Virtual LANs

VLANs let you segment a single physical network into multiple logical networks. On Linux, you create a VLAN sub-interface tagged with a VLAN ID.

# Using ip command (temporary)
sudo ip link add link eth0 name eth0.100 type vlan id 100
sudo ip addr add 10.100.0.1/24 dev eth0.100
sudo ip link set eth0.100 up

# Using nmcli (persistent)
sudo nmcli connection add \
  con-name "vlan100" \
  type vlan \
  ifname eth0.100 \
  dev eth0 \
  id 100 \
  ipv4.addresses 10.100.0.1/24 \
  ipv4.method manual

To verify:

ip -d link show eth0.100
cat /proc/net/vlan/eth0.100

Network Bonding: Combining Interfaces

Bonding (also called NIC teaming) combines two or more physical interfaces into one logical interface for redundancy or increased throughput.

+-------------------+
|    bond0          |  <-- logical interface (192.168.1.100)
|   /       \       |
| eth0     eth1     |  <-- physical interfaces
+-------------------+

Common bonding modes:

Mode	Name	Description
0	balance-rr	Round-robin, requires switch support
1	active-backup	One active, others standby (most common)
2	balance-xor	XOR-based hash
4	802.3ad	LACP, requires switch support
6	balance-alb	Adaptive load balancing

Creating a Bond with `nmcli`

# Create the bond
sudo nmcli connection add \
  con-name "bond0" \
  type bond \
  ifname bond0 \
  bond.options "mode=active-backup,miimon=100" \
  ipv4.addresses 192.168.1.100/24 \
  ipv4.gateway 192.168.1.1 \
  ipv4.method manual

# Add slave interfaces
sudo nmcli connection add \
  con-name "bond0-slave1" \
  type ethernet \
  ifname eth0 \
  master bond0

sudo nmcli connection add \
  con-name "bond0-slave2" \
  type ethernet \
  ifname eth1 \
  master bond0

# Activate
sudo nmcli connection up bond0

To check bond status:

cat /proc/net/bonding/bond0

Debug This

Scenario: A junior admin reports that a newly deployed server has no network connectivity. They show you this:

$ ip addr show eth0
2: eth0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN group default qlen 1000
    link/ether 52:54:00:a1:b2:c3 brd ff:ff:ff:ff:ff:ff

$ ip route show
(empty output)

What are the problems? How do you fix them?

Diagnosis:

The interface is DOWN -- notice the flags show <BROADCAST,MULTICAST> but no UP or LOWER_UP. Also, state DOWN and qdisc noop (no queuing discipline assigned).
There is no IP address assigned.
There is no routing table -- no default gateway.

Fix:

# Step 1: Bring the interface up
sudo ip link set eth0 up

# Step 2: Assign an IP address
sudo ip addr add 192.168.1.100/24 dev eth0

# Step 3: Add a default gateway
sudo ip route add default via 192.168.1.1

# Step 4: Make it permanent (using nmcli)
sudo nmcli connection add con-name "eth0-static" type ethernet ifname eth0 \
  ipv4.addresses 192.168.1.100/24 ipv4.gateway 192.168.1.1 \
  ipv4.dns "8.8.8.8" ipv4.method manual
sudo nmcli connection up "eth0-static"

What Just Happened?

+-------------------------------------------------------------------+
|                     Chapter 33 Recap                               |
+-------------------------------------------------------------------+
|                                                                   |
|  * `ip` is the modern tool for viewing and configuring            |
|    interfaces, addresses, and routes. Changes are temporary.      |
|                                                                   |
|  * NetworkManager (`nmcli`, `nmtui`) makes configuration          |
|    persistent and handles complex setups. Default on most distros.|
|                                                                   |
|  * systemd-networkd is a lighter daemon controlled via            |
|    .network files in /etc/systemd/network/.                       |
|                                                                   |
|  * Distro-specific methods:                                       |
|    - Debian: /etc/network/interfaces                              |
|    - RHEL legacy: /etc/sysconfig/network-scripts/                 |
|    - Ubuntu modern: Netplan (/etc/netplan/*.yaml)                 |
|                                                                   |
|  * Static IPs are for servers and infrastructure.                 |
|    DHCP is for clients and dynamic environments.                  |
|                                                                   |
|  * VLANs segment traffic on a single physical interface.          |
|  * Bonding combines multiple interfaces for redundancy.           |
|                                                                   |
+-------------------------------------------------------------------+

Try This

Basic configuration: Using nmcli, create a connection profile with a static IP of your choice. Verify it works with ping, then switch the same interface to DHCP.
Multi-address: Add three IP addresses to a single interface. Verify all three respond to ping from another machine on the same network.
Explore your system: Run nmcli connection show and examine every field in one of your connection profiles. Identify at least five settings you did not know about.
Compare methods: Configure the same static IP using ip commands, then using nmcli, then using a Netplan YAML file (on Ubuntu) or a manual config file for your distribution. Reboot after each method and verify which ones survive.
Bonus challenge: Set up a VLAN sub-interface on your machine. You will need a switch or virtual network that supports VLAN tagging. Verify the VLAN interface gets its own IP address in a different subnet than the parent interface.

Linux Book: From First Boot to Production