Networking / Stage 3 of 14

Persistent networking with Netplan

Not running Starting usually ~20 s

Step 7 of 9 · lab · 25 min

Netplan: restore the link to the neighbour

The training interface lab0 is up, yet the neighbour 192.0.2.1 does not answer. Find the cause and fix the persistent network configuration without touching anything else.

Situation

This machine has a training interface, lab0. A veth pair connects it to the neighbour 192.0.2.1, which lives in the lab-peer network namespace. The interface is up, but the neighbour does not answer ping.

According to the address plan, lab0 has 192.0.2.10/24, and the address must come from the persistent Netplan configuration: lab0 is described in /etc/netplan/50-lab.yaml. lab0 is an interface of the machine itself: it needs no -n or ip netns exec.

Task

In /etc/netplan/50-lab.yaml and in Netplan's merged configuration, lab0 has the single address 192.0.2.10/24, dhcp4: false and renderer: networkd; the file keeps its owner and mode, root:root and 600.
The up lab0 now has exactly one IPv4 address, 192.0.2.10/24.
The path to 192.0.2.1 goes directly through lab0 with the source address 192.0.2.10.
The neighbour 192.0.2.1 answers ping and keeps its address.
The rest of the machine's network, other network settings files and the lab0–peer0 pair are untouched, and there are no workaround services or scripts.

Constraints

A fix with a single ip command does not solve the task: after netplan apply or a reboot the address comes from the configuration again.
Don't recreate the neighbour or the interface pair, and don't set a global renderer.

The tools you need are already installed. There's no internet on this machine.

Lab · on your own

No commands are given in this lab: type them in the terminal yourself. Goals are checked on the machine after every command.

Goals

0 of 5

Netplan gives lab0 only 192.0.2.10/24, dhcp4 false, renderer networkd — not done yet
The up lab0 now has exactly one IPv4 address — 192.0.2.10/24 — not done yet
The path to 192.0.2.1 goes directly through lab0 with source 192.0.2.10 — not done yet
Neighbour 192.0.2.1 answers ping and keeps its address — not done yet
The rest of the network, its settings and the training interface pair are untouched — not done yet

Ctrl+Enter

Starts the machine node1 · Ubuntu 24.04

The machine is starting

Hints

0 of 3

Report a problem

Lab goals

0 of 5

Netplan gives lab0 only 192.0.2.10/24, dhcp4 false, renderer networkd — not done yet
The up lab0 now has exactly one IPv4 address — 192.0.2.10/24 — not done yet
The path to 192.0.2.1 goes directly through lab0 with source 192.0.2.10 — not done yet
Neighbour 192.0.2.1 answers ping and keeps its address — not done yet
The rest of the network, its settings and the training interface pair are untouched — not done yet

node1 Ubuntu 24.04

Lab goals

0 of 5

Netplan gives lab0 only 192.0.2.10/24, dhcp4 false, renderer networkd — not done yet
The up lab0 now has exactly one IPv4 address — 192.0.2.10/24 — not done yet
The path to 192.0.2.1 goes directly through lab0 with source 192.0.2.10 — not done yet
Neighbour 192.0.2.1 answers ping and keeps its address — not done yet
The rest of the network, its settings and the training interface pair are untouched — not done yet

Network · picked for this lab

ip — addresses, interfaces and routes

Cheat sheet

$_ — insert into the terminal

ip -br a addresses, brief
ip -br l interfaces, brief
ip r routing table
ip route get 8.8.8.8 path a specific packet would take
ip -s link show eth0 counters and errors
ip neigh ARP
sudo ip a add 192.168.1.10/24 dev eth0 add an address temporarily
permanent: /etc/netplan/*.yaml + sudo netplan apply

Output modifiers (-br, -s, -c, -d, -h) go right after ip and only affect how the result is printed, not which object is selected: ip -br a and ip a -br are the same. -4/-6 narrow the output to one protocol version — useful when an interface has both IPv4 and IPv6 and the extra lines get in the way.

ip objects

Object	About	Common
`ip address` (`a`)	IP addresses on interfaces	`ip -br a`
`ip link` (`l`)	the interfaces themselves (up/down, MTU, MAC)	`ip -br l`
`ip route` (`r`)	the routing table	`ip r`
`ip neigh` (`n`)	ARP/neighbors (IP↔MAC)	`ip n`
`ip -s link`	packet/error counters	per interface
`ip route get`	which route will be chosen	`ip route get IP`

Commands

bash

ip -br a                          # addresses, brief (name · state · address)
ip -c a                           # with color
ip a show eth0                    # a single interface
ip -br l                          # interface list and link state
ip -s link show eth0              # RX/TX counters, errors, drops
ip r                              # routing table
ip route get 8.8.8.8              # which route/interface a packet takes
ip neigh                          # the ARP table (neighbors)

# changes (TEMPORARY, until reboot):
sudo ip link set eth0 up          # bring the interface up
sudo ip link set eth0 mtu 1400    # change MTU
sudo ip a add 192.168.1.10/24 dev eth0    # add an address
sudo ip a del 192.168.1.10/24 dev eth0    # remove an address
sudo ip r add default via 192.168.1.1     # default gateway
sudo ip r add 10.10.0.0/16 via 192.168.1.254   # static route
sudo ip neigh flush all           # clear the ARP cache

Changes made with ip ... are temporary (until reboot). For permanent configuration on Ubuntu Server, use netplan (/etc/netplan/*.yaml, sudo netplan apply).

Address, prefix and subnet

192.0.2.10/24 is an address and a prefix: the first 24 bits (192.0.2) are the network number, the rest is the host number. Addresses with the same network number form the subnet 192.0.2.0/24: from 192.0.2.0 to 192.0.2.255, hosts get .1–.254. Neighbours in a subnet exchange packets directly; another subnet needs a route — without one ping prints Network is unreachable at once (see ping).

bash

ip -4 -br a                       # addresses with prefixes
ip r                              # subnets you have a path to
ip route get 192.0.2.1            # which way a packet will go

Network diagnostics

bash

ip -br a                          # is there an address at all
ip -br l                          # is the link up? (NO-CARRIER = cable/port)
ip route get 8.8.8.8              # is there a route out, and through what
ping -c3 192.168.1.1              # is the gateway reachable
getent hosts example.com          # does DNS work (see dns-resolv)
ss -tlnp                          # what ports listen locally

The natural ladder: address → link → route/gateway → external IP → DNS → port (see network-diagnostics).

No address (`ip -br a` empty/down)

The interface is down (sudo ip link set <if> up), or DHCP/netplan didn't run (sudo netplan apply, journalctl -u systemd-networkd).

`NO-CARRIER` in `ip -br l`

No physical link: cable/port/switch. Check the hardware with ethtool.

There's an address, but "Network is unreachable"

No default route: ip r has no default line → add a gateway (in netplan).

Two interfaces, traffic goes "the wrong way"

ip route get <IP> shows the actual choice. It's governed by route metrics and order; fix it in netplan (metric:) or with policy routing.

IP pings, names don't

That's DNS, not ip: resolvectl status, getent hosts name (see dns-resolv).

Persistent networking with Netplan

Situation

Task

Constraints

Signing in is needed to start the machine

ip objects

Commands

Address, prefix and subnet

Network diagnostics

No address (ip -br a empty/down)

NO-CARRIER in ip -br l

There's an address, but "Network is unreachable"

Two interfaces, traffic goes "the wrong way"

IP pings, names don't

No address (`ip -br a` empty/down)

`NO-CARRIER` in `ip -br l`