Ubuntu Server 16.04 LTS isn’t managed with NetworkManager and does not come installed by default like it did in prior releases. It is managed by /etc/networks/interfaces. In this tutorial we are going to learn how to restart network in Ubuntu Server 16.04 as well as previous releases.

Restart Network Service in Ubuntu Server 16.04

systemctl restart networking.service

Remove All IP Information from Interface

ip addr flush <if-name>

So to combine the two would look like this: ip addr flush eth0 && systemctl restart networking.service

We will start with the most recommended commands and move down to more obscure or obsolete commands in case the above commands fail from some reason.

systemctl restart NetworkManager.service

Next command service will restart a relevant networking System V init script:

service networking restart
service network-manager restart

The following command nmcli is a command-line tool for controlling NetworkManager which will restart the Network Manager directly:

nmcli networking off
nmcli networking on

The old fashion way using System V init scripts directly is still available on Ubuntu 16.04:

/etc/init.d/networking restart
/etc/init.d/network-manager restart

In the last example we will restart the network interface directly using ifup and ifdown commands. Note the -a option instructs the ifup and ifdown commands to restart all available network interfaces marked “auto” in /etc/networks/interfaces:

ifdown -a
ifup -a
#If you're connected via SSH it is best to combine the commands:
ifdown eth0 && ifup eth0

RTNETLINK answers: Cannot assign requested address

Network Interfaces

The first step in troubleshooting network issues will be to identify which network interfaces are present on the system. See Network configuration#Get current device names for details.

Link status

In the overview of ip a, the link status will already be displayed. But it can also be displayed by running:

ip link show dev eth0

This will provide an output along the lines of:

2: eth0: &lt;BROADCAST,MULTICAST,UP,LOWER_UP&gt; mtu 1500 qdisc pfifo_fast state DOWN qlen 1000
 link/ether 70:5a:b6:8a:a0:87 brd ff:ff:ff:ff:ff:ff

Bringing up an interface can be done by issueing:

# ip link set dev eth0 up

RTNETLINK answers: Cannot assign requested address

If you get this error when trying to set an interface up, its most probably because you’ve got an invalid MAC address. To set a working MAC, see MAC address spoofing.

IP address

In the overview provided by ip a, the ip address will already be displayed. But it can also be displayed by running:

ip addr show dev eth0

This will provide an output along the lines of:

eth0: &lt;BROADCAST,MULTICAST,UP,LOWER_UP&gt; mtu 1500 qdisc pfifo_fast state UP qlen 1000
 link/ether 70:5a:b6:8a:a0:87 brd ff:ff:ff:ff:ff:ff
 inet brd scope global eth0
 inet6 fe80::725a:b6ff:fe8a:a087/64 scope link
 valid_lft forever preferred_lft forever

Adding a temporary ip address:

ip addr add dev eth0

Removing an ip address:

ip addr del dev eth0

Route table

The route table can be displayed by running:

ip route show

Route table for a specific interface:

ip route show dev eth0

This will provide an output along the lines of:

default via proto static proto kernel scope link src

Configuring the default gateway:

ip route add 0/0 via dev eth0

Removing the default gateway:

ip route del 0/0 via dev eth0

DNS Servers

Dns is responsible for converting hostnames to an ip address. When connectivity towards ip addresses is working, but the system is unable to connect to a hostname; there is a fair chance that this will be related to the dns configuration. The configuration can be displayed by running:

cat /etc/resolv.conf

This will provide an output among the lines of:


  • The rule ‘nameserver’ is the relevant section. Configuring multiple nameservers is supported.
  • The ‘domain’ and ‘search’ rules are optional.
  • Often the ‘nameserver’ is the same as your default gateway.
  • In case of doubt there is always the possibility to use the Google DNS servers as your default DNS servers:


Testing your dns configuration can be done through the drill command (from the ldns package):

drill @

The above command will perform a DNS lookup of using the DNS server and return output as follows:

;; -&gt;&gt;HEADER&lt;&lt;- opcode: QUERY, rcode: NOERROR, id: 35144
;; flags: qr rd ra ; QUERY: 1, ANSWER: 2, AUTHORITY: 0, ADDITIONAL: 0
;; IN A
;; Query time: 38 msec
;; WHEN: Wed Jun 17 20:21:47 2015
;; MSG SIZE rcvd: 72

As an alternative to drill, there are also the dig, host, and nslookup tools from the bind-tools package.

Ping & Tracepath/Traceroute

The ping command can help test connectivity towards a specific host.

The first step would be verifying connectivity towards the default gateway (replace the ip address with your own default gateway):

ping -c4

When erasing the -c4 parameter, the ping will continue endlessly. It can be aborted by hitting Control-C.

PING ( 56(84) bytes of data.
64 bytes from icmp_req=1 ttl=64 time=0.193 ms
64 bytes from icmp_req=2 ttl=64 time=0.190 ms
64 bytes from icmp_req=3 ttl=64 time=0.192 ms
64 bytes from icmp_req=4 ttl=64 time=0.189 ms

The output above indicated the default gateway is reachable. When instead a “Destination Host Unreachable” message is displayed, doublecheck the ip address, netmask and default gateway config. This message can also be displayed when ICMP traffic is not permitted towards the default gateway (blocked by a firewall, router,…).

The next step is verifying connectivity towards the configured dns server(s). When no reply is received, tracepath or traceroute can be used to verify the routing towards said server and get an idea of where the issue lies.


Traceroute also used ICMP to determine the path and hence there can be “no reply” answers as well when ICMP traffic is blocked.

[credit goes to]

Page generated in 0.3724 seconds.
Advertisment ad adsense adlogger