Configuring network in RHEL7
As a Linux server administrator, you need to manage network interfaces assigning IP addresses to them. The IP addresses can be assigned
in two ways:
- Static IP addresses: Typically used for servers as they need to be always available at the same IP address.
- Dynamically assigned IP addresses: This method is usually used for desktops, tablets and cellphones. To dynamically assign
IP addresses, a Dynamic Host Configuration Protocol (DHCP) server is usually used. The latter automatically assigns an IP
address whenever a computer connects to the network for a certain amount of time (lease time). If computer is up and
running at the end of the lease the lease is automatically renewed. If it is down the after some period of time the
IP address will be reused.
To configure the static IP addressing you need to specify an IP address, a subnet mask, and a default gateway. On you home network you router is you default gateway and often have the first IP address in the network ID
or the last address.
Some recommendations fro creating home network or some test network in the cloud
- Try to use the IP addresses in some kind of sequential order.
- Try to separate servers IP space from clients IP-space. For example, servers could have higher half of host address space, while clients can have
lower half. So for servers in the network 192.168.0.0/24) you assign address starting from 192.168.0.254 down and
for clients starting from 192.168.0.2 up until those two meet.
- Always verify assigned addresses with
ping before and after — Before ping should fail,
after should return packets. it’s too easy to make a typo when you enter static IP addresses. This also prevent assigning
duplicate IPs.
- Create your IP-map document. For example as Excel spreadsheet. Write down whatever you choose to do so the person who comes after you understands.
These principles have become unofficial standards for network techs, and following them simplify troubleshooting and allow others faster
understand your network setup.
In any case you need to provide the computers on your network with IP address, subnet mask, and default gateway.
If you’ve verified that this IP is not used, entered an IP address and your ping is not successful, first check your IP settings. Odds are good you made a typo. Otherwise,
check your connections, driver, and so forth.
Static addressing has been around for a long time and is still heavily used for more critical
systems on your network. Static addressing poses one big problem, however: making any changes to the network is a serious pain.
That's why corporations usually use it only for servers. For desktop more flexible method to get their IP information: dynamic IP
addressing is used.
There are two special IP addresses you need to know about.
- The first special address is 127.0.0.1—the loopback address. When you tell a device to send data to 127.0.0.1, you’re telling that device to send the packets to itself.
The loopback address has a number of uses. One of the most common is to use it with the ping command. I use the command
ping 127.0.0.1
to test a NIC’s capability to send and receive packets. Even though, by convention, you use 127.0.0.1 as the loopback address, the entire 127.0.0.0/8 subnet is reserved for loopback addresses.
You can use any address in the 127.0.0.0/8 subnet as a loopback address.
- The second is the last IP address of your subnet, For example, doe the subnet 192.168.0.0/24, 192.168.0.255 -- broadcast address
For a long time, network cards in Linux have had default names eth0, eth1, and eth2. This naming was assigned based on the
order of detection of the network card. So, eth0 was the first network card that got detected, eth1 the second, and so on.
This approach works well for servers and was used in RHEL 4 till RHEL 6 (although in RHEL 6 network manager was also an option).
In RHEL 7 we have obligatory additional component -- an intermediary called Network Manager, which was initially designed for laptop
linux installation and now got into server space by stealth.
For laptops you typically need to take care of the "on the fly" change of the type of connection. When laptop is moved
from office to home or from cubical, or office the conference room, the type of connection can change. For example, people usually
use wired connection at office (via laptop docking station) and WiFi at conference room and home. All those things are not needed
for server, which is assigned static IP once and use it until decomissing in five or more years down the road.
There was a political struggle in Red Hat in which "Linux on desktop" faction won and now server admins need to learn and used
more complex setup which essentially was designed for the laptops and involves the component, we mentioned above -- the Network manager.
Another complication in RHEL 7 is that the default names for network cards are now based on firmware, device topology, and
device types. This leads to bizarre network card names. which are impossible to memorize and creates some problems for scripts, Now
names are dependent on the vendor of equipment and type of adapter.
In this, pretty bizarre, convention the name now consists of the following parts:
- Prefix: Ethernet interfaces begin with en, WLAN interfaces begin with wl, and so on.
- Middle part: The next part of the name represents the type of adapter. An o is used for onboard, s is
for a hotplug slot, and p is for a PCI location. Administrators can also use the x to create a device name that
is based on the MAC address of the network card.
- Suffix: a number, which is used to represent an index, ID, or port.
Apart from this default device naming, network cards can be named based on the BIOS device name as well. In this naming, names such
as em1 (embedded network card 1) or p4p1 (which is PCI slot 4, port 1) can be used.
Those name are generated by the biosdevname
package, developed by Dell and which screw up things even further. The idea was tosimplify matching interfaces with thir
marking on the server shassi, but the toad to hell is paved with good intentions.
First of all, it is idiotic to apply this to blades, in which all interfaces are hidden and any potential help in indentifying to which slot
cable was connected by the name of the interface is completely lost. But in its infinite wisdom the biosdevname
package is activated on blades in RHEL7 and you need to deactivate it after the installation unless you use Kickstart.
And the imposition of two different naming conventions on hapless sysadmins makes even less
sense and is a clear sign of the loss of architectural integrity.
If this "RHEL7-style" name cannot be generated, traditional names such as eth0 are used. You can also forcefully assign
old eth-type names by supplying parameters biosdevname=0 and net.ifnames=0 to GRUB2 bootloader.
In most case in RHEL7 after installation of the OS you need to deal with the this new Ethernet interface scheme and names such as ens2. From the name we
can guess that this is a hotplug slot Ethernet device, with its unique index number 2.
Generally this idea of "games with device names" look
counterproductive and complicate writing scripts for Linux.
By default not all necessary RPMs are installed in CentOS. Some of them should probably added immediately after the installation
(anaconda in RHEl 7 is a joke and does not allow specify individual packages unless you use Kickstart
The following can probably be added to create a more comfortable work environment, especially is you worked with previous version
of RHEL:
yum install bash-completion # packages below use bash completion, so it is useful to add it
yum install net-tools # provides "legacy" ifconfig, netstat, route, arp and several other commands
yum install traceroute.x86_64 # traceroute
yum install tcpdump # tcpdump -- sometimes needed but keeing it on the production server is not recommanded as represents some security risk
# so it should be de-installed after you finish installation
To set the network configuration for a new server you need to know the following five parameters:
- IP address
- Subnet mask
- Default router
- DNS server IP (ususally two) and search strings
- The name of the active Ethernet interface
The first thing we must do is find out the name of our Ethernet interface that is connected to network. You need also set it "up".
But which interface is connected might be a problem if you have 4, or more interfaces like most servers have. A static
IP address cannot be configured without the correct interface name. One way is to see the output of the command
nmcli d
in your terminal for quick identification of Ethernet cards installed in your machine. The output of this command can show which
interface is connected and which is not. You can try the command old ifconfig command: connected interface should get some packets while
disconnected do not.
The command ip a will show which interfaces are up and which are not. The other is to use tcpdump utility and see what packets each
interface gets
On small networks you select IP yourself. In large corporations there is usually some centralized IP allocation service, which
will provide you with IP that you should use for the particular server as well as subnet mask, default router and DNS server you should
use.
You need to check with ping if the Ip is already occupied in any case. That's a useful precaution
In you home network (192.168.0.0/24) router often has the address 192.168.0.1 or 192.168.0.254 (the
first or the last address in this subnet). Sometime a different subnet is used like 192.168.5.0/24. In this case router can be
192.168.5.1 or 192.168.5.254
This is where "Winfowization" of RHEL7 is full display. ip command is not available of Solaris or HP-UX or AIX.
While utility ip was available in RHEL6, RHEL 7 makes it the new standard way of configuring networking, although the old ifconfig
utility is deprecated but still available if you install the proper package (in CentOS7 it should be installed vi net-tools
package, which installs ifconfig, netstat, route, arp and several other commands; In RHEL7
I think it is installed by default).
The ip utility has different set of options/parameters in comparison with ifconfig. Which now need
to be memorized, as again this is "a new brave world" of RHEL7, quite different from the world of RHEL4-RHEL6. Some typical invocations:
- Use ip address to list configure and monitor network addresses. See ip address help for option
- to list interfaces (line ifconfig -a) use ip address (you can alias it ipconfig2 if you want)
- To assign static IP type ip addr add 10.0.0.10/24 dev <yourdevicename>.
- Use ip route to configure and monitor routing information. Similar to previous route command. See ip route help for option
- Use ip link to configure and monitor network link state. If produced the list of interfaces with MAC addresses.
Essentially cut down version of the output of the command ip adresss.
For Sysadmins who worked in RHEL6 this is usually pretty foreign utility and you need a cheatsheet to use it productively.
Among them:
So RHEL7 migrated away for earlier RHEL version in which the ifconfig utility was and is still used for validating network
configuration. and netstat for displaying routing table. They still can be used (and probably are preferred option
is you need to manage simultaneously both RHEL5, RHEL6 and RHEL7.) But as RHEL6 end of life is 2020 you also need to learn
new, unique for RHEL7 utilities, which represent another "new way of doing things", a new brave world, if you wish ;-)
TIP:
To show current network settings, you can use the ip addr show command (which can be abbreviated as ip a s or even
as ip a). The ip command does not always require you to type the complete option. The command shows the following
items about its current status:
- Current state: The most important part of this line is the text state UP, which shows that this network card is currently
up and available.
- MAC address configuration: This is the unique MAC address that is set for every network card. You can see the MAC address
itself (00:0c:29:b8:8c:eb), as well as the corresponding broadcast address.
- IPv4 configuration: This line shows the IP address that is currently set, as well as the subnet mask that is used. You
can also see the broadcast address that is used for this network configuration.
- IPv6 configuration: This line shows the current IPv6 address and its configuration.
If you are just interested in the link state of the network interfaces, you can use the ip link show command. This command
(of which you can see the output in Listing 8.2) repeats the link state information of the ip addr show command.
[0]d620@ROOT:~ # ip addr show
1: lo: mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: enp9s0: mtu 1500 qdisc mq state UP group default qlen 1000
link/ether 00:1c:23:06:cb:e7 brd ff:ff:ff:ff:ff:ff
inet 192.168.5.23/24 brd 192.168.5.255 scope global noprefixroute dynamic enp9s0
valid_lft 70434sec preferred_lft 70434sec
inet6 fe80::4dcf:908b:d93f:808d/64 scope link noprefixroute
valid_lft forever preferred_lft forever
3: wlp12s0: mtu 1500 qdisc noop state DOWN group default qlen 1000
link/ether ba:36:3f:be:e5:5f brd ff:ff:ff:ff:ff:ff
4: docker0: mtu 1500 qdisc noqueue state DOWN group default
link/ether 02:42:3e:20:c0:d0 brd ff:ff:ff:ff:ff:ff
inet 172.17.0.1/16 scope global docker0
valid_lft forever preferred_lft forever
In case the ip link show command the output looks like:
d620@ROOT:~ # ip link show
1: lo: mtu 65536 qdisc noqueue state UNKNOWN mode DEFAULT group default qlen 1000
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
2: enp9s0: mtu 1500 qdisc mq state UP mode DEFAULT group default qlen 1000
link/ether 00:1c:23:06:cb:e7 brd ff:ff:ff:ff:ff:ff
3: wlp12s0: mtu 1500 qdisc noop state DOWN mode DEFAULT group default qlen 1000
link/ether ba:36:3f:be:e5:5f brd ff:ff:ff:ff:ff:ff
4: docker0: mtu 1500 qdisc noqueue state DOWN mode DEFAULT group default
link/ether 02:42:3e:20:c0:d0 brd ff:ff:ff:ff:ff:ff
shows the current link state as down, you can temporarily bring it up again by using ip link set, which is followed by
dev devicename and up (for example, ip link set dev eno16777736 up).
We now know that ip addr add command allow us to set an IP address on a network interface. Bit settings
provided by the ip command are not persistent. If you want to make your configuration persistent in RHEL7 you need to
NetworkManager. It has two configuration utilities: nmtui or
nmcli.
Networking on RHEL 7 is managed by the NetworkManager service. You can use the systemctl status NetworkManager
command to verify its current status. When NetworkManager comes up, it reads the network card configuration scripts, which are in
/etc/sysconfig/network-scripts and have a name that starts with ifcfg and is followed by the name of the network card.
Two terms are used in describing the configuration of Ethernet interface:
- A device is a network interface card.
- A connection is the configuration that is used on a device.
You can create multiple connections for a single device. This can make sense on mobile computers, to make a difference between
settings that are used while connected to the home network and settings that are needed to the corporate network. Switching between
connections on devices is something that is common on end-user computers, and not so common on servers, except web servers.
To manage the network connections
you need to use iether the nmtui or the nmcli command.
Tip
- nmtui simpler to use. nmtui provides text based menues which are much easier to use and you will get things done a lot easier with this command.
- nmcli which requires command line parameters.
- The exact syntax of this command may be hard to remember. Fortunately, though, there is an excellent man page with examples.
Type man nmcli-examples to show this man page; you’ll notice that if you can find this man page, you can do almost
anything with nmcli.
- nmcli offers command line completion features. Just try it and type nmcli. Don’t press enter—but press
the Tab key twice—it will show all available options that nmcli expects at this moment. Choose an option, such as
connection, and press the Tab key twice. Using this approach helps you to comppose long commands without the need
to memorize anything!
If you assign ip address, netmask and the default route via ip command those settings will not survive the reboot. To make those
assignment permanent in RHL7 you typically use utility nmtui which provides user interface to Network Manager.
The nmtui interface consists of three menu options:
- Edit a Connection: Use this option to create new connections or edit existing connections. The option offers almost
all features that you might ever need to do while working on network connections. It sure allows you to do anything you need to be
doing on the RHCSA exam. You can use it to add any type of connection. Not just Ethernet connections, but also advanced connection
types such as network bridges and teamed network drivers are supported.
- Activate a Connection: Use this to (re)activate a connection.
- Set System Hostname: Use this to set the hostname of your computer.
If you have access to GUI you can use nm-connection-editor instead of nmtui. Not all setting are available via GUI but basic
settings are. It does, however, offer all you need to manage address configuration on a network connection. Start it by using the
nm-connection-editor command, or by using the applet in the GNOME graphical interface.
The result of setting parameter is your network interface(s) via mntui is the "per-interface" configuration file in the directory /etc/sysconfig/network-scripts. The
name of this generated configuration files has prefix ifcfg- which is followed by the name of the network interface.
[0]d620@ROOT:/etc/sysconfig/network-scripts # cat ifcfg-enp9s0
TYPE=Ethernet
PROXY_METHOD=none
BROWSER_ONLY=no
BOOTPROTO=dhcp
DEFROUTE=yes
IPV4_FAILURE_FATAL=yes
IPV6INIT=yes
IPV6_AUTOCONF=yes
IPV6_DEFROUTE=yes
IPV6_FAILURE_FATAL=no
IPV6_ADDR_GEN_MODE=stable-privacy
NAME=enp9s0
UUID=034c881a-b017-48d6-9398-9b02ee66e4e2
DEVICE=enp9s0
ONBOOT=no
USERS=root
TIPS:
- Normally, there should be no need to modify these configuration files manually. If you want to, though, you can. After making changes
to the configuration file, use the nmcli con reload command to activate the new configuration.
- You can disable Network manager for the particular interface and manage this interface directly via the network configuration
file (see below), but be careful and do not abuse this possibility without the necessity. In this sense RHEL7 is not
RHEL6 where disabling NetworkManager on servers was the standard practice.
This utility is used mainly in scripts.
nmcli without paramaters shows all connections. This shows active and inactive connections. You can easily
see the difference because inactive connections are not currently assigned to a device:
[0]d620@ROOT:~ # nmcli
enp9s0: connected to dhcp
"Broadcom and subsidiaries NetXtreme BCM5752"
ethernet (tg3), 00:1C:23:06:CB:E7, hw, mtu 1500
ip4 default
inet4 192.168.5.23/24
route4 0.0.0.0/0
route4 192.168.5.0/24
inet6 fe80::4dcf:908b:d93f:808d/64
route6 fe80::/64
route6 ff00::/8
docker0: connected to docker0
"docker0"
bridge, 02:42:3E:20:C0:D0, sw, mtu 1500
inet4 172.17.0.1/16
route4 172.17.0.0/16
wlp12s0: unavailable
"Intel PRO/Wireless 3945ABG"
wifi (iwl3945), BA:36:3F:BE:E5:5F, hw, mtu 1500
lo: unmanaged
"lo"
loopback (unknown), 00:00:00:00:00:00, sw, mtu 65536
DNS configuration:
servers: 4.2.2.3 4.2.2.2
domains: SSG5-ISDN-WLAN
interface: enp9s0
Use "nmcli device show" to get complete information about known devices and
"nmcli connection show" to get an overview on active connection profiles.
Consult nmcli(1) and nmcli-examples(5) manual pages for complete usage details.
After finding the name of the connection, you can use nmcli con show followed by the name of the connection to see all properties
of the connection. Notice that this... command shows many properties. Here is and example of the output of this command.
[127] d620@ROOT:~ # nmcli con show
NAME UUID TYPE DEVICE
dhcp ac522294-9aa2-4729-be22-f0f6ee6451af ethernet enp9s0
docker0 39461b65-4dc9-4e1e-8ec4-a8a1d2b6e725 bridge docker0
enp9s0 034c881a-b017-48d6-9398-9b02ee66e4e2 ethernet --
To find out what exactly these settings are doing, read man 5 nm-settings.
You can also use nmcli to show status of all devices
nmcli dev status
[0]d620@ROOT:~ # nmcli dev status
DEVICE TYPE STATE CONNECTION
enp9s0 ethernet connected dhcp
docker0 bridge connected docker0
wlp12s0 wifi unavailable --
As well as settings for the particular device, for example enp9s
[2] d620@ROOT:~ # nmcli dev show enp9s0
GENERAL.DEVICE:
enp9s0
GENERAL.TYPE:
ethernet
GENERAL.HWADDR:
00:1C:23:06:CB:E7
GENERAL.MTU:
1500
GENERAL.STATE:
100 (connected)
GENERAL.CONNECTION:
dhcp
GENERAL.CON-PATH:
/org/freedesktop/NetworkManager/ActiveConnection/1
WIRED-PROPERTIES.CARRIER: on
IP4.ADDRESS[1]:
192.168.5.23/24
IP4.GATEWAY:
192.168.5.1
IP4.ROUTE[1]:
dst = 0.0.0.0/0, nh = 192.168.5.1, mt = 100
IP4.ROUTE[2]:
dst = 192.168.5.0/24, nh = 0.0.0.0, mt = 100
IP4.DNS[1]:
4.2.2.3
IP4.DNS[2]:
4.2.2.2
IP4.DOMAIN[1]:
SSG5-ISDN-WLAN
IP6.ADDRESS[1]:
fe80::4dcf:908b:d93f:808d/64
IP6.GATEWAY:
--
IP6.ROUTE[1]:
dst = fe80::/64, nh = ::, mt = 100
IP6.ROUTE[2]:
dst = ff00::/8, nh = ::, mt = 256, table=255
TIPS:
You created network connections using nmcli con add. You can also change current connection properties
by using nmcli con mod. Here is an example of a simple script to configure a network interface:
# Make sure that the static connection does not connect automatically
nmcli con mod "static" connection.autoconnect no.
Add a DNS server to the static connection
nmcli con mod "static" ipv4.dns 10.0.0.10.
# To add a second item for the same parameters, use a + sign.
nmcli con mod "static" +ipv4.dns 8.8.8.8.
# Using nmcli con mod, you can also change parameters such as the existing IP address.
nmcli con mod "static" ipv4.addresses "10.0.0.20/24" 10.0.0.100/24.
# Add a second IP address you use the + sign again:
nmcli con mod "static" +ipv4.addresses 10.20.30.40/16.
# After changing connection properties, you need to activate them.
nmcli con up "static".
systemctl restart NetworkManager
See also
In RHEL 7 you usually use network manager to assign those addresses, and we will discuss it later. This is the preferable method,
but you also can do it directly by editing the file with settings in /etc/sysconfig/network-scripts directory.
One problem with the established method is that by default Network Manager overwrites /etc/resolv.conf file (the file
with DNS settings) and this is highly undesirable behaviour which you block only by editing specific configuration files.
See How do I
prevent Network Manager from controlling an interface CDRouter Support
Network Manager has a command line tool that can be used to see which interfaces it is controlling. Pull up a terminal window and
type the following command:nmcli dev status
This displays a table that lists all network interfaces along with their STATE. If Network Manager is not controlling
an interface, its STATE will be listed as unmanaged. Any other value indicates the interface is under Network
Manager control.
#Network Manager settings are controlled by a configuration file: #/etc/NetworkManager/NetworkManager.conf. We will
use this #file to tell Network Manager to stop controlling a particular interface.
ifcfg method
With CentOS, one way to tell Network Manager to stop controlling an interface is by editing the individual ifcfg-*
files. First, make sure the /etc/NetworkManager/NetworkManager.conf file has the following lines.
[main]
plugins=ifcfg-rh
This plugin tells Network Manager to look at the CentOS ifcfg-* files. Now for each interface you’d like Network
Manager to ignore, edit the individual /etc/sysconfig/network-scripts/ifcfg-* interface files, adding the following
lines:
NM_CONTROLLED=no
HWADDR=00:11:22:33:44:55
If present, set the HWADDR value to be the actual MAC address of this interface.
keyfile method
Another way to tell Network Manager to stop controlling an interface is by adding the following lines to the /etc/NetworkManager/NetworkManager.conf
file:
[main]
plugins=keyfile
[keyfile]
unmanaged-devices=mac:00:11:22:33:44:55;mac:66:77:88:99:00:aa
List the MAC address of each interface you want Network Manager to ignore, separated with a semicolon. Make sure that MAC addresses
listed here are LOWER CASE.
Newer versions of NetworkManager can also use a more generic alternate KEYFILE method that does not involve actual MAC addresses
and instead uses interface names. This syntax replaces the [keyfile] section above with the following:
[keyfile]
unmanaged-devices=interface-name:eth*,except:interface-name:eth0;interface-name:wlan*
udev rules method
A third way to tell Network Manager to stop controlling an interface is to add a udev rule. Create the file /etc/udev/rules.d/00-nta1000-net.rules
and add the following content:
#
# Interfaces that shouldn't be managed by NetworkManager
#
ACTION=="add", SUBSYSTEM=="net", KERNEL=="eth1", ENV{NM_UNMANAGED}="1"
ACTION=="add", SUBSYSTEM=="net", KERNEL=="eth2", ENV{NM_UNMANAGED}="1"
ACTION=="add", SUBSYSTEM=="net", KERNEL=="eth3", ENV{NM_UNMANAGED}="1"
You can list the interfaces (by name) that you want Network Manager to avoid.
You should reboot for these changes to take effect.
Manually stop Network Manager
If Network Manager is not needed during testing, it can be shut off completely.
systemctl stop NetworkManager
systemctl start NetworkManager
The above commands will stop/start NetworkManager, but will not affect whether it is automatically started when the machine boots.
To disable/enable NetworkManager automatically when the system boots, again use the systemctl command (as root):
systemctl disable NetworkManager
systemctl enable NetworkManager
Remove Network Manager from the system
If Network Manager will not be needed at all, it can be removed from the system completely.
yum remove NetworkManager
Restart Networking
You may need to restart networking to make changes take effect.
systemctl restart network
Within the directory /etc/sysconfig/network-scripts/ you should find the file ifcfg-INTERFACENAME (Where INTERFACENAME is
the name of your interface). In my instance, the file is ifcfg-enp9s03. The default configuration is DHCP configuration
and can look something lke:
[0]d620@ROOT:/etc/sysconfig/network-scripts # cat ifcfg-enp9s0
TYPE=Ethernet
PROXY_METHOD=none
BROWSER_ONLY=no
BOOTPROTO=dhcp
DEFROUTE=yes
IPV4_FAILURE_FATAL=yes
IPV6INIT=yes
IPV6_AUTOCONF=yes
IPV6_DEFROUTE=yes
IPV6_FAILURE_FATAL=no
IPV6_ADDR_GEN_MODE=stable-privacy
NAME=enp9s0
UUID=034c881a-b017-48d6-9398-9b02ee66e4e2
DEVICE=enp9s0
ONBOOT=no
USERS=root
We need to modify that file in order to not only change the protocol from dhcp to static, but to add the specific IP address. So
when you open up that file, you'll want to change:
BOOTPROTO=dhcp
To:
BOOTPROTO=static
Now you'll need to add the entries to set not only the IP address, but the netmask, gateway, and DNS addresses. At the bottom of
that file, add the following:
IPADDR=192.168.0.2
NETMASK=255.255.255.0
GATEWAY=192.168.0.254
NM_CONTROLLED=no
HWADDR=00:11:22:33:44:55
- Type link ip -s link. This shows all existing network connections, in addition to statistics about the number of packets
that have been sent and associated error messages.
- Verify IP adress type ip addr show. This shows the current network configuration. Note the IPv4 address
that is used. Notice the network device names that are used; you need these later in this exercise.
- Verify default gateway. Type ip route show to verify routing configuration.
- Verify that packets are forwarded correctly to other networks (or to Internet if you have no proxy). Use the ping command
.
- Type ip addr show. You’ll see the newly set IP address, in addition to the IP address that was already in use.
- Type ifconfig. Notice that you do not see the newly set IP address (and there are no options with the ifconfig
command that allow you to see it). This is one example of why you should not use the ifconfig command anymore.
- Type ss -tul. You’ll now see a list of all UDP and TCP ports that are listening on your server.
One important aspect of networking is routing. On every network that needs to communicate to nodes on other networks, routing is
a requirement. Every network has, at least, a default router (also called the default gateway) that is set, and you can see which router
is used as the default router by using the command ip route show Routers use network IDs to determine network traffic. Built
into this router is a routing table, the actual instructions that tell the router what to do with incoming packets and where
to send them.
You should always perform one quick check to verify that your router is set correctly: the default router at all times must be on
the same network as the local IP address that your network card is using. The first line shows the default router:
[130] d620@ROOT:~ # ip route show
default via 192.168.5.1 dev enp9s0 proto dhcp metric 100
172.17.0.0/16 dev docker0 proto kernel scope link src 172.17.0.1
192.168.5.0/24 dev enp9s0 proto kernel scope link src 192.168.5.23 metric 100
Network problems can be related to the local IP and router settings but can also be related to network ports that are not available
on your server or on a remote server. To verify availability of ports on your server, you can use the netstat command, or the
newer ss command, which provides the same functionality.
By typing ss -ls you will get summary of used ports
Total: 203 (kernel 264)
TCP: 5 (estab 1, closed 0, orphaned 0, synrecv 0, timewait 0/0), ports 0
Transport Total IP IPv6
* 264 - -
RAW 1 0 1
UDP 3 2 1
TCP 5 3 2
INET 9 5 4
FRAG 0 0 0
By typing ss -tul, you’ll see all listening TCP and UDP ports on the local system:
[0]d620@ROOT:~ #ss -tul
Netid State Recv-Q Send-Q Local Address:Port Peer Address:Port
udp UNCONN 0 0 *:bootpc *:*
udp UNCONN 0 0 127.0.0.1:323 *:*
udp UNCONN 0 0 ::1:323 :::*
tcp LISTEN 0 128 *:ssh *:*
tcp LISTEN 0 100 127.0.0.1:smtp *:*
tcp LISTEN 0 128 :::ssh :::*
tcp LISTEN 0 100 ::1:smtp :::* 192.168.5.14:58008
Notice where the port is listening on. Some ports are only listening on the IPv4 loopback address 127.0.0.1 or the IPv6 loopback
address ::1, which means that they are locally accessible only. Other ports are listening on *, which stands for all IPv4 addresses,
or on :::*, which represents all ports on all IPv6 addresses.
Softpanorama Recommended
How to configure a static IP
address in CentOS 7 - TechRepublic