How to Configure Network Teaming in CentOS/RHEL 7?

In earlier versions of RHEL, network bonding was the default method for creating aggregated network interfaces. In RHEL 7, network teaming has been added as a solution. The main difference between these two is that network bonding happened completely in user space, whereas, in network teaming, the teamd daemon is added to allow interaction in user space as well. Even if both methods are still valid, network teaming is the preferred method.

On earlier versions of Red Hat Enterprise Linux, network bonding was used to accomplish the same goals. Network teaming is new in Red Hat Enterprise Linux 7. The solution consists of a small kernel driver and a daemon that is available in userspace: teamd.

# rpm -qa | grep teamd
teamd-1.27-4.el7.x86_64

The kernel takes care of handling network packets, while the teamd driver handles logic and interface processing. To determine how exactly this is happening, different runners are used. Runners in teaming are equivalent to the bonding modes. They are used to define the logic of traffic handling between the interfaces that are involved in the configuration. The table below gives a summary of available runners.

1. Configure Network Team Using JSON-format definitions [Not Persistent]

1. Create a JSON-format definition file for the team and its component ports，in this example, /root/team_config/team0.conf.

# cat /root/team_config/team0.conf
#team0.conf
{
   "device": "team0",
   "runner": {"name": "activebackup"},
   "link_watch": {"name": "ethtool"},
   "ports": {
      "enp0s8": {
         "prio": -10,
         "sticky": true
       },
       "enp0s9": {
         "prio": 100
       }
   }
}

2. Use the ip command to bring down the component ports:

# ip link set enp0s8 down
# ip link set enp0s9 down

Note: Active interfaces cannot be added to a team.

3. Start an instance of the teamd daemon and have it create the teamed interface by reading the configuration file (in this example, /root/team_config/team0.conf):

# teamd -g -f /root/team_config/team0.conf -d
Using team device "team0".
Using PID file "/var/run/teamd/team0.pid"
Using config file "/root/team_config/team0.conf"

Note: The -g option displays debugging messages and can be omitted.

4. Use the ip command to set the IPaddr and network mask prefix length of the teamed interface:

# ip addr add 192.168.22.10/24 dev team0

5. Verify the configuration wit “ip addr” and “teamdctl” commands.

# ip addr
team0: <broadcast,multicast,up,lower_up style="box-sizing: inherit;"> mtu 1500 qdisc noqueue state UP qlen 1000
link/ether 08:00:27:5a:d8:50 brd ff:ff:ff:ff:ff:ff
inet 192.168.22.10/24 scope global team0
valid_lft forever preferred_lft forever
inet6 fe80::a00:27ff:fe5a:d850/64 scope link
valid_lft forever preferred_lft forever</broadcast,multicast,up,lower_up>

# teamdctl team0 state
setup:
runner: activebackup
ports:
enp0s8
link watches:
link summary: up
instance[link_watch_0]:
name: ethtool
link: up
down count: 0
enp0s9
link watches:
link summary: up
instance[link_watch_0]:
name: ethtool
link: up
down count: 0
runner:
active port: enp0s9

2. Configure Network Team Using ifcfg Files [Persistent]

/etc/sysconfig/network-scripts directory and create “ifcfg-team0” file as shown below:

# cat /etc/sysconfig/network-scripts/ifcfg-team0
DEVICE=team0
DEVICETYPE=Team
ONBOOT=yes
BOOTPROTO=dhcp
NM_CONTROLLED=no
#IPADDR=192.168.11.1
PREFIX=24
TEAM_CONFIG='{"runner": {"name": "activebackup"}, "link_watch": {"name": "ethtool"}}'

Note:

The teaming interface team0 can also have static ip with BOOTPROTO=none

2. Edit the files for respective interface (here I have added two interface enp0s8 & enp0s9):

# cat /etc/sysconfig/network-scripts/ifcfg-enp0s8
DEVICE=enp0s8
#HWADDR=D4:85:64:01:46:9E
DEVICETYPE=TeamPort
ONBOOT=yes
NM_CONTROLLED=no
TEAM_MASTER=team0
TEAM_PORT_CONFIG='{"prio": 100}'

# cat /etc/sysconfig/network-scripts/ifcfg-enp0s9
DEVICE=enp0s9
#HWADDR=D4:85:64:01:46:9E
DEVICETYPE=TeamPort
ONBOOT=yes
TEAM_MASTER=team0
NM_CONTROLLED=no
TEAM_PORT_CONFIG='{"prio": 100}'

3. Make sure both the interface are down:

# ip link set enp0s8 down
# ip link set enp0s9 down

Note: Active interfaces cannot be added to a team.

4. Now you can bring up your interface

# ifup team0

5. Verify the teaming configuration with “ip addr”, “nmcli device status” and “teamdctl” commands.

# ip addr
team0: <broadcast,multicast,up,lower_up style="box-sizing: inherit;"> mtu 1500 qdisc noqueue state UP qlen 1000
link/ether 08:00:27:2e:4c:65 brd ff:ff:ff:ff:ff:ff
inet 192.168.57.101/24 brd 192.168.57.255 scope global dynamic team0
valid_lft 1178sec preferred_lft 1178sec
inet6 fe80::a00:27ff:fe2e:4c65/64 scope link tentative dadfailed
valid_lft forever preferred_lft forever</broadcast,multicast,up,lower_up>

# nmcli device status
DEVICE TYPE STATE CONNECTION
enp0s8 ethernet unmanaged --
enp0s9 ethernet unmanaged --
team0 team unmanaged --

# teamdctl team0 state
setup:
runner: activebackup
ports:
enp0s8
link watches:
link summary: up
instance[link_watch_0]:
name: ethtool
link: up
down count: 0
enp0s9
link watches:
link summary: up
instance[link_watch_0]:
name: ethtool
link: up
down count: 0
runner:
active port: enp0s8

Removing a Team

To terminate, or kill, an instance of the team daemon use the below command:

# teamd -t team0 -k