linux_wiki:use_network_teaming_or_bonding_to_configure_aggregated_network_links_between_two_red_hat_enterprise_linux_systems

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

linux_wiki:use_network_teaming_or_bonding_to_configure_aggregated_network_links_between_two_red_hat_enterprise_linux_systems [2018/05/29 22:07]
billdozor
linux_wiki:use_network_teaming_or_bonding_to_configure_aggregated_network_links_between_two_red_hat_enterprise_linux_systems [2019/05/25 23:50]
Line 1: Line 1:
-====== Use Network Teaming Or Bonding To Configure Aggregated Network Links Between Two Red Hat Enterprise Linux Systems ====== 
- 
-**General Information** 
- 
-Network interface teaming is a new method of what RHEL 6 referred to as bonding. There are more available options in RHEL 7's network teaming. 
- 
----- 
- 
-====== Lab Setup ====== 
- 
-The following virtual machines will be used: 
-  * server1.example.com (192.168.1.150) -> Configure a team interface backed by to slave interfaces 
-  * server2.example.com (192.168.1.151) -> Configure a team interface backed by to slave interfaces 
- 
----- 
- 
-====== Help ====== 
- 
-Finding help in this section. 
-  * nmcli examples<code bash>man nmcli-examples</code> 
-  * Team 'runner' config examples<code bash>/usr/share/doc/teamd-1.17/example_configs/</code> 
- 
----- 
- 
-====== Team Creation ====== 
- 
-Creating a virtual team interface that is backed by two physical network interfaces. 
- 
-===== Virtualbox Notes ===== 
- 
-To each virtual machine, **add two network interfaces**.  
- 
-Each VM should end up with 3 total interfaces: 
-  * Main Management NIC: 192.168.1.x (server1=150 and server2=151) 
-  * Additional NIC for team-slave: Can be bridged or private 
-  * Additional NIC for team-slave: Can be bridged or private 
- 
-\\ 
-Due to the limitations of the virtualbox hypervisor, the virtual NICs must be configured to be in promiscuous mode.  
- 
-\\ 
-When NICs are added to a team, the MAC address of the first slave gets duplicated to the team interface and any other slave interfaces added. This confuses the underlying virtualbox hypervisor. 
- 
-\\ 
-In Virtualbox (Example from Virtualbox 5.2): 
-  * Select the EL7 virtual machine 
-  * Click Settings 
-  * Click Network 
-  * Under each Adapter tab, change the Promiscuous Mode drop down to "Allow All" 
-  * Click OK after changing all adapters 
- 
-In the EL7 OS 
-  * Once the VM is booted up, set the interfaces to promiscuous mode<code bash>ip link set enp0s8 promisc on 
-ip link set enp0s9 promisc on</code> 
- 
----- 
- 
-===== Packages ===== 
- 
-Install Teamd package (if not already installed) 
-<code bash> 
-yum install teamd 
-</code> 
- 
-\\ 
-Install Network Manager team plugin (if not already installed) 
-<code bash> 
-yum install NetworkManager-team 
-</code> 
- 
----- 
- 
-===== Interface Teaming ===== 
- 
-**On each VM**; Create a new team connection 
-<code bash> 
-nmcli con add type team ifname team0 config '{"runner": {"name": "loadbalance"}}' 
-</code> 
-  * Connection name is auto-created by combining type and ifname. (team-<ifname>) 
-    * If you really want to, you can define the con-name during creation 
-  * Other team "name" types that can be passed in the JSON for runner configuration (man teamd.conf): 
-    * broadcast -> transmit packets via all ports 
-    * roundrobin -> transmit in round robin 
-    * activebackup -> 1 active port, 1 on standby in case of failure 
-    * loadbalance -> balance between ports 
-    * lacp -> 802.3ad LACP protocol 
-  * **Example JSON configs** are available at: 
-    * /usr/share/doc/teamd-1.17/example_configs/ 
-    * man teamd.conf -> "EXAMPLES" section at the bottom 
- 
-\\ 
-Assign a static address to the virtual team0 interface 
-  * server1 team<code bash>nmcli con mod team-team0 ipv4.method manual ipv4.addresses 10.0.0.10/24</code> 
-  * server2 team<code bash>nmcli con mod team-team0 ipv4.method manual ipv4.addresses 10.0.0.20/24</code> 
- 
-\\ 
-**On each VM**; Add two **unconfigured** interfaces (enp0s8 and enp0s9) to the virtual team0 interface 
-<code bash> 
-nmcli con add type team-slave ifname enp0s8 master team0 
-nmcli con add type team-slave ifname enp0s9 master team0 
-</code> 
-  * Connection name is auto-created by combining type and ifname. (team-slave-<ifname>) 
-    * If you really want to, you can define the con-name during creation 
-  * If the interfaces have any previous configuration, you may get errors when attempting to set the connections to "up" later on. ("Error: Connection activation failed: Active connection removed before it was initialized") 
-    * Remove configurations: nmcli con delete <con-name> 
- 
-\\ 
-**On each VM**; Configure ports and team as up 
-<code bash> 
-nmcli con up team-slave-enp0s8 
-nmcli con up team-slave-enp0s9 
-nmcli con up team-team0 
-</code> 
- 
-\\ 
-View team status (by team device name) 
-<code bash> 
-teamdctl team0 state 
-</code> 
- 
-\\ 
-**Team Config Command Examples**<code bash>man nmcli-examples</code> 
- 
----- 
- 
-===== Interface Bonding ===== 
- 
-If you run into issues with teaming not working correctly, you can always set up a bonded connection in order to achieve the objective on the exam. 
- 
-This is straight from 'man nmcli-examples':<code bash>nmcli con add type bond ifname mybond0 mode active-backup 
-nmcli con add type ethernet ifname eth1 master mybond0 
-nmcli con add type ethernet ifname eth2 master mybond0</code> 
- 
----- 
  
  • linux_wiki/use_network_teaming_or_bonding_to_configure_aggregated_network_links_between_two_red_hat_enterprise_linux_systems.txt
  • Last modified: 2019/05/25 23:50
  • (external edit)