Enable OVHcloud Link Aggregation in your SLES 15 server
Objective
OVHcloud Link Aggregation (OLA) technology is designed by our teams to increase your server's availability, and boost the efficiency of your network connections. In just a few clicks, you can aggregate your network cards and make your network links redundant. This means that if one link goes down, traffic is automatically redirected to another available link. The available bandwidth is also doubled thanks to aggregation.
Aggregation is based on IEEE 802.3ad, Link Aggregation Control Protocol (LACP) technology.
This guide explains how to bond your interfaces to use them for OLA in SLES 15.
Navigation path:Bare Metal Cloud > Dedicated servers > Select your server
Instructions
Because you have a private-private configuration for your NICs in OLA, you will be unable to SSH into the server. Thus, you will need to leverage the IPMI tool to access the server.
Click the IPMI tab (1).
Next, click the From a Java applet (KVM) button (2).
A JNLP program will be downloaded. Open the program to enter the IPMI. Log in using valid credentials for the server.
By default, using an OVHcloud template, the NICs will be named eth0 and eth1. If you are not using an OVHcloud template, you can find the names of your interfaces using the following command:
ip a
Info
This command will yield numerous "interfaces." If you are having trouble determining which ones are your physical NICs, the first interface will still have the server's public IP address attached to it by default.
Retrieving MAC addresses
Switch to the tab Network Interfaces and take note of the MAC addresses for each interface (public/private) which are displayed at the bottom of the menu.
Info
You can use any private IP address and subnet you wish.
If your server has more than 2 network interfaces, you can add them in the configuration, by incrementing the number of the BONDING_SLAVE_ parameter, for example, BONDING_SLAVE_2='eth2'.
Now that you know which MAC addresses are associated to each type (public/private) of interface, you need to retrieve the interface names.
Retrieving interface names
Warning
The IP addresses will be different for each server.
To retrieve the names of the interfaces, execute the following command:
ip a
Info
The hardware address (MAC address) of the NIC can be found using the ip a command that you used earlier. It will be the number next to link/ether in the output.
Here's an output example:
1: lo: <LOOPBACK,UP,LOWER_UP> 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 noprefixroute valid_lft forever preferred_lft forever2: ens22f0np0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000 link/ether a1:b2:c3:d4:e5:c6 brd ff:ff:ff:ff:ff:ff inet 203.0.113.1/32 metric 100 scope global dynamic ens22f0np0 valid_lft 71613sec preferred_lft 71613sec inet6 2001:db8:1:1b00:203:0:112:0/56 scope global valid_lft forever preferred_lft forever inet6 fe80::a6b2:c3ff:fed4:e5c6/64 scope link valid_lft forever preferred_lft forever3: ens22f1np1: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN group default qlen 1000 link/ether a1:b2:c3:d4:e5:c7 brd ff:ff:ff:ff:ff:ff4: ens33f0np0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN group default qlen 1000 link/ether a1:b2:c3:d4:e5:d6 brd ff:ff:ff:ff:ff:ff5: ens33f1np1: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN group default qlen 1000 link/ether a1:b2:c3:d4:e5:d7 brd ff:ff:ff:ff:ff:ff
Once you have determined the names of your interfaces, you can configure interface bonding in the OS.
Configuring interface bonding
Select the tab below that matches your server configuration:
Two interfaces: Advance servers with two physical NICs.
Four interfaces - Double LAG: Scale and High-Grade servers with OLA in Active - Double LAG mode (public + private aggregates). This requires OLA to be enabled in the OVHcloud Control Panel.
Four interfaces - Fully Private: Scale and High-Grade servers with OLA in Active - Fully Private mode (single private aggregate for vRack). This requires OLA to be enabled in the OVHcloud Control Panel.
Two interfaces
Four interfaces - Double LAG
Four interfaces - Fully Private
Create the bond configuration file /etc/sysconfig/network/ifcfg-bond0:
The private bond (ifcfg-bond1) and all physical interface configuration files remain the same as above.
This configuration aggregates all physical interfaces into a single bond for vRack use only. There is no public IP connectivity.
Warning
Following the implementation of OLA in Fully Private mode, the public IP is no longer accessible. Make sure you have an alternative means of access (e.g. through another server in the vRack, or via KVM/IPMI) before applying this configuration.
Create the bond configuration file /etc/sysconfig/network/ifcfg-bond0:
In Fully Private mode, the bond uses the MAC address of the main private interface. The IPADDR field should be set to your vRack private IP.
Applying the configuration
Apply the configuration by reloading all interfaces with wicked:
wicked ifreload all
This may take several seconds since it is building the bond interface. To test that the bond is working, ping another server on the same vRack. If it works, you are all set. If it does not, double-check your configurations or try rebooting the server.
You can also verify the bonding parameters using the following command:
