You configure bond interface properties at the network interface bond level.
To create a bond interface, use the following command:
Admin@nodename# create network interface bond
You need to specify the following parameters:
Parameter
|
Description
|
---|
enabled
|
Enable/disable the interface:
|
interface-name
|
Enter a number to include in the interface name (for example, if you enter 1 the interface name will be bond1).
|
description
|
Interface description.
|
alias
|
The interface's alias.
|
node-name
|
Cluster node where the bond interface is created.
|
zone
|
Zone to which the bond belongs.
|
link-info
|
Settings for network interface parameters:
-
bc_forwarding: control forwarding the directed broadcast packets arriving at the specified interface.
-
proxy_arp, proxy_arp_vlan: Proxy ARP mechanism. With proxy_arp, UserGate will respond to ARP requests for addresses outside the interface's network; with proxy_arp_vlan, UserGate will respond to ARP requests for addresses that belong to the interface's network.
To specify them, use the following format:
Admin@nodename# create network interface <iface-type> ... link-info [ key/value ]
where key is the parameter name. which can include lowercase Latin letters (a-z) and underscore (_), and
value is the parameter value. Parameter values can only be integers.
For example, use proxy_arp/1 to enable the Proxy ARP mechanism and proxy_arp/0 to disable it.
The link-info field is displayed only when adding parameters.
Important! You cannot delete the specified parameters.
|
netflow-profile
|
The Netflow profile to send statistical data to the Netflow collector. For more details on Netflow profile settings, see Configuring Netflow Profiles.
|
bonding
|
Additional bond interface parameters:
-
aggr-mode: bond operation mode. The available options:
-
round-robin: Round robin mode (packets are sent sequentially starting with the first available interface and ending with the last one. This policy is used to provide load balancing and high availability.)
-
active-backup: Active backup mode (only one network interface in the bond will be active. Another slave interface can become active only if the currently active interface fails. With this policy, the MAC address of the bond interface is only visible externally through one network port to avoid problems with the switch. This policy is used to provide high availability).
-
xor: XOR mode (the transmission is allocated among the NICs using the following formula: [( XOR ) MOD ]. This means that the same NIC sends packets to the same recipients. Optionally, the transmission allocation can also be based on the xmit_hash policy. The XOR policy is used for load balancing and high availability).
-
broadcast: Broadcast mode (broadcasts everything to all network interfaces. This policy is used for high availability).
-
802.3ad: IEEE 802.3ad mode (the default mode supported by most network switches. Creates aggregated groups of NICs with identical speed and duplex settings. When combined like this, all links in the active aggregation participate in transmission as per IEEE 802.3ad. The choice of interface for packet transmission is determined by the policy. By default, the XOR policy is used, with the xmit_hash policy as a possible alternative).
-
transmit: Adaptive transmit load balancing mode (outgoing traffic is distributed depending on the loading of each NIC (determined by the load speed). No additional configuration on the switch is required. The incoming traffic is received by the current network card. If this card fails, another card assumes the MAC address of the failed one).
-
load: Adaptive load balancing mode. Includes the previous policy plus incoming traffic balancing. No additional configuration on the switch is required. The incoming traffic is balanced through ARP negotiation. The driver intercepts ARP responses sent from the local NICs to the outside and overwrites the source MAC address with one of the unique MAC addresses of the NIC in the bond. Thus, different peers use different server MAC addresses. The incoming traffic is balanced sequentially (round-robin) among the interfaces.
-
mii-monitoring: MII monitoring period in milliseconds. Determines how often the link state will be checked for failures.
-
down-delay: delay time (in milliseconds) before an interface is disabled if a connection failure occurs. This option is only valid for MII monitoring (miimon). The parameter value must be a multiple of miimon,
-
up-delay: delay time in milliseconds before deploying the channel if it is detected to be restored. This parameter is only valid with MII monitoring (miimon). The parameter value must be a multiple of miimon,
-
lacp-rate: interval with which the partner transmits LACPDU packets in 802.3ad mode. Enumerated options:
-
failover-mac: define the assignment type of MAC addresses to bond interfaces in Active backup mode when switching interfaces. Enumerated options:
-
disabled: the same MAC address is set on all interfaces during switching.
-
active: the MAC address on the bond interface will always be identical to that on the currently active slave. The MAC addresses on the backup interfaces are not changed. The MAC address on the bond interface changes during the failover processing.
-
follow: the MAC address on the bond interface will be the same as that on the first slave added to the bond. This MAC is not set on the second and subsequent interfaces while they are in backup mode. That MAC address gets assigned during a failover: when a backup slave interface becomes active, it assumes a new MAC (the one on the bond interface), and the formerly active slave is assigned the MAC that the currently active one used to have.
-
xmit-hash: define a hash policy for sending packets over bond interfaces in XOR or IEEE 802.3ad mode. Enumerated options:
-
l2: use only MAC addresses to generate the hash. With this algorithm, the traffic for a particular network host is always sent over the same interface. This algorithm is compatible with IEEE 802.3ad.
-
l2-3: use both MAC and IP addresses to generate the hash. This algorithm is compatible with IEEE 802.3ad.
-
l3-4: uses IP addresses and transport layer protocols (TCP or UDP) to generate the hash. This algorithm is not universally compatible with IEEE 802.3ad, as both fragmented and non-fragmented packets can be transmitted within a single TCP or UDP interaction. Fragmented packets lack the source and destination ports. As a result, packets from the same session can reach the recipient in an order other than the intended one because they are sent via different slaves.
-
interface: interfaces to be bonded.
|
iface-mode
|
IP address assignment mode:
Static mode is set automatically when an IP address is assigned to the interface.
|
iface-type
|
The type of interface to be created:
|
ip-addresses
|
Assign an IP address to the interface.
The IP addresses are specified as [ <ip_address/mask> ] or [ <ip_address/mask> <ip_address/mask> ]. In case of several IP addresses (with space used as the separator), the subnet mask is entered in the decimal format.
Important! Make sure to separate the square brackets with spaces on both sides.
|
mac
|
Interface MAC address.
|
mtu
|
Specify the MTU size.
|
dhcp-relay
|
Settings for the DHCP relay on the interface. You need to specify the following:
-
enabled: enable/disable the relay:
-
utm-address: IP address of the UserGate interface on which the relay function is added.
-
server-address: addresses of DHCP servers where DHCP requests from clients should be redirected.
|
To update an existing bond interface, use the following command:
Admin@nodename# set network interface bond <bond-name>
The parameters available for setting are the same as those for creating a bond interface, except for interface-name and node-name (you cannot change the values of these parameters).
To delete a bond interface or its parameters, use the following command:
Admin@nodename# delete network interface bond <bond-name>
You can delete the following parameters:
Parameter
|
Description
|
---|
ip-addresses
|
Specified IP address.
|
dhcp-relay server-address
|
DHCP server IP address.
|
bonding interface
|
Bonded interfaces.
|
To display information about all bond interfaces, use the following command:
Admin@nodename# show network interface bond
To display information about a single interface, use the following command:
Admin@nodename# show network interface bond <bond-name>