Assuming that Eth0,eth1,eth2,eth3 is bound to bond0, the steps are as follows:
(1) #vi/etc/modprobe.conf, add the following 2 lines: Alias bond0 bonding options Bonding miimon=100 MoD E=4
When mode=0, the corresponding port of the switch does not need trunk, but the Ping Gateway (192.*.*.*) from the node has a large delay. The solution to this problem is to change mode to 4, while configuring the switch corresponding port to turn on link aggregation and bundle as trunk.
In the modprobe.conf file,
The Miimon parameter is used to monitor the physical connection of the NIC and is recommended at least 100
The mode parameter is used to set the binding mode
0 Load balancing (Round robin strategy)
1 active/backup mode, only one NIC is used at the same time.
2 Load Balancing (XOR algorithm)
3 Broadcast Mode
4 802.3AD mode, LAGP protocol, double the bandwidth.
mode cannot be written wrong
Here is prone to unstable situation, most of the mode is wrong to choose, so in the implementation of the need to be careful, the mode is determined, and then in the hands.
(2) Modify the/etc/sysconfig/network-scripts/directory ifcfg-bond0, Ifcfg-eth0, Ifcfg-eth1,ifcfg-eth2,ifcfg-eth3. The specific contents are as follows:
[Email protected] network-scripts]# Pwd/etc/sysconfig/network-scripts[[email protected] network-scripts]# cat ifcfg-bond0device=bond0bootprote=staticonboot=yesipaddr=192.*.*.*netmask=255.255.255.0usertcl=nogateway=192.*. *.*
[email protected] network-scripts]# cat ifcfg-eth0device=eth0bootproto=noneonboot=yestype=ethernetusertcl= Nomaster=bond0slave=yes
[email protected] network-scripts]# cat ifcfg-eth1device=eth1bootproto=noneonboot=yestype=ethernetusertcl= Nomaster=bond0slave=yes
[email protected] network-scripts]# cat ifcfg-eth2device=eth2bootproto=noneonboot=yestype=ethernetusertcl= Nomaster=bond0slave=yes
[email protected] network-scripts]# cat ifcfg-eth3device=eth3bootproto=noneonboot=yestype=ethernetusertcl= Nomaster=bond0slave=yes
(3) View the status of Bond0
[[Email protected] ~]# more /proc/net/bonding/bond0ethernet channel bonding Driver: v2.6.3-rh (june 8, 2005) bonding mode: ieee 802.3ad dynamic link aggregationmii status: upmii polling interval (ms): 100Up Delay (ms): 0down delay (ms): 0802.3ad infolacp rate: slowactive Aggregator Info: Aggregator ID: 1 Number of ports: 4 actor key: 17 partner key: 480 partner mac address: 00:12:f2:cd : 68:00slave interface: eth0mii status: uplink failure count: 0permanent hw addr: 00:1e:0b:73:56:26Aggregator id: 1slave interface: eth1mii status: uplink failure count: 0permanent hw addr: 00:1e:0b:73:56:a2aggregator id: 1slave interface: eth2mii status: uplink failure count: 0permanent hw addr: 00:1e : 0b:d5:34:d2aggregator id: 1slave interface: eth3mii status: uplink failure count: 0permanent hw addr: 00:1e:0b:d5:34:d0aggregator id: 1
The trunk is then done on the switch (where the switch is foundry)
The use of bundling multiple network cards of the same server into a single logical line, increasing the bandwidth between the server and the switch, and being able to load balance traffic on multiple network cards, and redundancy between multiple network cards when one of the NICs is bad, the other NIC will split the traffic on the NIC. Does not affect data forwarding between the server and the switch. Note If the switch is configured as trunk then the server segment should also be configured with the appropriate trunk to be used correctly
[email protected]_SX_1600#show trunk Configured trunks:Trunk ID: 897hw trunk id: 6ports_configured: 4primary port monitored: jointlyports 15/1 15/2 15/3 15/4 port names none none none none port_ status enable enable enable enable monitor off off off off Rx Mirr Port N/A N/A N/A N/A tx mirr port n/a n/a n/a n/a monitor dir n/a n/a n/a n/a trunk id: 901Hw Trunk ID: 11Ports_Configured: 4Primary Port Monitored: jointlyports 15/5 15/6 15/7 15/8 Port Names none none none none Port_Status enable enable enable enable monitor off off off off Rx Mirr Port N/A N/A N/A N/A tx mirr port n/a n/a n/a N/A Monitor Dir N/A n/a n/a n/a
View the red callout information as above port status is enable this group of trunks includes 15/5 15/6 15/7 15/8 These 4 ports this group of trunks has been in effect
The trunk is then done on the switch (where the switch is foundry)
The use of bundling multiple network cards of the same server into a single logical line, increasing the bandwidth between the server and the switch, and being able to load balance traffic on multiple network cards, and redundancy between multiple network cards when one of the NICs is bad, the other NIC will split the traffic on the NIC. Does not affect data forwarding between the server and the switch. Note If the switch is configured as trunk then the server segment should also be configured with the appropriate trunk to be used correctly
[email protected]_sx_1600#show trunkconfigured trunks:trunk id: 897hw Trunk ID: 6Ports_Configured: 4Primary Port Monitored: JointlyPorts 15/1 15/2 15/3 15/4 Port Names none none none none Port_Status enable enable enable enable Monitor off off off off Rx Mirr Port N/A N/A n/a n/a tx mirr port n/a &nBsp; n/a n/a n/a monitor Dir N/A N/A N/A n/a trunk id: 901hw trunk id: 11ports_ configured: 4primary port monitored: jointlyports 15/5 15/6 15/7 15/8 Port Names none none none none port_status enable enable enable enable monitor off off&nbsP; off off rx mirr port N/A N/A N/A N/A Tx Mirr Port N/A N/A N/A N/A Monitor Dir N/A N/A N/A n/a
View the red callout information as above port status is enable this group of trunks includes 15/5 15/6 15/7 15/8 These 4 ports this group of trunks has been in effect
In the modprobe.conf file,
The Miimon parameter is used to monitor the physical connection of the NIC and is recommended at least 100
The mode parameter is used to set the binding mode
0 Load balancing (Round robin strategy)
1 active/backup mode, only one NIC is used at the same time.
2 Load Balancing (XOR algorithm)
3 Broadcast Mode
4 802.3AD mode, LAGP protocol, double the bandwidth.
mode cannot be written wrong
This article from "11726068" blog, declined reprint!
Linux under multiple NIC bindings