Configuring Link-Group Using the CLI
This section describes how to configure the link-group function using the CLI.
Context
The link group function binds the status of several interfaces to form a logical group. If one interface in the logical group is faulty, the system changes the status of the other interfaces to Down. After all the interfaces recover, the system changes the status of the interfaces to Up. The link group function ensures that the status of the upstream and downstream interfaces are consistent with each other, avoiding the inconsistency of upstream and downstream paths upon active/standby switchover.
The link group function supports interface status management in the same virtual system and interface status management of the root system and different virtual systems.
After an Eth-Trunk interface is added to a link group, the system considers the Eth-Trunk interface to be faulty only when all Eth-Trunk member interfaces go Down.
If some interfaces are added to both the Eth-Trunk and Link-group and you need to delete one of them from the Eth-Trunk, you must run the undo link-group command on the interface to be deleted and then delete the interface from the Eth-Trunk. Otherwise, services will be impacted.
To associate the interface status of different link-groups, configure a link-group monitor group. When configuring a link-group monitor group, properly plan the relationship between sub-groups and link-groups to achieve the association of interface status in different link-groups. The relationships between the link-group monitor group, sub-groups in the link-group monitor group, link-groups, and interfaces in the link-groups are as follows:
- A link-group can be added to only one sub-group of a link-group monitor group.
- Link-groups in a sub-group of a link-group monitor group are independent of each other and do not affect each other.
- Different sub-groups in a link-group monitor group affect each other only in the following situation: The device forcibly changes all interfaces in link-groups of other sub-groups to Down only when the state of all link-groups in a sub-group is fault.
- Different link-group monitor groups are independent of each other and do not affect each other.
Procedure
- Access the system view.
- Access the interface view.
interface interface-type interface-number
- Add the interface to the link-group.
link-group [ public ] link-group-id
By default, the system is not configured with the link-group management group.
If the status of the interface belonging to a virtual system needs to be associated with the interface status in the root system, you need to configure public. This parameter can be configured only in the root system and cannot be configured in a virtual system.
- Optional: Bound a link-group to the CPU of SPU.
link-group link-group-id binding spu-cpu-limit [ limit-number ]
After the link-group is bound to the CPU of SPU, if the CPU faults occur, the device is triggered to check the number of current CPU. If the number of current CPU is smaller than limit-number, the device turns the status of all valid interfaces in the link-group to Down.
Only the USG6635E/6655E, USG6680E and USG6712E/6716E support this command.
- Optional: Set an interface status monitoring timer for a link-group.
link-group link-group-id timer timer
By default, The value of the interface status monitoring timer is 10 seconds.
After a faulty interface in the link-group recovers, the system sets the state of other interfaces in the link-group to Up and checks whether all interfaces in the link-group are Up after twice of the interface status monitoring timer. If any interface is still Down, the system sets the state of other interfaces in the link-group to Down. Therefore, before configuring a link-group, check whether the delayed Up function is configured on the local interface to be added to the link-group and the peer interface. If yes, ensure that the interface status monitoring timer is longer than half of the interface Up delay to prevent frequent interface flapping, which interrupts services.
- Optional: Configure a link-group monitor group to associate the status of interfaces in different link-groups.
Example
# Add the GigabitEthernet0/0/1, GigabitEthernet0/0/2, and GigabitEthernet0/0/3 that belong to the root system, vsysa, and vsysb respectively to link-group 1.
<FW> system-view [FW] interface GigabitEthernet 0/0/1 [FW-GigabitEthernet0/0/1] link-group 1 [FW-GigabitEthernet0/0/1] quit [FW] interface GigabitEthernet 0/0/2 [FW-GigabitEthernet0/0/2] link-group public 1 [FW-GigabitEthernet0/0/2] quit [FW] interface GigabitEthernet 0/0/3 [FW-GigabitEthernet0/0/3] link-group public 1 [FW-GigabitEthernet0/0/3] quit
- If Eth-Trunk 1 fails, Eth-Trunk 2, Eth-Trunk 3, and Eth-Trunk 4 go Down as well.
- If Eth-Trunk 2 or Eth-Trunk 4 fails, the other interface goes Down, but Eth-Trunk 1 and Eth-Trunk 3 are not affected.
- If Eth-Trunk 3 or Eth-Trunk 4 fails, Eth-Trunk 1 is not affected. If both Eth-Trunk 3 and Eth-Trunk 4 fail, Eth-Trunk 1 goes Down.
To meet this requirement, add Eth-Trunk 1 and Eth-Trunk 3 to different link-groups (link-group 1 and link-group 3 in this example). Eth-Trunk 2 and Eth-Trunk 4 are added to the same link-group (link-group 2 in this example) different from Eth-Trunk 1 and Eth-Trunk 3. Create a link-group monitor group, create sub-group 1 and add link-group 1, and create sub-group 2 and add link-group 2 and link-group 3.
<FW> system-view [FW] interface Eth-Trunk 1 [FW-Eth-Trunk1] link-group 1 [FW-Eth-Trunk1] quit [FW] interface Eth-Trunk 2 [FW-Eth-Trunk2] link-group 2 [FW-Eth-Trunk2] quit [FW] interface Eth-Trunk 4 [FW-Eth-Trunk4] link-group 2 [FW-Eth-Trunk4] quit [FW] interface Eth-Trunk 4 [FW-Eth-Trunk4] link-group 2 [FW-Eth-Trunk4] quit [FW] interface Eth-Trunk 3 [FW-Eth-Trunk3] link-group 3 [FW-Eth-Trunk3] quit [FW] link-group-monitor name test [FW-link-group-monitor-test] monitor 1 link-group 1 [FW-link-group-monitor-test] monitor 2 link-group 2 [FW-link-group-monitor-test] monitor 2 link-group 3
Follow-up Procedure
Run the display link-group link-group-id command to check the configuration of the link-group and the status of each member interface.
<FW> display link-group 1 link group 1, total 3, fault 0 ------------------------------------------------------------------------------- Interface Vsys State ------------------------------------------------------------------------------- GigabitEthernet0/0/1 public up GigabitEthernet0/0/2 vsysa up GigabitEthernet0/0/3 vsysb up -------------------------------------------------------------------------------
Run the display link-group-monitor [ name link-group-monitor-name ] command to check the configuration and member status of the link-group monitor group.
<FW> display link-group-monitor name test link-group-monitor name test ----------------------------------------------------------- Monitor Link-group State ----------------------------------------------------------- 1 1 up ----------------------------------------------------------- 2 2 up 3 fault -----------------------------------------------------------