Supported Platforms
Example: Configuring Automatic VLAN Administration Using MVRP
As the numbers of servers and VLANs attached to a QFabric systems increase, VLAN administration becomes complex and the task of efficiently configuring VLANs on multiple redundant server Node group devices becomes increasingly difficult. To partially automate VLAN administration, you can enable Multiple VLAN Registration Protocol (MVRP) on your QFabric system. If your QFabric system connects to servers that host many virtual machines that require their own VLANs, using MVRP can save you the time and effort that would be required to manually configure and administer the VLANs on the interfaces that connect to the servers. For example, if a virtual machine moves between servers—and therefore connects to a different redundant server Node group interface—MVRP can configure the appropriate VLAN membership on the new server Node group interface.
![]() | Note: Only trunk interfaces can be enabled for MVRP. |
This example describes how to configure MVRP on a QFabric system.
Requirements
This example uses the following hardware and software components:
- One QFabric system
- Junos OS Release 13.1 for the QFX Series
Overview and Topology
MVRP ensures that the VLAN membership information on the trunk interface is updated as the switch’s access interfaces become active or inactive in the configured VLANs in a static or dynamic VLAN creation setup.
You do not need to explicitly bind a VLAN to the trunk interface. When MVRP is enabled, the trunk interface advertises all the VLANs that are active (bound to access interfaces) on that switch. An MVRP-enabled trunk interface does not advertise VLANs that have been configured on the switch but that are not currently bound to an access interface. Thus, MVRP provides the benefit of reducing network overhead—by limiting the scope of broadcast, unknown unicast, and multicast (BUM) traffic to interested devices only.
When VLAN access interfaces become active or inactive, MVRP ensures that the updated information is advertised on the trunk interface. Thus, in this example, distribution Switch C does not forward traffic to inactive VLANs.
A redundant server Node group device is connected to a server that hosts virtual machines for three customers, each of which requires its own VLAN.
- customer-1: VLAN ID 100
- customer-2: VLAN ID 200
- customer-3: VLAN ID 300
Table 1 explains the components of the example topology.
Table 1: Components of the Example Topology
Settings | Settings |
---|---|
Hardware |
|
VLAN names and IDs |
|
Interfaces | Redundant server Node group device interfaces:
Network Node group device interface:
|
Configuring VLANs and Network Node Group Interfaces
To configure VLANs, bind the VLANs to the server-facing trunk interface, and enable MVRP on the trunk interface of the network Node group device, perform these tasks:
CLI Quick Configuration
To quickly configure VLANs on the QFabric system, assign VLAN membership to the uplink port on the network Node group device, and configure the uplink port to be trunk:
[edit]
set vlans customer-1 vlan-id
100
set vlans customer-2 vlan-id 200
set vlans customer-3 vlan-id 300
set interfaces NNG:xe-0/0/1 unit 0 family ethernet-switching
port-mode trunk
set interfaces NNG:xe-0/0/1 unit 0 family ethernet-switching
vlan members [customer-1 customer-2 customer-3]
![]() | Note: As recommended as a best practice, default MVRP timers are used in this example, so they are not configured. The default values associated with each MVRP timer are: 200 ms for the join timer, 1000 ms for the leave timer, and 10000 ms for the leaveall timer. Modifying timers to inappropriate values might cause an imbalance in the operation of MVRP. |
Step-by-Step Procedure
To create the VLANs and configure the network Node group device for MVRP, follow these steps. Note that you are creating VLANs for the entire QFabric system, so you do not need to create them on specific QFabric devices.
- Configure the VLAN for customer 1:
[edit]
user@qfabric# set vlans customer-1 vlan-id 100 - Configure the VLAN for customer 2:
[edit]
user@qfabric# set vlans customer-2 vlan–id 200 - Configure the VLAN for customer 3:
[edit]
user@qfabric# set vlans customer-3 vlan–id 300 - Configure an uplink interface (one that connects to an
interconnect device) to be a trunk:
[edit]
user@qfabric# set interfaces NNG:xe-0/0/1 unit 0 family ethernet-switching port-mode trunk - Configure the uplink interface to be a member of all three
VLANs:
[edit]
user@qfabric# set interfaces NNG:xe-0/0/1 unit 1 family ethernet-switching vlan members [customer-1 customer-2 customer-3]Note: If you want the uplink interface to be a member of all the VLANs in the QFabric system, you can enter all instead of specifying the individual VLANs.
Results
Check the results of the configuration on the network Node group device:
user@qfabric# show interfaces NNG:xe-0/0/1.0
user@qfabric# show vlans
Configuring the Redundant Server Node Group
CLI Quick Configuration
To quickly configure the redundant server Node group device for MVRP:
[edit]
set interfaces RSNG:xe-0/1/1 unit 0 family ethernet-switching
port-mode trunk
set interfaces
RSNG:xe-0/1/1 unit 0 family ethernet-switching vlan members [customer-1
customer-2 customer-3]
set
interfaces RSNG:xe-0/0/1 unit 0 family ethernet-switching port-mode
trunk
set protocols mvrp
interface RSNG:xe-0/0/1.0 passive
Step-by-Step Procedure
To configure the redundant server Node group device, follow these steps. Note that you do not need to configure the VLANs on the server-facing interface (RSNG:xe-0/0/1), but you do need to configure the VLANs on the uplink interface. Also notice that in this example you configure the server-facing interface to be passive, which means that it will not announce its membership in a VLAN or send any VLAN declarations (updates) unless it receives registration for that VLAN from the server.
- Configure an uplink interface (one that connects to the
interconnect device) to be a trunk:
[edit]
user@qfabric# set interfaces RSNG:xe-0/1/1 unit 0 family ethernet-switching port-mode trunk - Configure the uplink interface to be a member of all three
VLANs:
[edit]
user@qfabric# set interfaces NNG:xe-0/1/1 unit 0 family ethernet-switching vlan members [customer-1 customer-2 customer-3] - Configure an interface that connects to the server that
hosts multiple virtual machines to be a trunk:
[edit]
user@qfabric# set interfaces RSNG:xe-0/0/1 unit 0 family ethernet-switching port-mode trunk - Enable MVRP on the server-facing trunk interface and configure
it to be passive:
[edit]
user@qfabric# set protocols mvrp interface RSNG:xe-0/0/1.0 passive
Results
Check the results of the configuration for the redundant server Node group:
user@qfabric# show interfaces RSNG:xe-0/0/1.0
user@qfabric# show interfaces RSNG:xe-0/1/1.0
user@qfabric# show protocols mvrp interface RSNG:xe-0/0/1.0;
Verification
To confirm that the configuration is updating VLAN membership, perform these tasks:
Verifying That MVRP Is Enabled On The QFabric System
Purpose
Verify that MVRP is enabled on the appropriate interfaces
Action
Show the MVRP configuration:
user@qfabric> show mvrp
MVRP configuration MVRP status : Enabled MVRP timers (ms): Interface Join Leave LeaveAll -------------- ----- -------- ----------- NNG:xe-0/0/1.0 200 1000 10000 RSNG:xe-0/0/1.0 200 1000 10000 RSNG:xe-0/1/1.0 200 1000 10000 Interface Status Registration Mode -------------- -------- ----------------- NNG:xe-0/0/1.0 Enabled Normal
RSNG:xe-0/1/1.0 Enabled Normal
RSNG:xe-0/0/1.0 Enabled Passive
Meaning
The results show that MVRP is enabled on the appropriate network Node group and redundant server Node group interfaces and that the default timers are used.