Configuring a vLAG
If you want to connect the two cluster nodes to an upstream switch or to a downstream host in a redundant fashion, you can configure a vLAG between the uplinks or the downlinks on the cluster nodes.
For example, if switch1 has port 53 connected to an upstream switch and switch2 has port 19 connected to the same upstream switch, create a vLAG by executing the vlag-create command on either of the cluster switches (switch1 in this example):
CLI (network-admin@switch1) > vlag-create name uplink port 53 [peer-switch switch2] peer-port 19
To verify the configuration, use the following command:
CLI (network-admin@switch1) > vlag-show
name cluster mode switch port peer-switch peer-port status local-state lacp-mode
------ --------- ------------- ------- ---- ----------- --------- ------ ----------- ---------
uplink cluster1 active-active switch1 53 switch2 19 normal enabled active
Note: Before you can create a vLAG, you must configure the two switches in a cluster.
Starting from release 7.0.0, NetVisor OS supports an enhancement for active-standby mode, for example when a single main vLAG link needs to be active while the backup link is required to become (and remain) active only when the main one fails.
You can enable this mode with the command:
CLI (network-admin@switch1) > vlag-create name <name> port <id> peer-switch <peer-name> peer-port <id> mode active-standby
Note: lacp-mode must be set to active for mode active-standby to work. lacp-mode off is not supported with mode active-standby.
In this release, in active-standby mode the active port election would happen through LACP message exchanges and would be based on the first port to become operationally active (i.e., in up state). Consequently, the other port would become the standby port and would be in phy-up,vlag-blocked state, as shown in the example below:
CLI (network-admin@switch1) > port-show port 121
port bezel-port status config
---- ---------- ------------------------------------------ ------
121 31 up,host,LLDP,LACP-PDUs,vlag-active,vlan-up fd,10g
CLI (network-admin@switch2) > port-show port 122
port bezel-port status config
---- ---------- ---------------------------------- ------
122 31.2 phy-up,LLDP,vlag-blocked,LACP-wait 10g
CLI (network-admin@switch1) > vlag-show
name cluster mode switch port peer-switch peer-port status local-state peer-state lacp-mode
----- ------- -------------- ------- ---- ----------- --------- ------ ----------- ---------- ---------
vlag1 cluster active-standby switch1 121 switch2 122 normal enabled,up active
Note: The active-standby mode and release 7.0.0’s enhancements work with both traditional clusters and clusters over Layer 3.
Note: With active-standby mode, for Layer 3 connectivity and redundancy, a single VTEP VIP (VRRP VIP) would still be used along with active-active L3 forwarding. Due to the asymmetric nature of active-standby vLAG mode, upstream traffic arriving on the cluster node with the standby vLAG port will have to take an extra physical (or virtual) hop (i.e., it will have to traverse the cluster link, or a VXLAN tunnel in case of a cluster over Layer 3) to reach its destination over an asymmetric forwarding path.
In release 7.0.0 VXLAN bridge domains are supported on vLAG members configured in mode active-standby. It is recommended that the bridge domains be configured before configuring the vLAG(s) to carry them.