Brocade Multi-Service IronWare Switching Configuration Guide (Supporting R05.6.00) - Manual (Page 178)

148 Multi-Service IronWare Switching Configuration Guide
53-1003036-02
Deploying a LAG
6
If the packet is not IP:
• If the packet is MPLS, the hash is further calculated as the following:
• hash[5:0] = hash[5:0]^MPLS_L0[5:0]^MPLS_L1[5:0]^MPLS_L2[5:0];
• The 4-bit hash value is: hash[3:0] = hash[5:0]% num_of__members;
NOTE
The hashing in CES/CER is based on the values present in the header of the incoming traffic (L2,
L3, and L4). If the incoming traffic is equal to the total bandwidth of the outgoing LAG, there is no
guarantee that the packet will be equally shared among the LAG links.
Not supported:
• Speculate UDP or TCP headers
• Mask Layer-4 source
• Hash diversification
Deploying a LAG
After configuring a LAG, you must explicitly enable it using the deploy command before it begins
aggregating traffic. Once the deploy command is executed, the LAG enters aggregating mode. Only
the primary port within the LAG is available at the individual interface level. Any configuration
performed on the primary port applies to all ports within the LAG.
To deploy a LAG, at least one port must be in the LAG and the primary port must be specified for
non keepalive LAGs. Once a non keepalive LAG is deployed, a LAG is formed. If there is only one
port in the LAG, a single-port LAG is formed. For a dynamic LAG, LACP is started for each LAG port.
For a keepalive LAG, no LAG is formed and LACP is started on the LAG port.
Refer to the Multi-Service IronWare Administration Guide for additional information on LAG naming
conventions.
You can deploy a LAG as shown for the “blue” LAG.
Brocade(config)# lag blue static
Brocade(config-lag-blue)# deploy
Syntax: [no] deploy [forced | passive]
When the deploy command is executed:
• For a static and dynamic LAGs, the LAG veto mechanism is invoked to make sure the LAG can
be formed. If the LAG is not vetoed, a LAG is formed with all the ports in the LAG.
• For dynamic LAGs, LACP is activated on all LAG ports. When activating LACP, use active mode if
passive is not specified; otherwise, use passive mode.
• For keepalive LAGs, no LAG is formed, and LACP is started on the LAG port.
• Once the deploy command is issued, all LAG ports will behave like a single port.
• If the no deploy command is executed, the LAG is removed. For dynamic LAGs, LACP is
deactivated on all LAG ports.
• If the no deploy command is issued and more than 1 LAG port is not disabled, the command is
aborted and the following error message is displayed: “Error 2 or more ports in the LAG are not
disabled, un-deploy this LAG may form a loop - aborted.” Using the forced keyword with the no
deploy command in the previous situation, the LAG deployment is cancelled.