Configuring Advanced FabricPath Features
This chapter contains the following sections:
- Information About Advanced FabricPath Layer 2 IS-IS Configurations
- Setting Advanced FabricPath Layer 2 IS-IS Parameters
- Multi-Destination Tree
- Verifying the FabricPath Advanced Configurations
- Feature History for Configuring FabricPath Advanced Features
Note
For information about prerequisites, guidelines and limitations, and licensing requirements for FabricPath, see Chapter1, “Overview”
Information About Advanced FabricPath Layer 2 IS-IS Configurations
We recommend that you run the FabricPath network using the default Layer 2 IS-IS configurations.
However, you can change many of the IS-IS settings as follows:
- Globally on the entire switch and on each switch in the FabricPath network
- On specified FabricPath switches within the FabricPath network
If you change any of the FabricPath Layer 2 IS-IS settings, ensure that you make the same changes for those global parameters on every switch in the FabricPath network and for those parameters on every applicable FabricPath switch in the network.
Layer 2 IS-IS is based on Layer 3 IS-IS with enhancements to run on Layer 2. The commands for Layer 2 IS-IS and Layer 3 IS-IS are not the same. Layer 2 IS-IS is the control plane in FabricPath and a single protocol controls all unicast and multicast traffic. From a forwarding standpoint, FabricPath Layer 2 IS-IS forwards traffic for unicast, unknown unicast, broadcast, and multicast frames. Using Layer 2 IS-IS, the software maintains loop-free paths throughout the FabricPath network. (see “Configuring FabricPath Switching,” for information on default FabricPath Layer 2 IS-IS behavior and “Configuring FabricPath Forwarding,” for information on FabricPath forwarding.)
You can use these advanced FabricPath Layer 2 IS-IS configurations to fine-tune the operation of the FabricPath network.
Overload Bit
IS-IS uses the overload bit to tell other routers not to use the local router to forward traffic but to continue routing traffic destined for that local router.
You may want to use the overload bit in these situations:
- The router is in a critical condition.
- Graceful introduction and removal of the router to/from the network.
- Other (administrative or traffic engineering) reasons such as waiting for BGP convergence.
The Overload Bit feature is not backward compatible. For this feature to work properly, all FabricPath switches must support this feature.
You can configure the router to signal other routers not to use this router as an intermediate hop in their shortest path first (SPF) calculations. You can optionally configure the overload bit temporarily on startup, until BGP converges.
In addition to setting the overload bit, you might also want to suppress certain types of IP prefix advertisements from LSPs for Level 1 or Level 2 traffic.
Setting Advanced FabricPath Layer 2 IS-IS Parameters
Although the Layer 2 IS-IS protocol works automatically once you enable FabricPath, you can optionally configure parameters. Some FabricPath Layer 2 IS-IS parameters you configure globally and some you configure per switch. This section includes the following topics:
- Setting Advanced FabricPath Layer 2 IS-IS Parameters Globally
- Setting Advanced FabricPath Layer 2 IS-IS Parameters per Interface
- Clearing Advanced FabricPath Layer 2 IS-IS Counters
Setting Advanced FabricPath Layer 2 IS-IS Parameters Globally
Although the FabricPath Layer 2 IS-IS protocol works automatically once you enable FabricPath, you can optionally configure the global parameters.
Note
Graceful restart of IS-IS is not supported. While the graceful-restart command exists in the CLI, it is not supported at this time.
SUMMARY STEPS
3.
(Optional) authentication-check
4.
(Optional) authentication key-chain auth-key-chain-name
5.
(Optional) authentication type { cleartext | md5 }
6.
(Optional) log-adjacency-changes
7.
(Optional) lsp-gen-interval msecs [ msecs msecs ]
9.
(Optional) max-lsp-lifetime secs
10.
(Optional) maximum-paths max-paths
11.
(Optional) reference-bandwidth { ref-mbps [Mbps] | ref-gbps [Gbps] }
12.
(Optional) spf-interval msecs [ msecs msecs ]
13.
(Optional) topology topology_number
14.
(Optional) hostname dynamic
15.
(Optional) root-priority value
16.
(Optional) set-overload-bit {always | on-start-up { seconds }}
DETAILED STEPS
See the Cisco Nexus 6000 Series NX-OS Unicast Routing Configuration Guide, Release 6.0 for more information on IS-IS commands.
Setting Advanced FabricPath Layer 2 IS-IS Parameters per Interface
Although the FabricPath Layer 2 IS-IS protocol works automatically once you enable FabricPath, you can optionally configure the interface parameters.
SUMMARY STEPS
2.
interface { ethernet mod/slot | port-channel channel-number }
3.
(Optional) fabricpath isis authentication-check
4.
(Optional) fabricpath isis authentication key-chain auth-key-chain-name
5.
(Optional) fabricpath isis authentication type { cleartext | md5 }
6.
(Optional) fabricpath isis csnp-interval seconds
7.
(Optional) fabricpath isis hello-interval seconds
8.
(Optional) fabricpath isis hello-multiplier multiplier
9.
(Optional) fabricpath isis hello-padding
10.
(Optional) fabricpath isis lsp-interval milliseconds
11.
(Optional) fabricpath isis metric metric
12.
(Optional) fabricpath isis retransmit-interval seconds
13.
(Optional) fabricpath isis retransmit-throttle- interval milliseconds
DETAILED STEPS
For more information about IS-IS commands, see the Unicast Routing configuration guide and command reference for your platform.
Clearing Advanced FabricPath Layer 2 IS-IS Counters
SUMMARY STEPS
Step 1
(Optional) clear fabricpath isis adjacency [* | system-id | interface {ethernet mod/slot | port-channel channel-number }]
Step 2
(Optional) clear fabricpath isis statistics *
Step 3
(Optional) clear fabricpath isis traffic [* | interface {ethernet mod/slot | port-channel channel-number }]
DETAILED STEPS
Multi-Destination Tree
Information About MDT
The Multi-Destination Tree (MDT) feature removes the restriction of two MDTs per topology and permits a user to configure up to 16 MDTs per topology. This feature allows load-sharing of multicast traffic across different multi-destination trees within a topology, thus improving the network performance.
Guidelines and Limitations for MDT
An MDT, also referred to as a Forwarding Tag or ftag, is a spanning-tree used for forwarding packets within a topology. By default, a topology has two MDTs ftags: topology 0 has ftag 1 and 2, topology 1 has ftag 3 and 4, topology 2 has ftag 5 and 6.In a default configuration, the first ftag of a topology is used for broadcast and unknown unicast. The second ftag is used for multicast only. In addition, multicast traffic might use the first ftag for load-sharing.
Note
The 256 and 257 ftags are reserved ftags used for a CE VPC topology.
Configuration and usage of MDT on the Cisco Nexus device needs to follow the following rules:
- Minimum of Maximum MDTs: A contiguous fabricpath network has the maximum number of multi-destination trees (MDT) configured on every node in the network. Each node advertises the maximum number of trees it can support. The minimum value of the max trees across the nodes is used as the final operational count for MDT.
For example, if there are ten switches in a fabricpath network, and if nine switches are configured with 8 MDTs but the tenth switch is not configured with MDT, then by this rule, the tenth switch advertises its maximum supported trees as two. Each switch has two MDTs which is the minimum number of MDTs in the fabricpath network.
- Number of zero rooted trees with regards to the number of nodes: The number of multi-destination trees with zero root is equal to the number of MDTs configured minus the number of nodes in the fabricpath network.
For example, if there are three switches in the fabricpath network and each of these switches is configured with 8 MDTs, each switch would have (8 - 3 = 5) zero rooted trees. Only three trees are used with each node acting as the root for each tree.
- Maximum number of MDTs per topology: The maximum number of MDTs supported per topology is 16. If no MDT is configured on a topology, the default of two MDTs takes effect.
- Total number of MDTs across all topologies: The maximum number of MDTs supported functionally is 16 across all topologies. For example, if topology-0 has 8 trees, topology-1 is allowed to have a maximum of eight trees functionally.
Note
The configuration of 16 MDTs for 16 topologies passes through, However, functionally we plan to support maximum 16 MDTs across all topologies
- You can configure more than two MDTs for topologies 0 through 15.
- MDT Root Unreachable: Root re-calculation for the trees is triggered whenever any node in the setup becomes unreachable. This can also affect traffic in the network. For example, if there are four nodes with four MDTs configured with root priority as Node1>Node2>Node3>Node4 and Node 2 (Root for tree 2) reboots, root re-calculation is triggered as per the node priority. In this case, Node 3 becomes root for Tree2, Node 4 becomes root for Tree 3 and Tree 4 will be zero-rooted. Traffic for all the trees whose path went through Node 2 is affected temporarily in this case. The higher the priority of the disabled root, the more traffic is disrupted.
- Number of nodes is less than the number of MDTs: If the MDT configuration changes during run-time such that the number of MDTs configured is greater than the number of nodes in the network, the root-calculation of traffic is not affected. This adheres to the minimum of maximum rule.
For example, if there are four nodes and four MDT configured and later the admin configures eight MDTs on all four nodes, we see four MDTs and the root calculation remains the same.
- Mismatch in MDT Configuration in the beginning: If each switch in the fabricpath network has a different number configured as the number of MDTs, again the minimum of maximum rule applies and we see the least number of trees on all nodes.
For example, if there are three switches in the fabricpath network and switch 1 = 3 MDTs, switch 2 = 5 MDTs, switch 3 = 8 MDTs, the total number of MDTs seen on the setup is three.
Configuring MDT for Default Topology
You can configure MDT for a default topology.
|
|
|
|
|---|---|---|
switch(config-fabricpath-isis)# multi-destination trees tree-num |
Specifies the number of multi-destination trees for a topology. |
Configuring Per-Topology MDT
You can configure MDT per topology.
Verifying the FabricPath Advanced Configurations
To display FabricPath information for advanced configurations perform one of the following tasks:
Feature History for Configuring FabricPath Advanced Features
Table 5-1 lists the release history for these features.
|
|
|
|
|---|---|---|
Feedback