Help us improve your experience.

Let us know what you think.

Do you have time for a two-minute survey?

Guide That Contains This Content
[+] Expand All
[-] Collapse All

    Implementing MC-LAG Active/Active with VRRP

    To allow all the links to forward traffic without being blocked by spanning-tree, multi-chassis link aggregation (MC-LAG) is configured on the edge routers and core switches. The edge routers use MC-LAG toward the edge firewalls, and the core switches use MC-LAG toward each POD switch, application load balancer (F5), and OOB management switch. MC- is a feature that supports aggregated Ethernet (AE) LAG bundles spread across more than one device. LACP is used for dynamic configuration and monitoring on links.

    Summary of Implementation Details for MC-LAG Active/Active

    MC-LAG is a key component of the MetaFabric 1.0 solution architecture. MC-LAG is configured using the following design considerations:

    • Do not mix Layer 2 next generation CLI syntax (L2NG, or family ethernet-switching) and non-l2ng (flexible-ethernet-serivces) syntax on the same interface.
    • Mac-learning is disabled on inter-chassis link (ICL).
    • Arp learning is disabled on ICL.
    • Static arp is required for integrated routing and bridging (IRB)-to-IRB connectivity across the ICL.

      This is configured to support OSPF over the Inter-Control Center Communications Protocol (ICCP).

    • Load balancing between mc-lag peers is 100 percent local bias by default.
    • Load balancing within local peer is the same as normal lag hashing.
    • Two possible options for Layer 3 gateway is Virtual Router Redundancy Protocol (VRRP) based or irb-mac-sync.
    • If irb-mac-sync is used, routing protocols on IRB are not supported.
    • In a VRRP-based Layer 3 solution, even the VRRP backup node forwards traffic.
    • Prefer separate link aggregation group (LAG) links for ICL and ICCP.
    • ICCP peering with loopback IP is preferred to use all available links through interior gateway protocol (IGP).
    • Configure backup-liveness detection to get sub-second traffic loss during MC-LAG peer switch reboots.
    • Spanning Tree Protocol (STP) is not supported on ICL or MC-LAG interfaces.
    • Access security features are not supported on ICL or MC-LAG interfaces.
    • Configure mcae with “prefer status control active” on both provider edge routers (PEs) to avoid lacp system ID flap during active node (SW) reboot.

    MC-LAG Configuration for Better Convergence

    To improve the network convergence in this solution, the following configuration statements are configured:

    • prefer-status-control active on both the MC-LAG nodes for all MC-AE interfaces. With this configuration, LACP system-id will be retained on both ICCP/ICL failures to improve convergence.
    • Loopback IP peering is configured for the ICCP protocol. The ICCP peer can be reached over protocols in case of direct ICCP link failure.
    • The 1-second BFD timer is configured for the ICCP protocol and all IRB/VRRP interfaces.
    • More than 1-second hold-down timer is configured on the ICL links to prevent ICL link start-up before the ICCP during failure events.
    • init-delay-time is configured on EX9214 core switches to delay the start-up of the MC-AE interfaces before the protocol convergence. This knob removes any packet loss during the recovery of failed links and devices.

    Published: 2015-04-20