Help us improve your experience.

Let us know what you think.

Do you have time for a two-minute survey?

 

Overview of Internal Fabric Monitoring

 

Internal fabric monitoring is a feature of the Operation, Administration, and Maintenance (OAM) of the QFabric system. This feature enables you to validate the flow path of protocol data units (PDUs) across a given VLAN on the QFabric system using the unicast ping, multicast ping, and traceroute operations.

Internal fabric monitoring is useful for fault detection on the QFabric system. For example, if a PDU reaches a destination that is not part of the VLAN configuration, the operation (unicast ping, multicast ping, or traceroute) displays the exception on the console at runtime.

The unicast and multicast ping operations send PDUs from a source interface (called the source fabric maintenance endpoint [FMEP]) to a destination FMEP. The destination FMEP sends a response to the source FMEP when the PDUs are received.

The traceroute operation, also called a flow linktrace, traces the path taken by a specific, learned unicast flow from a source FMEP to a destination FMEP in a VLAN. The source and destination FMEPs may be on the same Node device, different Node devices connected to the same Interconnect device, or different Node devices connected to different Interconnect devices. The flow path is the sequence of Packet Forwarding Engine forwarding hops along which the PDU travels. The hop-by-hop sequence and number of hops are reported in terms of fabric maintenance intermediate points (FMIPs), which are interfaces on the Packet Forwarding Engine of the Interconnect device. An FMIP sends a response to the source FMEP when the traceroute PDU is received.

The following internal fabric monitoring commands are supported:

  • show oam fabric flow specification

  • show oam fabric interfaces

  • ping fabric unicast-flow

  • ping fabric multicast-flow

  • traceroute fabric unicast-flow

The following are the main components of the internal fabric monitoring feature:

  • FMEP—Represents the source or destination point (endpoint) in the monitoring operations. An FMEP is an interface through which PDUs are sent (source FMEP) or received (destination FMEP). Upon receipt of the PDUs, the destination FMEP sends a response to the source FMEP to validate the monitoring flow. FMEPs are associated with a VLAN in the fabric maintenance association (FMA) configuration, and the source and destination FMEP addresses are configured in the flow specification.

  • FMA—Associates a set of FMEPs with a VLAN. The FMA defines the VLAN and FMEP parameters, including the VLAN name, FMEP identifiers, FMEP names, and the interface names of the FMEPs. The FMEPs defined in the FMA are the source and destination FMEPs used in the monitoring commands.

    Note

    A default FMA is automatically created for each Node group in the QFabric system. The default FMA is used to send error response PDUs (for example, in the case of a VLAN leak) and responses to PDUs that are not mapped to a specific interface in the QFabric system.

  • Flow specification—Configures the flow type and FMEP addressing parameters. Unicast flow types include an Ethernet type (other than IPv4) and Ethernet IPv4. Multicast flow types include the Ethernet IPv4 and VLAN flood type. The flow specification also defines parameters within each flow type, such as MAC or IPv4 addresses of the source and destination FMEPs. The names and identifiers of these FMEPs are configured in the FMA parameters.

To enable internal fabric monitoring, configure the fabric-maintenance-associations and flow-specs statements at the [edit protocols oam fabric] hierarchy level.

The ping fabric unicast-flow, ping fabric multicast-flow, and traceroute fabric unicast-flow commands require that you specify the flow specification and FMA names, as well as the source and destination FMEP names.