RIFT Overview and Set Up

Enabling RIFT Using Load Command

To load the RIFT configuration manually:

1. After you successfully install the junos-rift package, run the following command to load and combine the RIFT configuration from a specified file with the current configuration in the CLI:

   Here, platform is the host device, and can be any one of the following values—mx, qfx, or vmx.

2. Commit the configuration.

Enabling RIFT in CLOS-based Topology (ZTP Mode)

Note:

For activating the RIFT software package on a CLOS topology, additional configuration is required. You must identify the nodes that are the top-of-fabric in the topology, and configure all the top-of-fabric devices to override the auto level in the default configuration.

To activate the RIFT software package in a CLOS-based topology:

1. Override the auto level in the default configuration, and optionally, specify the levels manually.

2. Commit the additional configuration.

Traceoptions for RIFT

Although enabling the RIFT protocol automatically inherits the necessary configuration, you can additionally configure minimal tracing as optional configuration.

To configure traceoptions for RIFT:

1. Specify the traceoptions and proxy-process parameters under the rift statement.

   For example:

Verifying RIFT Configuration

You can verify the RIFT protocol configuration from the following hierarchy levels:

• [groups rift-defaults]

• [interfaces interface-range rift-interfaces]

• [protocols rift]

You can also verify the RIFT configuration by viewing the defaults applied from the junos-rift package. To do this, run the show configuration protocols rift | display inherited command.

For example:

Table 1 list the commands you can use to verify the RIFT protocol configuration and status.

Table 1: Commands to Verify RIFT Protocol Configuration

Command	Description
show rift	Inspect the runtime state of the RIFT protocol.
show route protocol rift	The RIFT protocol can be used in policies and commands where other protocols are accepted.
show route protocol rift extensive display-client-data	View detailed RIFT-installed routes.
clear rift database content	Clear the RIFT database.
restart rift-proxyd	Restart the RIFT protocol.

For more information, see the FAQ file in the software package distribution.

Troubleshooting the RIFT Protocol

The RIFT protocol does not produce core files except in very extreme cases. It reports every failure by extensive logging and sometimes backtraces on exit. The RIFT process provides configurable tracing events that can be collected using traceoptions configuration.

To troubleshoot the RIFT implementation, see:

• Forming Adjacency

  Problem

  RIFT adjacency flapping up and down, showing rejects with Multiple Neighbors or Remote Uses Our Own SystemID errors.

  Solution

  The RIFT protocol does not support more than two neighbors on an Ethernet link forming a point-to-point adjacency, or a node's own interfaces looped back. Check and correct the cabling.

• Undefined Level

  Problem

  All the switches show undefined level and do not form three-way adjacencies, but link information elements (LIEs) are being sent and received.

  Solution

  There is a possibility that there is no top-of-fabric level configuration. All top-of-fabric devices must be configured with the top-of-fabric level to provide an anchor for ZTP.

• Loopback Address

  Problem

  Not able to get loopback addresses to my nodes in RIFT.

  Solution

  To have all the loopback addresses in top-of-fabric, the simplest way is to configure a loopback address in every node necessary and redistribute it in the northbound direction into RIFT. The following configuration is required for doing this:

  This configuration allows every leaf to ping the loopback address of all other nodes except the top-of-fabric devices.

  If the top-of-fabric devices should also be reachable from the leaf devices, or vice-versa, the top-of-fabric loopback addresses need to be exposed to one level below (that is, southbound). The following configuration is required for doing this:

  Note:

  To enable the top-of-fabric addresses to be propagated to all the leaf nodes, configure the allow-rift-routes option under the [edit protocols rift export southbound] hierarchy level.

• System Log Error Messages

  The RIFT process generates system log messages to record errors related to the integration of the RIFT protocol into Junos OS. To interpret system log messages, refer to the following:

  • RIFT_PROXYD_ALREADY_RUNNING—Another instance of the RIFT process is already running.

  • RIFT_PROXYD_CONNECT_RIFT—Attempts to connect to the local RIFT process failed.

  For more information on system log error messages, see System Log Explorer.