Enabling OSPF IP FRR

Procedure

1. Run system-view

   The system view is displayed.

2. Run ospf [ process-id | router-id router-id | vpn-instance vpn-instance-name ] ^*

   An OSPF process is started, and the OSPF view is displayed.

3. Run frr

   The OSPF FRR view is displayed.

4. Run loop-free-alternate

   OSPF IP FRR is enabled, and a loop-free backup link is generated.

   

   OSPF can generate a loop-free backup link only when the OSPF IP FRR traffic protection inequality is met. For detailed description of OSPF IP FRR, see HUAWEI NetEngine 8000 F Series Feature Description- OSPF IP FRR.

5. (Optional) Run frr-policy route { route-policy route-policy-name | route-filter route-filter-name }

   An OSPF IP FRR filtering policy is configured.

   After the OSPF IP FRR filtering policy is configured, only the OSPF backup routes that match the filtering conditions of the policy can be added to the forwarding table.

6. If you want to configure remote LFA OSPF IP FRR, perform the following steps:
   1. Run remote-lfa tunnel ldp [ maximum-reachable-cost cost-value ]

      Remote LFA OSPF IP FRR is enabled.

   2. (Optional) Run remote-lfa available-tunnel-destination ip-prefix ip-prefix-name

      A filtering policy is configured to filter PQ nodes.

      Only the PQ node that matches the filtering policy can be used as the next hop of an LFA link.

   3. Run quit

      Return to the OSPF view.

   4. (Optional) Run avoid-microloop frr-protected

      The OSPF anti-microloop is enabled.

   5. (Optional) Run avoid-microloop frr-protected rib-update-delay rib-update-delay

      The delay after which OSPF delivers routes is configured.

      If OSPF remote LFA FRR is enabled and the primary link fails, traffic is switched to the backup link. If route convergence occurs again, traffic is switched from the backup link to a new primary link. During the switchover, microloop may occur. To prevent this problem, OSPF anti-microloop is enabled and delays the switching. To configure the delay, run the avoid-microloop frr-protected rib-update-delay command. After the command is run, OSPF does not switch traffic to the backup link until the delay elapses.

   6. Run frr

      The OSPF FRR view is displayed.

7. (Optional) Run tiebreaker { node-protecting | lowest-cost | ldp-sync hold-max-cost | srlg-disjoint } preference preference

   The solution of selecting a backup path for OSPF IP FRR is set.

   By default, the solution of selecting a backup path for OSPF IP FRR is node-protection path first. In some cases, the solution needs to be changed to smallest-cost path first because of data forwarding capacity or link cost consideration. In Figure 1, the primary path is Link-1 (Device S -> Device E -> Device D), and Link-2 and Link-3 (Device S -> Device N -> Device D) are backup path candidates. By default, Link-3 is selected as the backup path. To change the solution of selecting a backup path for OSPF IP FRR to smallest-cost path first, run the tiebreaker command. After the command is run, Link-2 is selected as the backup path.

   Figure 1 Solution of selecting a backup path for OSPF IP FRR
   
   Figure 2 shows an inter-board scenario, where Link-1 (Device A -> Device D) is the primary path, and Link-2 (Device A -> Device E -> Device D) is the backup path. If Link-1 fails, Link-2 functions as the new primary path, and Link-3 (Device A->Device B->Device C->Device D) functions as the new backup path. If Link-1 goes Up again but the LDP session has not gone Up, OSPF enters the Hold-max-cost state. Consequently, the primary path is still Link-2, and the backup path is still Link-3. If the LDP session goes Up but ldp-sync hold-max-cost is not configured, OSPF exits from the Hold-max-cost state when the timer used to delay sending an LDP session Up message expires. In this case, OSPF switches the primary path back to Link-1. Because the upstream and downstream entries reside on different boards and the downstream entry has not been updated when downstream traffic arrives, packet loss occurs. To resolve the problem, configure ldp-sync hold-max-cost so that OSPF preferentially selects the path with the maximum cost set by LDP-IGP synchronization when OSPF is in the Hold-max-cost state. Then OSPF switches the backup path to Link-1 and delivers the backup forwarding entry in advance. When the timer used to delay sending an LDP session Up message expires, OSPF exits from the Hold-max-cost state and switches the primary path to Link-1. Because the downstream backup entry is available, no packet loss occurs.

   Figure 2 Maximum-cost (set by LDP-IGP synchronization) path first solution
   

8. Run commit

   The configuration is committed.