Configuring an EVPL Instance

Procedure

1. Run system-view

   The system view is displayed.

2. Run evpn vpn-instance vpn-instance-name vpws

   A VPWS EVPN instance is created, and its view is displayed.

3. Run route-distinguisher route-distinguisher

   An RD is configured for the EVPN instance.

   An EVPN instance takes effect only after an RD is configured for it. The RDs of different EVPN instances on the same PE must be different.

   

   After being configured, an RD cannot be modified, but can be deleted. After the RD of an EVPN instance is deleted, VPN targets configured for the EVPN instance are also deleted.

4. Run vpn-target vpn-target &<1-8> [ both | export-extcommunity | import-extcommunity ]

   VPN targets are configured for the EVPN instance.

   VPN targets are BGP extended community attributes used to control the acceptance and advertisement of EVPN routes. You can configure a maximum of eight import VPN targets and eight export VPN targets using the vpn-target command each time. To configure more VPN targets in the EVPN instance address family, run the vpn-target command multiple times.

5. (Optional) Run timer revert delay delay-value

   A switchback delay time is set.

   In dual-homing scenarios, if a fault on the access-side link of the master PE or the master PE is rectified, per-EVI A-D routes are generated on the master PE. Then, per-EVI A-D routes can be advertised to the remote PE based on the BGP EVPN peer relationship. After receiving the per-EVI A-D routes and generating forwarding entries, the remote PE switches traffic from the path destined for the backup PE to the path destined for the master PE. In this case, a few packets may be lost on the master PE because some forwarding entries are not generated. To prevent this problem, run the timer revert delay delay-value command on the remote PE to configure a proper switchback delay time. After receiving per-EVI A-D routes, the remote PE delays generating forwarding entries. Specifically, the remote PE generates forwarding entries only after the forwarding entries on the master PE become stable. After the delay times out, the remote PE generates new forwarding entries and sends traffic to the master PE, preventing packet loss.

6. Run quit

   Return to the system view.

7. Run evpl instance evpl-id

   An EVPL instance is created, and its view is displayed.

8. Run evpn binding vpn-instance vpn-instance-name

   The specified VPWS EVPN instance is bound to the current EVPL instance.

9. Run local-service-id service-id remote-service-id service-id

   The current EVPL instance is configured to carry the local and remote service IDs in packets.

10. (Optional) Run load-balancing ignore-esi

    The function to check ESI validity during EVPL instance load balancing is disabled.

    In an EVPN VPWS scenario where active-active protection is deployed, if each access-side device is single-homed to an aggregation-side device and no ESI is configured on the access interface, to implement active-active load balancing, you can run this command on the aggregation-side device to enable the device to ignore the ESI validity check during EVPL instance load balancing.

11. (Optional) Run control-word enable

    The control word function is enabled for the EVPL instance.

    In load-balancing mode, packet disorder may occur during the in-depth parsing of MPLS packets. In this case, you can enable the control word function in the EVPL instances on both ends, so that MPLS packets can be reassembled using the control word.

12. Run commit

    The configuration is committed.