Reset Search
 

 

Article

ICMP Redirect feature's possible hindrance to a 'VE over VPLS' configuration

« Go Back

Information

 
TitleICMP Redirect feature's possible hindrance to a 'VE over VPLS' configuration
Symptoms
  • Routing of control/data packets from a host in a VPLS domain to another router/host outside of its VPLS domain/subnet may fail.
Environment
  • Netiron
  • MLX/XMR
  • CER/CES
Cause
By default, routers have 'ICMP redirect' feature enabled to tell the downstream host/router that there is a better path to reach the destination.
  • Although this feature is used as a way to optimize the flow of traffic across the network, it also is used to prevent packets from traversing the same outgoing link as the one that they came in through; to prevent L3 (routing) loops.
In 'VE over VPLS' configurations, it is common for a data packet (with next hop as the 'VE interface') to traverse a link to a router, to only get routed by the router (to a destination) back along the same link. This router usually contains the 'VE interface' for a VPLS domain.
  • The reason the packet even comes to this router is to exit the VPLS domain through the VE interface.
  • In some cases, the packet's next-hop/destination may be along the same router (with the mentioned VPLS domain) from which it came in.
If 'ICMP redirect' feature is enabled, these packets are dropped by the router in order to prevent a loop from forming in the network.
Resolution
Disabling the 'ICMP Redirect' feature globally on the MLX hosting the VE interface for the VPLS domain, would allow packets to be routed along the same outgoing interface as their incoming interface; hence solving the routing issue:
  • MLX(config)#no ip icmp redirects
Additional notes

Feedback

 

Was this article helpful?


   

Feedback

Please tell us how we can make this article more useful.

Characters Remaining: 255