MPLS Lab Tips

  • Router ID for BGP, LDP and OSPF should ALL match and be routable form the P routers..
  • As a best practice and to be on the safe side, always set your loopbacks for MPLS to /32 (especially in ospf using ip ospf network type point-to-point).
  • If using TDP be sure to include mpls label protocol tdp under the interface as well.
  • Can use mpls ldp discovery transport-address interface/ip to specify the source peering address which by default will be the highest loopback address.
  • mpls ldp router-id loopback 0 force is important otherwise higher loopback will reset the session.
  • RD keeps the routing table unique (overlapping subnets etc).
  • Route-Target assigns an import and export tag through use of dynamic routing e.g. BGP using extended community. You then use redistribute-connected under the address-family ipv4 unicast  VRF BGP process to bring those routes into that VRF domain. Don’t forget to activate the neighbour with neighbour x.x.x.x activate and add send-community both under the VPNV4 process / address-family vpnv4 unicast process.
  • Troubleshooting pings – Check mpls is enabled on all interface and the LDP/TDP relationship is formed on each interface and that show mpls fowarding-table shows a label or pop for the next hop.
Advertisements

Trunk Links Lab Tips

  • If the Lab requires that all traffic be tagged with a vlan header when sent over a trunk, they are talking about ISL, especially if they restrict you from issuing any global commands.
  • If global commands are not restricted then the above can be achieved with 802.1q by issuing the vlan dot1q tag native in global command mode.

Ether Channel Lab Tips

  • If configuring mode as on then you must configure the other end on within60 seconds! otherwise Spanning Tree will consider this as a loop. Better, shut down the interface and then configure and unshut once both sides are done.
  • show etherchannel load-balance will show the current load balancing algorithm.
  • From global command, port-channel load-balance will allow you to change the algorithm.
  • port-channel and etherchannel commands and sub-commands for link aggregation.
  • All interfaces in the port-channel will inherit the configuration. so just configure logical to save time as physical interfaces will inherit config i.e. configure port-channel 12 instead of fa0/16 etc

RIP Lab Tips

  • Show key chain and show ip protocols to see authenentication stuff.
  • Watch for spaces / trailing spaces e.g. show run | inc Cisco$ or Cisco $
  • Show ip protocols to check for filters and timers etc and passive interfaces.
  • Debug ip rip may show auth and version mismatches ete.

Troubleshooting – Link Layer

  • Debug ip icmp
  • Debug ip packet
  • Debug interface
  • Debug arp
  • Debug frame packet (if debug ip packet shows encapsulation failed)
  • Passive interfaces – can bypass with neighbour statement as it only stops multicasts NOT unicasts

Frame Relay Lab Tips

  • Shut down port first and then configure – this can help against unwanted dynamic maps.
  • Set the hub as the DR for everything e.g. PIM, OSPF.
  • Inverse arp and no arp frame to avoid dynamic entries.
  • Show frame map and make sure nothing says ‘dynamic’. If it does, do ‘clear frame map’, reload or remove the encapsulation frame-relay command which removes everything!
  • No Need to do ‘BROADCAST’ from spoke to spoke if going via hub.
  • enable frame-relay switching in global configuration for switching.
  • frame-relay intf-type dce required on the FR Switch to bring the line protocol up.
  • RIP –

– Check hub has broadcast option on map to each spoke.

– Disable split horizon on hub for spoke to spoke comms or – Neighbour statements or GRE tunnel!