OTV Site VLANs and AED Election Scenario

Introduction

This document discusses a common Overlay Transport Virtualization (OTV) scenario, where the site-VLAN does not come online or the Authoritative Edge Device (AED) does not forward a set of VLANs, and the steps to troubleshoot this scenario.

Background Information

This document assumes a multi-AED setup for either unicast or multicast OTV.

Scenario

AED does not become active for a VLAN.

1. Enter the show otv command and check for errors. For example, 'Site-ID mismatch' is problematic.
2. Enter the show otv site command and confirm the 'System-ID'.

   This is used in order to elect the AED so you know what should be active for each VLAN.

   RTP-OTV1# show otv site
   [snip]
   System-ID 8478.ac0b.6044
   Site-Identifier aaaa.aaaa.aaaa
   Site-VLAN 6 State is Up 
   
   
   Site Information for Overlay1:
   
   
   Local device is AED-Capable
   Neighbor Edge Devices in Site: 1
    
   Hostname   System-ID   Adjacency-   Adjacency-     AED-
                             State     Uptime         Capable
   -------------------------------------------------------------
   RTP-OTV1  8478.ac0c.7b44   Full       1w6d           Yes

3. Confirm that the site-VLAN is the same on both OTV Virtual Device Contexts (VDCs) in that site.

   Nexus-otv# show run | in "site-vlan"
      otv site-vlan 7

   The best practice is to use different site-VLANs per site.

4. Verify the Spanning Tree Protocol (STP) state of the site-VLAN along all switches in the path between OTV VDCs. It should be forwarded on the Layer 2 (L2) ports to the agg and the agg should have its ports forwarded between VDCs as well.

   Nexus-otv# show spanning-tree vlan 7
   [snip]
   Interface        Role Sts Cost      Prio.Nbr Type
   ---------------- ---- --- --------- -------- --------------------------------
   Po1              Root FWD 1         128.4096 Network P2p

5. Check the event history in order to verify that hellos come in from the other AED at the local site through both the overlay and the site-VLAN and that the local AED sends two out.

   RTP-OTV1# show otv isis internal event-history iih | in "IIH"
   2015 Mar 26 14:35:09.227474 isis_otv default [14312]: [14321]:
    Send L1 LAN IIH over Overlay1 len 1397    prio6,dmac 0100.0cdf.dfdf
   2015 Mar 26 14:35:08.803200 isis_otv default [14312]: [14325]:
    Receive L1 LAN IIH over Overlay1 from      SJ-OTV1 (8478.ac0c.7b47) len 1397 prio 7
   2015 Mar 26 14:35:08.434157 isis_otv default [14312]: [14325]:
    Receive L1 LAN IIH over site-vlan from    RTP-OTV1 (8478.ac0c.7b44) len 1497 prio 7
   2015 Mar 26 14:35:08.345369 isis_otv default [14312]:
    [14321]: Send L1 LAN IIH over site-vlan len 1497    prio6,dmac 0100.0cdf.dfdf

   Note: If the site VLAN is the same on both sites and hello packets are switched from one site to another, then OTV detects the different site ID and inactivates the VLANs in order to shut down one site. This is completed in order to avoid an L2 loop. Enter the show otv output command for verification. A 'Site-ID mismatch' should be displayed when this occurs. You can change the site-VLAN or stop the loop before the site-VLAN in order to correct the issue. AED can then be brought up.

6. Check the statistics in order to verify whether the hellos are received in the counters.

   RTP-OTV1# show otv isis site statistics | begin PDU
   OTV-IS-IS PDU statistics for site-vlan:
   
   PDU         Received     Sent          RcvAuthErr    OtherRcvErr   ReTransmit
   LAN-IIH     91697        91700             0              0             n/a <<<<<<
   
   
   
   RTP-OTV2# show otv isis traffic
   OTV-IS-IS process: default
   VPN: Overlay1
   OTV-IS-IS Traffic:
   PDU           Received     Sent    RcvAuthErr  OtherRcvErr ReTransmit
   LAN-IIH        85530   23298                0               0           n/a <<<<<<

7. Check the ordinal numbers in order to ensure they are assigned correctly:

   ////////////////////////////////////// Note //////////////////////////////////////
     Lower IS-IS System-ID (Ordinal 0) = AED for EVEN VLANs
     Higher IS-IS System-ID (Ordinal 1) = AED for ODD VLANs
   //////////////////////////////////////////////////////////////////////////////////
    
   RTP-OTV1# show otv internal site
   Group:239.1.1.1 Queue-size:2 num_site_adjs:2 fwd_rdy_cnt:2 local-present:1
      System-ID           Up Time   Ordinal  Fwd-Rdy
      8478.ac0b.6044      1w6d       0        1
      * 8478.ac0c.7b44  1w6d       1        1 <<<<< active for ordinal 1,
      so this VDC should be AED for odd VLANs
   
   
   
   RTP-OTV2# show otv internal site
   Group:239.1.1.1 Queue-size:2 num_site_adjs:2 fwd_rdy_cnt:0 local-present:1
      System-ID            Up Time   Ordinal  Fwd-Rdy
      * 8478.ac0b.6044  1w6d       0        1 <<<< active for ordinal 0,
      so this VDC should be AED for even VLANs
      8478.ac0c.7b44       1w6d       0        1
   
   
   
   RTP-OTV1# show otv vlan
   
   
   OTV Extended VLANs and Edge Device State Information (* - AED)
   
   
   Legend:
   (NA) - Non AED, (VD) - Vlan Disabled, (OD) - Overlay Down
   (DH) - Delete Holddown, (HW) - HW: State Down
   (NFC) - Not Forward Capable
   
   
   VLAN   Auth. Edge Device      Vlan State     Overlay
   ---- -------------------- ----------- -------
   190     RTP-OTV2               inactive(NA)  Overlay1
   191*    RTP-OTV1              active           Overlay1 <<< This AED
    is active for odd vlans as expected

   Note: Keep Cisco bug ID CSCur30401 in mind when you look at Ordinals!

Related Cisco Bug IDs

• Cisco bug ID CSCum54509 - Release 6.2(6), site adjacency stuck in partial state because F Series linecard packets destined to 0100.0cdf.dfdf are not flooded to Bridge Domain (BD) 
  
  
• Cisco bug ID CSCur30401 - Ordinal in 'show otv site detail' is 0 for all switches