Verify HW VoQ Programming on a Multichassis NCS4K

Introduction

This document describes the procedure to verify Hardware (HW) Virtual Output Queueing (VoQ) programming on a multichassis NCS4K.

Procedure

This is a step-by-step approach to verify HW VoQ programming on NCS4K nodes.

Multichassis nodes in 6.5.26 release are potentially affected by Cisco bug ID CSCvz41459, where VoQ is misprogrammed on the ingress Line Card (LC) Virtual Machine (VM).

Step 1. First, you need to check the ingress and egress interface for the traffic flow.

In this example, the ingress interface is TenGigE1/3/0/0/1.4001 (rack 1), and the egress interface is Hu0/9/0/11/2.4001(rack 0). So, ingress and egress interfaces are on different racks, and you could potentially be affected by Distributed Defect Tracking System (DDTS) Cisco bug ID CSCvz41459

For example, see tunnel 117:

#show mpls forwarding tunnels 117
Wed Nov  9 13:15:47.159 UTC
Tunnel        Outgoing    Outgoing     Next Hop        Bytes
Name          Label       Interface                    Switched
------------- ----------- ------------ --------------- ------------
tt117             24764       Hu0/9/0/11/2.4001 172.16.13.170   0

Step 2. Check active LC VM in rack 0 and rack 1:

#show redundancy summary
Wed Nov  9 13:16:59.309 UTC
    Active Node    Standby Node
    -----------    ------------
          1/LC0           1/LC1 (Node Ready, NSR:Not Configured)
          0/RP1           1/RP0 (Node Ready, NSR:Ready)
          0/RP0           1/RP1 (Node Ready, NSR:Not Configured)
          0/LC0           0/LC1 (Node Ready, NSR:Not Configured)

In this example, 0/LC0 is an active VM in rack 0 and 1/LC0 is in rack 1.

Step 3. As a next step, check SysPort number for the egress interface in egress LC VM:

#show controllers npu voq-usage interface HundredGigE 0/9/0/11/2.4001 instance all location 0/lc0
Wed Nov  9 13:16:45.149 UTC

-------------------------------------------------------------------
Node ID: 0/LC0
Intf         Intf     NPU NPU  PP   Sys   VOQ   Flow   VOQ    Port
name         handle    #  core Port Port  base  base   port   speed
             (hex)                                     type
----------------------------------------------------------------------
Hu0/9/0/11/2.4001 8000bd4  18   1  448 24655   1336   5152 local   100G
Hu0/9/0/11/2.4001 8000bd4   0   0    0 24655   1144      0 remote    0M

In this case, SysPort is 24655 and VoQ base is 1336.

Step 4. Check the same SysPort on the ingress line card and ingress VM.

The command is:

show controllers fia diagshell <ingress Line card number> "dump IRR_DESTINATION_TABLE <SysPort>" location <ingress LC VM>

For example:

#show controllers fia diagshell 3 "dump IRR_DESTINATION_TABLE 24655" location 1/lc0
Wed Nov  9 13:18:00.684 UTC

Node ID: 1/LC0
IRR_DESTINATION_TABLE.IRR0[24655]: <VALID_1=1,VALID_0=1,TC_PROFILE=0,QUEUE_NUMBER=0x538,ECC=0>

In this example, QUEUE_NUMBER is 0x538 in HEX, that is matching to your VoQ base 1336 (dec) from the previous output.

If you hit  Cisco bug ID CSCvz41459 the QUEUE_NUMBER is 0x1ffff.

For example:

#show controllers fia diagshell 3 "dump IRR_DESTINATION_TABLE 24655" location 1/LC0
Wed Nov  9 12:44:54.270 UTC
 
Node ID: 1/LC0
IRR_DESTINATION_TABLE.IRR0[24655]: <VALID_1=0,VALID_0=0,TC_PROFILE=0,QUEUE_NUMBER=0x1ffff,ECC=0>

Recovery Method

The recovery method for this issue is to remove the affected egress subinterface, and re-apply it back.

Do not forget to check the flow path in the reverse direction as well.