Nexus 7K Optimised multicast flooding enabled causing out of order pac

[Larry Sullivan]

Hi,

I have a strange one for you and I know this is a long shot.  We have two Nexus 7Ks for our core and recently started getting out of order packets on our UDP multicast video streams when traversing our Nexuses.  We discovered that disabling optimised multicast flooding on our Nexus B eliminates the issue even when we use a static to send the test stream through our Nexus A.  We've tried switching supervisors on the Nexus B but are still getting the errors.  TAC hasn't been able to resolve it yet and we are at our wits end.  Has anyone ran into issues like this that involves Nexus 7Ks, OMF, and UDP multicast video?  There doesn't seem to be any queueing issues, multi-pathing issues, congestion issues, or any of the typical out of order packet issues.  Also unicast doesn't have any issues.  Strictly multicast video with OMF enabled globally.

[balaji.bandi]

what version of NEXUS OS and what Sup card ?

check below information may help you :

https://www.cisco.com/c/en/us/support/docs/switches/nexus-7000-series-switches/113619-nexus7000-error-messages.html

[Larry Sullivan]

Yeah we were getting the below log error but that was resolved a week ago but we are still seeing the out of order packets with OMF enabled globablly.

%L2MCAST-SLOT2-2-L2MCAST_MAC_FULL_LC:

system: version 8.2(6)

N77-SUP2E HW 1.2

[balaji.bandi]

I would prefer to open a TAC case to work with you to investigate :

other hand  if you still looking to pursue the issue :

https://community.cisco.com/t5/switching/nexus-7000-error-message/td-p/2247954

[Larry Sullivan]

We already have a TAC case ongoing on this and yes I saw the post you linked there already.  Due diligence on my behalf has been done in regards to Googlefu.  Appreciate the replies and if a solution is found I will post it here.  Thanks again.

[marce1000]

 - You may find this thread informational : https://community.cisco.com/t5/switching/l2mcast-slot2-2-l2mcast-mac-full-lc/td-p/2584109

 M.

[Larry Sullivan · Accepted solution]

I found the issue, ended up being we we're getting output drops on our Nexus to Nexus port-channel.  Keep in mind this port channel has 3x40Gb interfaces in it and total traffic was only around 5-10 Gb.  Our ingress interfaces from the LAN had CoS marking the video streams as CoS 4.  A look at the queuing showed CoS4 as 14 percent of remaining queue with no priority.  CoS5 showed priority Queuing.  So I changed the QoS policy to mark traffic as CoS5 and bam errors cleared up.

Below didn't format right but I highlighted the relevant parts.  the 4 and 5 are the CoS values and the High is priority while the 14 is the BW percentage.

Egress Queuing for Ethernet1/1 [System]
-------------------------------------------
Template: 8Q8E
-----------------------------------------------------------------------------
Queue Group Bandwidth% PrioLevel Shape% CoSMap
-----------------------------------------------------------------------------
8e-4q8q-out-q4 3 14 - - 4
8e-4q8q-out-q5 4 14 - - 3
8e-4q8q-out-q6 5 14 - - 2
8e-4q8q-out-q7 6 14 - - 1
8e-4q8q-out-q1 0 - High - 5
8e-4q8q-out-q2 1 14 - - 7
8e-4q8q-out-q3 2 14 - - 6
8e-4q8q-out-q-default 7 14 - - 0