Nexus 56128 - SFP behavoir varies on SUP and modules

lauresfrank · ‎09-23-2018

Dear Community,

we see a weird behavoir of 1G multimode GLC-SX-MMD SFPs on every port of the modules (except N5K-C56128P-SUP).

The link comes up when using SUP SFP slots (Mod1), but links remain in down state when inserting SFPs inside slot 2+3 (N56-M24UP2Q).

Mod Ports Module-Type                         Model                  Status
--- ----- ----------------------------------- ---------------------- -----------
1   52    Nexus 56128P Supervisor             N5K-C56128P-SUP        active *
2   26    Nexus 24x10GE Unified Port + 2xQSFP N56-M24UP2Q            ok
3   26    Nexus 24x10GE Unified Port + 2xQSFP N56-M24UP2Q            ok

This behavoir is seen when we connect this switch to our Nexus: Datasheet of switch

I also created a video where you can see this behavoir:

Some more information you may find useful:

Ethernet3/1
  Model:                 N56-M24UP2Q
  Type (SFP capable):    SFP-1000BASE-SX
  Speed:                 1000,10000
  Duplex:                full
  Trunk encap. type:     802.1Q
  Channel:               yes
  Broadcast suppression: no
  Flowcontrol:           rx-(off/on),tx-(off/on)
  Rate mode:             none
  QOS scheduling:        rx-(6q1t),tx-(1p6q0t)
  CoS rewrite:           no
  ToS rewrite:           no
  SPAN:                  yes
  UDLD:                  yes
  MDIX:                  no
  Link Debounce:         yes
  Link Debounce Time:    yes
  dot1Q-tunnel mode:     yes
  Pvlan Trunk capable:   yes
  TDR capable:           no
  FabricPath capable:    yes
  Port mode:             Switched
  FEX Fabric:            yes

 show int ethernet 1/43 capabilities 
Ethernet1/43
  Model:                 N5K-C56128P-SUP
  Type (SFP capable):    SFP-1000BASE-SX
  Speed:                 1000,10000
  Duplex:                full
  Trunk encap. type:     802.1Q
  Channel:               yes
  Broadcast suppression: no
  Flowcontrol:           rx-(off/on),tx-(off/on)
  Rate mode:             none
  QOS scheduling:        rx-(6q1t),tx-(1p6q0t)
  CoS rewrite:           no
  ToS rewrite:           no
  SPAN:                  yes
  UDLD:                  yes
  MDIX:                  no
  Link Debounce:         yes
  Link Debounce Time:    yes
  dot1Q-tunnel mode:     yes
  Pvlan Trunk capable:   yes
  TDR capable:           no
  FabricPath capable:    yes
  Port mode:             Switched
  FEX Fabric:            yes


Mod  Sw              Hw      World-Wide-Name(s) (WWN)
---  --------------  ------  ---------------------------------------------------
1    7.3(3)N1(1)     2.0     --
2    7.3(3)N1(1)     2.0     --
3    7.3(3)N1(1)     2.0     --



!Command: show running-config interface Ethernet1/43
!Time: Thu Sep 20 08:28:21 2018

version 7.3(3)N1(1)

interface Ethernet1/43
  switchport access vlan 22
  speed 1000
  no shutdown

WSF-SWCORE01# sh run int ethernet 3/1

!Command: show running-config interface Ethernet3/1
!Time: Thu Sep 20 08:28:26 2018

version 7.3(3)N1(1)

interface Ethernet3/1
  switchport access vlan 22
  speed 1000
  no shutdown

I also noticed that the switchport raimains in down state, although you can see Layer 1 RX state values on the interface (show int transeiver details). Thats weird, ladies and gentlemen.

You may have an idea and can help? Thanks a lot.

ADP_89 · ‎09-23-2018

Hello,

I saw this behaviour on non supported transcievers, are yours original or compatible? Have you tried the command "service unsupported-transceiver". Eventually we ended up by using a different model of SFPs..

HTH,

ADP

lauresfrank · ‎09-23-2018

Thank you! We use cisco sfps. I showed them in the video. Same behavoir with different multimode sfps from different vedors.

pieterh · ‎09-24-2018

do you have UDLD enabled? (you plug fibers in one-by-one)

lauresfrank · ‎09-24-2018

Thanks for your reply. Udld is disabled on both ports.