issue _vPC on N9K-C93108TC-FX3P with QSFP-100G-CU1M cable

[herve.leon]

 I'd like to create a vPC between two Nexus N9K-C93108TC-FX3P . I have two QSFP-100G-CU1M DAC cables compatibles (https://tmgmatrix.cisco.com/?si=N9K-C93108TC-FX3P)

However, the status of my ports is always 'DISCONNECTED'.   

Note: I have no issue when I create the vPC between two N9K-C93240YC-FX2 with the same cables and configuration.

Is there anything particular on this C93108 ?

# show int status

Eth1/53 -- notconnec trunk auto auto QSFP-100G-CR4
Eth1/54 -- notconnec trunk auto auto QSFP-100G-CR4

The small configuration applied:

feature vpc
feature lacp

interface mgmt0
description **vPC peer keepalive**
vrf member management
ip address 10.99.1.1/30

vrf context management

vpc domain 5
role priority 8192
peer-keepalive destination 10.99.1.2 source 10.99.1.1
exit

interface ethernet 1/53 - 53
channel-group 5 mode active
no shutdown

interface port-channel 5
switchport
switchport mode trunk
spanning-tree port type network
vpc peer-link
exit

[Daniel Kemper]

Hi,

did you already got a solution for this problem? I have the same issue with actual NX-OS release 10.2.5(M) and the same hardware.

I already managed to fix the vPC by setting speed on the port channel to 100000. Now one of the ports comes up, but the other one is not able to start.

[katerina.dardoufa]

Hi,

I am having the same issue. C93108TC-FX3P release 10.4(1).

I have connected 4 QSFP-100G-CR4, 3 come up, but the last one Eth1/54 refuses to obey!!!

I have tried the following:

• Different cable
• Different port
• manually configure speed and duplex settings

nothing seems to work. Any other ideas? I wouldn't like to downgrade to 40G, but maybe I can try this...

Eth1/51 ** LINK TO MK-N9K- connected trunk full 100G QSFP-100G-CR4
Eth1/52 ** LINK TO MK-N9K- connected trunk full 100G QSFP-100G-CR4
Eth1/53 ** LINK TO MK-N9K- connected trunk full 100G QSFP-100G-CR4
Eth1/54 ** LINK TO MK-N9K- notconnec trunk auto auto QSFP-100G-CR4

[Daniel Kemper · Accepted solution]

Hi, we fixed this by setting fixed speed on the port channel. The not upcoming ports mentioned in my post were not correctly connected by our customer.

[herve.leon]

Thanks. I also fixed the speed of the two links. 

[katerina.dardoufa]

Hi again!

In my case port eth1/54 would not function with any combination (speed/new cable/different SFP etc). I resulted to moving the **bleep** cables to other ports (1/49 & 1/50), setting the speed settings on port-channel and running "channel-group x force mode active" on the interfaces. This resolved the issue. Greatful for all the ideas listed in this topic.