Saturday, July 11, 2020

Troubleshooting a 10GBase-T Link

Hi all,

I have a Netgear M4300 connected to a Netgear M4200 via 2x 10Gtek ASF-10G-T SFP's. The link is CAT6A UTP and is only 45ft. It isn't shielded but I do not believe that is the problem here. It's also not near any common interference sources.

The connection is stable *most* of the time. Occasionally, I will notice devices connected to the M4200 are offline. When I pull up the Monitoring page of the M4300, it will show hundreds of Link Down Events and it will count up at about one additional per second. All it takes to fix the problem is pulling either end of the CAT6A and replugging, or rebooting either switch. It's not a loose connection or a bad termination. It will work again stable for up to a day, and then act up again.

Th M4300 logs "SFP interrupt received on the unit" every time this happens. What does this error mean?

Given how much it DOES work, it doesn't seem like a physical wiring problem. The link works fine on two gigabit ports.

Right now I am suspecting one of two things: 1) a bad SFP module, or 2) one of the SFP modules is overheating? I took out the one from the M4200 and it was VERY hot.

The setup is used for AV production multicast video (NDI) and the link is actively delivering around 800 Mbps continually. I'm not sure if throughput increases a module's likeliness to fail. But, 800 Mbps on a 10G link shouldn't be breaking a sweat, right? I understand that link speed is not the same as throughput.

Before it is mentioned, fiber is not a good idea for this setup. It will be connected and disconnected several times a week and I don't want the client to have to clean fiber ports every time. Also, industrial/tactical fiber connections like Neutrik OpticalCON are pricey.

Any pointers for troubleshooting this would be great.



No comments:

Post a Comment