Messages

Update to this.  Errors came back, I dug into the logs and found out they were coming from the IPv6 network.  I turned off my IPv6 setup and since then I haven't had any errors.  I probably misconfigured something but that will have to wait till I get a better understanding of setting up IPv6.  That said the large error number now makes sense as that is pretty close to a /64 subnet size.

In gigabit ethernet and beyond "auto" is actually a separate mode, not a "let's try what we have and what the other side can do and use the highest setting possible" like with e.g. an old school modem.

So if one side is set to a fixed speed and duplex setting depending on the implementation it will silently ignore autonegotiation requests. If the other side is standard compliant and set to "auto" and it does not get any response from the side using the fixed setting, it will drop to 10 Mbit/s half duplex.

So either both sides "auto" or both sides fixed but never mix!

This is great to know. Thanks!

That's ambitious ;-)

😂 ... its a work in progress. 

Any sources of interference? Have you tried different patch cables? Can you bring the Unifi next to the DEC and use a pre-made CAT6A cable?

It seems switching the connection from 10Gbe to Auto was the culprit.  Its still connected as 10Gbe but I'm assuming it's using a different mode.  So far no errors today.

Check your physical connections and make sure you're not doing jumbo frames on one side only.

Are you using fibre or copper for 10Gbe?

Everything is set to 1500.  My fiber to SFP is also set to 1500.  For the 10Gbe I'm using copper.  On my unifi switch I did notice the connection to the router was set at 10Gbe instead of Auto Negotiate, which I just switched but I don't think thats going to make an issue here.

I'm using opnsense with the DEC3862.  On my LAN I've attached the ax1 to my Unifi Switch via a Unifi Uplink SFP cable which can negotiate between 1-25G.  Between the opnsense router and the switch it's set at 10G.  However, and only, on my LAN I'm getting these ridiculously high error outs in the quintillions and thats one day after a reboot.  Are these related to pause frames for the flow control?  How can I fix this?