Just to double check terminology - I thought "patch cable" referred to the short cables used in a network room to interconnect switches with patch panels. I thought they were wired differently to the standard network cables that connect your PC to the RJ45 wall socket (or at least they can be depending on which "standard" of wiring is chosen (eg T568A or T568B). (see here). Now that the cabling path has been changed it is possible that there may have been an inadvertent swap of transmit and receive pairs.Snail wrote:I have been running an early version of Sulu2 on my old HP desktop. Until a few days ago, I was connected to my gigabit-ethernet-capable router by a long patch cable.
Most modern devices have the ability to autoswap their Tx and Rx pairs "on-the-fly" so maybe the router is now having to do that for your PC where it did not have to do it before. Maybe this could introduce timing alterations that are affecting the connection process.
It might be worth using a multimeter to verify that the end-to-end wiring using the new method is exactly identical to what it was before. If it is pin-to-pin identical then the issue must be due to cable length or resistance.
Unless of course the change has introduced different electronic hardware - are you still connecting to the same router or has that hardware been swapped as part of your ISP deal?
- and is there now a hub or switch that was not previously in the data path to the PC?
- how is the house wiring distributed? Is there a single cable from the router feeding a central electronic hub/switch that splits off to individual room wiring, or do you have a "house-wiring breakout panel" next to the router with individual wires from each patch connector to each port on the router?
.
.
This sounds more like a kernel problem directly effecting your hardware !
