On 13/06/2020 21:11, Richard Kettlewell wrote:
> Adrian  writes:
>> So the only thing in this that I can see is unique is the source
>> machine.
>
> Cable and switch port are presumably unique too, and easy to rule out.

I've just found a faulty cable on the to the Raspberry Pi attached to
the bedroom TV, used for web browsing and media playing. I've just
upgraded it from a Pi 3B+ to a Pi 4B, and everything seemed to be ok.

However I ran a speed test against from another machine, accessing a
USB3 HD on the Pi, and while it was managing 86MB/s down, it was only
doing about 5MB/s up. After replacing the cable it did 86MB/s down and
92MB/s up.

As it was normally only used to consume data, and the return path wasn't
completely broken, this hadn't been noticed up to now.

---druck

--- SoupGate-Win32 v1.05