Denver Gingerich <den...@ossguy.com> wrote:
>> It turns out that the switch in question is a process by which it
>> turns itself off momentarily to change to the correct vlan. This is a
>> local, IETF hackthon setting. This cuts the PoE power, and so a cycle
>> ensues when depending upon PoE only.
>>
>> There is still a bug though, because the loss of PoE power seems to
>> reset the One even though there is USB power present.
> To be clear, would you want it to always use whichever power source was
> plugged in first?
That would an option, but I think it would be surprising to many people.
By "plugged in", you would really mean which source provides power first.
(My USB was plugged in. But I had turned the switch on the power bar off,
because I wanted my serial console ready. I didn't expect PoE to be available)
> Or are you hoping for some kind of UPS-style
> automatic switch-over when both are connected and one power source
> disappears?
I think this probably makes more sense for most people and would match
people's expectations better. However, I recognize that PoE requires some
time (maybe a few dozen ms) to activate, so switching from USB to PoE
probably would not work.
So I'd conclude: if there is USB power, it should be used, and PoE should NOT
be negotiated. That wouldn't have helped me, since I had USB turned off, and
the PoE was unexpected, but at least, when I disconnected the PoE, it would
have stuck to the USB.
I think that when the PoE got negotiated, it used that, even though USB was
also present, and when the PoE bounced, the router bounced.
> I'm not sure how complicated the latter would be, but it's helpful to
> know the expected use case so we can investigate the hardware options
> accordingly (for the One, Two, or otherwise).
Whatever decision: I guess just document it :-)
signature.asc
Description: PGP signature
_______________________________________________ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
