[Nut-upsdev] Re: [Nut-upsuser] Ablerex 625L USB version

Carlos Rodrigues carlos.efr at mail.telepac.pt
Thu Feb 1 01:56:12 CET 2007


On 1/31/07, Alexander I. Gordeev <lasaine at lvk.cs.msu.su> wrote:
> 2. As you see in the output 2 other prolems exist:
>
> Asking for UPS status ("Q1" command)...
> UPS status: 229.0 165.0 229.0 000 50.2 13.6 00.0 00001000
> Unsupported battery voltage (11.5V).
> This UPS has an unsupported combination of battery voltage/number of batteries.
>
> Is this ok?


This means there is no data for that in the driver. I can add it, but
I need the "battery.voltage" for a full battery and the
"battery.voltage" for an empty battery (the voltage at the moment the
UPS goes into low battery is good enough).


> and
>
> Asking for UPS status ("Q1" command)...
> UPS doesn't return any information about its status.
>
> This means UPS sent "UPS NO ACK". If I have upsd running it would print
> that "data is stale". I think, although this situation is quite
> normal for my UPS, it's not the normal behaviour for a driver, am I right?
> There are 2 ways, I guess: to repeat the previous successful response
> and to ask UPS once more without a delay. Which is prefered?


Repeating the previous successful response is covering up for stale
data. Repeating the command without a delay will probably fail again.
I think the correct behavior is to declare stale data until the next
poll (that's what stale means... "got nothing this time").

Usually a driver should hide this kind of weird behavior, but in NUT's
case, there is no point in doing that. Nothing will happen if the data
gets stale for one or two poll cycles, only after a while will that
have an effect (if the UPS was LB the last time it was polled
successfully, the machine will power off, if I'm not mistaken).

The only drawback is a message in the logs, but that's another thread
altogether.

If this happens a lot, the correct procedure is to set a larger
"pollinterval" in "ups.conf".

When using the serial port, stale data is very rare. But over USB
there is probably more stuff going in and out. Maybe the poll interval
should be bigger by default, like in "usb-hidups" (haven't checked how
this driver sets the poll interval, but if it can be set only when
using megatec-over-usb, it should be done).

-- 
Carlos Rodrigues



More information about the Nut-upsdev mailing list