[Nut-upsuser] MGE Evolution and programmable outlets configuration
Arjen de Korte
nut+users at de-korte.org
Sat Dec 6 16:37:27 UTC 2008
Citeren Marco Chiappero <marco at absence.it>:
> First of all I think I have to use upssched for shutting down the
> computers, both connected to the programmable outlets, haven't I?
Yes.
> But
> how does the driver communicate to the computers that outlet.X is
> cutting the power since is reaching the
> outlet.X.autoswitch.charge.low/outlet.X.delay.shutdown condition? Does
> it cut the power without any notice? And how the computers get aware
> from witch outlet they are drawing power?
At the moment we're working on making switchable outlets more useful.
For now, you'll indeed have to setup 'upssched' to shutdown the
computers. After shutting down, it makes little difference to also cut
power to their outlets. The difference in runtime for the UPS will be
insignificant in this case.
Where switchable outlets *are* useful, is when power is restored.
You'll need to cycle the power to restart the systems up again. Here
having multiple switched outputs is an advantage, since that allows
you to only cycle power on the outlets where needed. This is all still
in development and frankly speaking, I don't know when we'll have
something available that is production quality.
> Without these information it
> sounds that programmable outlets are almost unuseful and just marketing.
See above.
> I mean, if I have to set up a counter in my computers for starting the
> halt sequence it does not make any difference the outlet it is connected to.
> Then, since I *do* not want the server to shutdown the UPS, should I run
> upsmon as slave on both the PC and the server?
For the moment, yes. It will also mean that if power returns before
the UPS batteries are empty, the servers won't restart automatically.
For the moment, there is nothing we can do about that.
Best regards, Arjen
--
Please keep list traffic on the list
More information about the Nut-upsuser
mailing list