[Nut-upsdev] [Powerman-discuss] Joining forces with the Network UPS Tools

[Arnaud Quette]

2008/10/6 Jim Garlick <garlick at llnl.gov>:
> On Mon, Oct 06, 2008 at 03:43:38PM +0200, Arnaud Quette wrote:
> (snip)
>> To conclude this first "meeting", the best would be:
>> - to have a common ground for comparison (ie having 1 same device
>> supported by both, and check how things works on each side),
>> - possibly try to make a quick and dirty bridge between nut and
>> powerman to see if it make sense (is there a client API to bind with
>> Powerman?),
>
> There is a text-based protocol between client and server that looks very
> much like a power controller (in fact the powerman daemon can control
> a remote powerman daemon as though it were a power controller).
> So perhaps this fits into an infrastructure you already have for
> talking to devices on TCP sockets?  Alternatively, it would not take too
> much arm twisting to convince me to refactor the client code so that a
> C based API is exposed for external use.

it shouldn't be that hard to spawn a powerman driver in nut, using
libpowerman and possibly a small bunch of doc/help ;-)

the only problem ATM is that a driver only expose 1 device.

>> - in the meantime, check how hard it would be to merge Powerman's
>> devices into NUT.
>>
>> I'll also check on my side your simulators, since I don't own any of
>> the Powerman's supported devices...
>>
>> Does all this make sense for you?
>
> Sure, though I would need to see a very compelling benefit to powerman
> users to consider a project merge.  I don't see that based on my limited
> understanding of NUT.  However, I'm very open to the other forms of
> collaboration discussed thus far.

so, let's go that way for the moment, and we'll see later on if this
can/should be improved.

On my side, I'm about to release a preliminary branch for supporting
Eaton SNMP PDUs (1 on 4 ATM).
I'll send a more detailed mail soon about that.

Arnaud