[Nut-upsdev] MGE NMC and NutShutdownModule (and other stuff)

[Marco Chiappero]

Arjen de Korte ha scritto:
>> Sure, but a single driver could contain all the (vendor) specific 
>> logic without tainting the upsd code (starting from a netxml-ups clone 
>> and my patch).
> 
> Your proposed changes to the netxml-ups driver

I have never proposed changes to the netxml-ups driver, I proposed to 
include a simple NSM client in the mge-xml subdriver (single UPS 
configuration only), OR a totally new NSM driver dealing with just 
ALARMS and some other basic info (could easily poll data from many UPSes 
and support the same redundancy configurations suported by the Eaton/MGE 
software [1]). I simply asked for the better approach and inclusion 
chances. That's all I said from the beginning.
I liked also the idea of having the netxml-ups and, let's say, 
mgensm-ups drivers talking to each other side by side, but it looked too 
complex/unfeasible (that's why I asked you about how to realize 
communication).
That's all.

> can only monitor a single UPS

That's how the Eaton NSM works: if multiple UPSes are involved you list 
them, select the redundancy schema [2] and use a single virtual UPS as 
"global sum" of the UPSes involved [3]. Obviously the "sum" must be 
calculated accordingly to the schema (which is the difficult part) [4].
A possible sample configuration could be:

[globalups]
	driver = mgensm-ups
	port = parallel
	minimum = 1
	ups1.port = ups1.domain.com
	ups1.outlet = main
	ups2.port = ups2.domain.com
	ups2.outlet = 2
	...

As I said we assume that we are using only Eaton/MGE products, otherwise 
the NSM does not look the way to go. About machines with multiple PSU, 
if I'm not wrong NUT should already deal with them. Moreover you can use 
multiple mgensm-ups instances (eg: virtualups1 -> PSU1 , virtualups2 -> 
PSU2). I think that such a driver could complete NUT support for MGE 
products. The downside of the separate driver is just a more complex 
config in a single UPS-PSU scenario and the informations splitted into 
two UPSes.

> Not in the netxml-ups (or mge-xml subdriver). Given the above 
> difficulties, this *must* be a separate driver, because we already have 
> production systems that would break if this was added.

I can't see what would break but I'm for the separate driver too.

> I would consider 
> a NSM driver only, but without the parsing of additional XML data from 
> the UPS. It should only parse the ALARM messages, nothing more. For 
> logging purposes, one would then have to run a SNMP or XML enabled 
> driver on the side and use the NSM one to shutdown attached clients.

And this is what I proposed since the beginning (however some other xml 
informations have to be parsed once a while).
So, at the end, is it interesting or not? May I spend some time on this 
in the future? Or am I going to waste my time?


Regards,
Marco

[1] http://i34.tinypic.com/2ynms9h.png
[2] http://i34.tinypic.com/2q04sbk.png
[3] http://i33.tinypic.com/xm3hg9.png
[4] http://i36.tinypic.com/vg07e9.png