[Nut-upsuser] Emne: Eaton Powerware 5110 - some stats not reported

Arnaud Quette aquette.dev at gmail.com
Sat Feb 15 21:16:50 UTC 2014


2014-02-15 20:07 GMT+01:00 Charles Lepple <clepple at gmail.com>:

> On Feb 15, 2014, at 12:54 PM, Arnaud Quette wrote:
>
> > 2014-02-13 6:55 GMT+01:00 Alf Høgemark <alf at i100.no>:
> >> Hi
> >
> > Hi Alf,
> >
> >> On
> >> http://nutwiki.kanonbra.com/wiki/Category:Eaton_Powerware_5110
> >
> > cool thing!
> > But duplicating the NUT Device Dumps Library [1].
> > The DDL can also serve both users and developers purposes, by providing
> these dumps as .dev files (usable with dummy-ups).
> > I'm looking for someone to help me completing this effort: would you be
> interested in?
> > It would mostly be collecting dumps posted on the lists and the web, and
> calling users to massively send theirs...
>
> Arnaud,
>
> here's the discussion that led to Alf's tests with MediaWiki:
>
> http://news.gmane.org/find-root.php?message_id=51D93E42.60500%40i100.no
>
> (Note "Mike."'s mention of read/write variables, which are not addressed
> by upsc output alone.)
>

thanks for pointing this to me Charles.
no surprise, I've completely missed it!

read/write vars and instant commands are still the weak point, since this
effort was spawned from the dummy-ups effort...
I've also created another tracker for that part:
https://github.com/networkupstools/nut/issues/94

The way I see it, there are several use cases that overlap a bit:
>
> * dummy-ups data files for debugging
>

   * clients development
   * QA regression testing (whole NUT framework)
(as in Ubuntu), currently use the default static device from dummy-ups,we
could there also imagine
   * device non regression testing
 by testing such outputs


> * Answering the question "what variables do I get from this UPS with NUT
> vX.Y.Z?" (or, "what is the benefit of upgrading the driver from the
> distribution's old version?")
> * Answering the question "what UPS should I buy to be able to monitor
> these variables?"
>

indeed.


> The thought was that tagging pages with something like MediaWiki
> categories could build up the cross-references needed to answer the second
> and third questions, but that might get unwieldy. With something like the
> DDL, it should be possible to programmatically generate those
> cross-reference lists.
>

that's also what I'm thinking, with github tasks 3 and 4:

   -  Create a dedicated repository and website (
   http://devdumps.networkupstools.org)?
   -  Create a better page, or set of pages, to document / search / submit
   dumps ( @zykh <https://github.com/zykh>, would you be able to help?)


Organizing the information in something like a wiki allows additional
> comments from users of the equipment. Of particular note is whether or not
> a given variable can be relied upon (which might depend on the NUT driver
> version, or even the UPS firmware version).
>
> Unfortunately, it isn't clear what the best way forward would be.
> Shoe-horning the second and third use cases into MediaWiki would probably
> require a custom plugin.
>

a dedicated website with some JS filtering / searching, as in the HCL,
could probably do it (hem, I'm unsure that it can scale however).
maybe some DB with pre compilation would be needed (manifest, indexing,
...).
but that's a bigger project.

a tool would also be needed to drive collecting data on the user's system,
à la "nut-recorder"


> Another set of useful metadata is the USB VID:PID, plus dumps from
> "lsusb". This overlaps nut-scanner a bit, but considering that lsusb is a
> much lighter-weight dependency than nut-scanner (and it's even in the
> FreeBSD ports tree now), it's a quick lookup for users to determine whether
> it is worth their time to try installing NUT.
>

that may be an addressable point the nut-scanner Gen2 and weak deps:
https://github.com/networkupstools/nut/issues/14


> > This however made me realize that it's still not referenced on the web
> and in the docs.
>
> helps if you link it from somewhere :-)
>

sure ;)
in my defense, it goes back around the time I disappeared...
iirc, I only announced it on the mailing list (not sure which one however)


> update: I see you're working on that:
> https://github.com/networkupstools/nut/issues/93
>

yup, started to track and share the effort, and finally dump things that my
little brain was holding for far too long...

cheers,
Arno
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/nut-upsuser/attachments/20140215/339b8825/attachment.html>


More information about the Nut-upsuser mailing list