Nut upsdev - Feb 2012 - How to cause a local system shutdown...

Hi all.  I'm working on an update to the ONEAC driver and have a 
request from a user that I am not sure how to handle.

The UPS can be communicated with through a couple different methods 
(e.g. Serial, network).   If a UPS shutdown command comes through the 
network interface, the user wants NUT (serial connection) to signal 
the supported systems to shutdown.

What is the "right" way to do this?  One thought was to trigger the 
LB status if the shutdown counter was ever found to be 
non-zero.  Would that work?  What other options should be considered?

Thanks,

Bill

On Feb 9, 2012, at 5:44 PM, William R. Elliot wrote:
> Hi all.  I'm working on an update to the ONEAC driver and have a
request from a user that I am not sure how to handle.
> 
> The UPS can be communicated with through a couple different methods (e.g.
Serial, network).   If a UPS shutdown command comes through the network
interface, the user wants NUT (serial connection) to signal the supported
systems to shutdown.
> 
> What is the "right" way to do this?  One thought was to trigger
the LB status if the shutdown counter was ever found to be non-zero.  Would that
work?  What other options should be considered?

Hi Bill,

I'll be honest - all of my NUT experience has been with UPS hardware that
only support a single active control connection (and all of the fan-out is
handled by upsd).

That said, I think a slightly different status flag (FSD) would be the right way
to go:

   http://www.networkupstools.org/docs/developer-guide.chunked/ar01s09.html#_fsd

That links to a chapter on the network protocol, where an appropriately
privileged NUT client can set the FSD flag in upsd's status. But based on
the commentary there, it seems to be equivalent to the on-battery/low-battery
status-- without the minor confusion of seeing "low battery" messages
in the logs (when the battery might not actually be low).

There is additional code in the server to reflect FSD to other clients in the
status (since the server-side flag is kept in a separate per-UPS variable), but
I don't think it would conflict with using it in a driver.

Hope this helps. Feel free to post patches if you have any questions about
implementing this.

-- 
Charles Lepple
clepple at gmail