Thanks Charles for the clarification.

If NUT cannot control it then removing it would, IMO add value.

On Sun, 10 Mar 2024 at 00:14, Charles Lepple <
[email protected]> wrote:

> On Mar 9, 2024, at 5:16 PM, [email protected] wrote:
> >
> > The posting at
> https://github.com/networkupstools/nut/issues/432#issuecomment-405371395
> references ups.delay.start and suggests setting it to -1.
> >
> > upsrw -l pve1@localhost listed this as modifiable but any attempt at
> modifying the value failed.  I used the override.ups.delay.start = -1 in
> nut.conf and this is now being reported.
>
> My original suggestion (which implies the use of override.*) is mainly for
> setting values that the UPS driver uses to decide when to shut down (if it
> is told to ignore the UPS LB flag, and must calculate the LB condition in
> the driver).
>
> ups.delay.start is a value that is only interpreted by the UPS hardware
> (the driver just passes it through to upsd/upsc when override.* is not in
> play). So IMHO it doesn't make sense to override it, unless you have
> something else that needs to read it. In most cases, the startup delay
> happens when NUT is not running yet (assuming the machine with upsd/driver
> is powered by that UPS).
>
> If the UPS won't let you change a delay variable, then I don't think
> override is going to help here.
>
> > Two observations are that once a variable setting is overridden, upsrw
> -w will not show these values.  Is it by design?
> >
> Others will have to speak to the design rationale of override.* and
> default.*, but I guess I tend to think of override.* as hard-coding a value
> in the driver (and therefore something that I wouldn't attempt to change
> with upsrw, but with the conf file instead). I guess that probably means
> the variable should be removed from the list of read/write variables.

_______________________________________________
Nut-upsuser mailing list
[email protected]
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/nut-upsuser

Reply via email to