On Thursday, April 18, 2019 at 7:03:40 PM UTC-7, Greg Troxel wrote: > On NetBSD, by default, in the absence of a hardware clock, the system > clock is initialized to the last sync time of the root filesystem. I > gather that this is similar behavior to what fake-hwclock does. But, I > thought that in general waiting until the clock was known to be right > was good strategy. > > Sure. You could alternately hard-set the date to some ancient time as part of your bootup sequence. That would let the normal weewx 'is the date newer than the conf file' mechanisms work as-is, although I don't know what it would do to other stuff on your box.
(Also, I don't understand why bad times are a problem, because it seems > that archive records from the hardware have hardware timestamps. But, > with a 1.5h power outage, I experienced what I think was recording those > records at the wrong time.) > > Hardware doesn't emit archive records as far as I know. Weewx aggregates things periodically 'into' archive records from the measurements it received in that time period. And everything is based on the clock. You would probably get fascinatingly bad archive data and graphs if the clock is moving wildly (example - a decade of people complaining about DST transitions) Bottom line is always work off a good clock, no matter how you get it good. -- You received this message because you are subscribed to the Google Groups "weewx-user" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
