meteohub.de

I'm currently running MeteoHub (v5.0i) on an RPI and am uploading weather data to HAMWeather (PWS Weather).

In reviewing a daily total rainfall issue in their data, I noticed that they had recorded a value shortly after midnight that represented the previous days total rainfall (i..e should have been ZERO but was not). They kindly provided the errant post and the post immediately following which appear as follows:

Sent at 12:02 AM MST (6:02 AM GMT)…

76.187.xx.xxx - - [09/Mar/2016:06:02:02 +0000] "GET /pwsupdate/pwsupdate.php?act
ion=updateraw&ID=HLFRISCOTX&PASSWORD=xxxxxxx@m&dateutc=2016-03-09%2006:02:02&win
ddir=28&windspeedmph=3.4&windgustmph=4.3&humidity=94&tempf=59.4&rainin=0.00&dail
yrainin=1.15&monthrainin=1.20&yearrainin=3.88&baromin=29.80&dewptf=57.6&UV=0.0&s
olarradiation=0&softwaretype=meteohub HTTP/1.1" 200 131 "-" "meteohub"

Then at the next send, at 12:07 AM MST (6:07 AM GMT)…

76.187.xx.xxx - - [09/Mar/2016:06:07:02 +0000] "GET /pwsupdate/pwsupdate.php?act
ion=updateraw&ID=HLFRISCOTX&PASSWORD=xxxxxxxx@m&dateutc=2016-03-09%2006:07:02&win
ddir=32&windspeedmph=3.1&windgustmph=3.1&humidity=94&tempf=59.4&rainin=0.00&dail
yrainin=0.00&monthrainin=1.20&yearrainin=3.88&baromin=29.79&dewptf=57.6&UV=0.0&s
olarradiation=0&softwaretype=meteohub HTTP/1.1" 200 131 "-" "meteohub"

Is this a known problem and/or has it been fixed in a release after 5.0i?

Thanks
mark

please check your time zone settings and that you are working with local timezone for evaluations.

I'll review them for accuracy, but I believe that this is an issue that cannot influenced by time zone settings as the reset occurred within an hour -- not across an hour boundary.

It's clear from the data that was posted that the Daily Rain (and perhaps other daily-based counters -- have not attempt to verify those) was reset to zero sometime between xx:02 and xx:07 -- it does not matter what xx is based on timezone settings -- it should have been reset to zero at xx:00 and was not...

Processes that compute daily rain are started very few minutes, so a delay in the magnitude you
describe can happen. By that at 0:04 values from last day might still be there. Calculations
are done correctly but may be a bit later than expected.

Thanks for clarifying the process/schedule of how those values get reset.

With that in mind, I would rather not publish bogus values to weather services during that potential windows of having values from the previous day.

How might I go about automatically (i.e. time based) turning off all the exports for a period of time (say between 23:59 and 00:10) and then again automatically turning them back on again to avoid posting old/stale day-count values?

Any suggestions on this issue and how to better address it or is the weather network export process not extensible/flexible in this way?

I am not aware of any problems any weather network does have when this data (when used at all by the network) is lagging behind a few minutes.