Page 1 of 3

b3182

Posted: Sat 30 Apr 2022 8:44 pm
by HansR
This version does not seem to give valid winddata anymore.
And it appears to be worse than that.
Device: Ecowitt WS80 / http protocol

logfile attached.

Re: b3182

Posted: Sat 30 Apr 2022 9:19 pm
by mcrossley
Odd as the Ecowitt protocol has not been changed for the last few betas...

Your device is not sending wind data, outdoor temperature or humidity.

The rain I will have to check when I get back to my computer. The value in the response (12.787 inches) matches what your CMX is expecting.

You could try switching the Local API station type and see if that is working.

Are your sensors still registered and the data is showing in the phone app?

Re: b3182

Posted: Sun 01 May 2022 1:26 am
by BCJKiwi
Hi, Have Davis VP2 on Win 10

Build 3182
In stationsettings.html
Data Series Visibility
Temperature Data
and
Humidity Data

When changing Indoor temperature to ticked ( or unticking )
or
When changing Indoor Humidity to ticked ( or unticking )

When clicking Save Settings, this message is displayed:-
Error: 500(error) - Error processing Ecowitt sensor mapping: Object reference not set to an instance of an object.
However Save Settings still works.

I do NOT have an Ecowitt Station.

Re: b3182

Posted: Sun 01 May 2022 3:18 am
by HansR
mcrossley wrote: Sat 30 Apr 2022 9:19 pm Odd as the Ecowitt protocol has not been changed for the last few betas...

Your device is not sending wind data, outdoor temperature or humidity.

The rain I will have to check when I get back to my computer. The value in the response (12.787 inches) matches what your CMX is expecting.

You could try switching the Local API station type and see if that is working.

Are your sensors still registered and the data is showing in the phone app?
Everything on the Ecowitt side is fine, running and delivering data.

Please note that everything was fine, CMX gathering data etc... Just before the update.
I shutdown CMX, copied the new CMX in place and started CMX.
No changes on my side.

Note: not sure what rain value that should be... Yearly?
And why inches when working in mm.

I am not changing to local API device, it is the HTTP I am working with. If that is not wanted it should not be available.

Re: b3182

Posted: Sun 01 May 2022 7:14 am
by AndyKF650
I downloaded the ver B3182, stopped CMX, extracted the new version and rebooted the RPi to get things started again.

CMX appeared to start automatically, (systemctl auto start is in place) but it stalled halfway.

I stopped CMX and then restarted using systemctl stop/start and all now seems fine.

For completeness I ran CUtils website with no problems, all my customisations remain in place and working.

My setup is a Davis console running through a Meteo-Pi logger to a RPi.

Re: b3182

Posted: Sun 01 May 2022 7:47 am
by Buford T. Justice
I also cannot get 3182 to work with my Ecowitt GW1000 setup. I reinstalled 3173 and all is well again.

Re: b3182

Posted: Sun 01 May 2022 7:56 am
by HansR
@AndyKF650, @Buford T. Justice : Please post the corresponding logfiles!

Re: b3182

Posted: Sun 01 May 2022 8:08 am
by AndyKF650
Hi Hans

As requested here is my current working log file for info. The MySQL errors are being worked on, this is my current development project.

Re: b3182

Posted: Sun 01 May 2022 8:17 am
by HansR
Thanks @Andy, but that is the log where all was OK (apart from the SQL).
Do you still have the log where it stalled?

Re: b3182

Posted: Sun 01 May 2022 8:28 am
by AndyKF650
Unfortunately no Hans! All cleaned up and tidied away since the system was going well. Next time maybe?

Re: b3182

Posted: Sun 01 May 2022 9:18 am
by mcrossley
HansR wrote: Sun 01 May 2022 3:18 am Everything on the Ecowitt side is fine, running and delivering data.
Ok, so if the apps work, then the local API is working.

All I can suggest is you reboot your gateway and see if that fixes it.
HansR wrote: Sun 01 May 2022 3:18 am Please note that everything was fine, CMX gathering data etc... Just before the update.
I shutdown CMX, copied the new CMX in place and started CMX.
No changes on my side.
Maybe, but your log shows that your station is just not sending the data. :(
HansR wrote: Sun 01 May 2022 3:18 am Note: not sure what rain value that should be... Yearly?
And why inches when working in mm.
CMX uses two values the rainratein, and yearlyrainin.
Inches because Ecowitt chose to use imperial units on the HTTP protocol.

All I can think of is your rain sensor is set to use the piezo sensor in Station Settings | Ecowitt Sensor Maps?
HansR wrote: Sun 01 May 2022 3:18 am I am not changing to local API device, it is the HTTP I am working with. If that is not wanted it should not be available.
OK, your choice, I just suggested it as a test, but if the Ecowitt app is working the Local API is working.
I recommend using the local API over the HTTP as it offers more data, more timely. But if you have something against using it I really don't mind :)

The HTTP station type is offered as an alternative because not all Ecowitt stations have the local API. Also some people may already be using the local API for other software and do not wish to overload their gateway.

Re: b3182

Posted: Sun 01 May 2022 9:20 am
by mcrossley
BCJKiwi wrote: Sun 01 May 2022 1:26 am Hi, Have Davis VP2 on Win 10

...

When clicking Save Settings, this message is displayed:-
Error: 500(error) - Error processing Ecowitt sensor mapping: Object reference not set to an instance of an object.
However Save Settings still works.
Thanks, I've had another report of this.

The error message is benign, as you have found the other settings will be saved correctly.
The error will not occur if your station type is set to either of the Ecowitt stations.

Fixed for the next patch release.

Re: b3182

Posted: Sun 01 May 2022 9:29 am
by mcrossley
AndyKF650 wrote: Sun 01 May 2022 8:28 am Unfortunately no Hans! All cleaned up and tidied away since the system was going well. Next time maybe?
Not much I can do without any logs :(

Re: b3182

Posted: Sun 01 May 2022 9:42 am
by HansR
mcrossley wrote: Sun 01 May 2022 9:18 am Ok, so if the apps work, then the local API is working.

All I can suggest is you reboot your gateway and see if that fixes it.
OK. Rebooted (through the app) and that seems to work.
Could the cause of the failure be that a transmission was cut-off during the stopping of CMX?

I asked this before, but would it not be a good idea to reboot the GW1x00 from CMX (if possible) at error situations? It reboots within 20 seconds!
mcrossley wrote: Sun 01 May 2022 9:18 am CMX uses two values the rainratein, and yearlyrainin.
Inches because Ecowitt chose to use imperial units on the HTTP protocol.

All I can think of is your rain sensor is set to use the piezo sensor in Station Settings | Ecowitt Sensor Maps?
No piezo sensor mapping. I think just the prior point: the GW1100 was confused.
mcrossley wrote: Sun 01 May 2022 9:18 am OK, your choice, I just suggested it as a test, but if the Ecowitt app is working the Local API is working.
I recommend using the local API over the HTTP as it offers more data, more timely. But if you have something against using it I really don't mind :)

The HTTP station type is offered as an alternative because not all Ecowitt stations have the local API. Also some people may already be using the local API for other software and do not wish to overload their gateway.
I will think about this. There is no hard reason for me to stick with HTTP but habit and that I am OK with it. Just the last time I switched I ran into other problems with double data at some point. Anyway... irrelevant for the current issue apparently.

Problem solved.
Next time I will reboot before I signal an issue but still think an automatic reboot originated by CMX might solve this type of problems.

Re: b3182

Posted: Sun 01 May 2022 9:54 am
by HansR
In addition to my prev post and possibly related is the following chart where apparently there appeared to be missing some temperature (after I shutdown CMX and gave up until the reboot described above). the GW1100 seems to recover itself at 23h15 :groan:

And beside that: during catch up FeelsLike and WindChill apparently are calculated but the Dewpoint is not, that only starts showing values when the catchup is completed at 11h30 just now. Humidity is not shown either but that is not in this chart.

EDIT: and it probably is the absence of Humidity which causes the dewpoint to be absent.
    Schermafbeelding 2022-05-01 114517.jpg