CU v7.3.2 Upgrade Problems with Extra Sensors

[HansR]

OK. Looked at the logfiles.

I have the following questions:

1. Does it have to be 1 hour for a cycle? Can't you reduce that to 10 minutes or so?
2. You can create a 'sensor gone' situation by disabling it in the app. Then restart CMX so it understands the sensor is gone

I'll try to replicate the failing upload situation over here if succesful I'll work from there.

If not successfull I will ask you to just disable the soilmoisture sensor and when the upload fails leave it running failing. It may run until tomorrow morning, the result does not have to be tonight.
Than send me:

1. The log for the useraskeddata run
2. The ExtraLog202501.txt (the datafile)
3. The corresponding json: the erroneous one which must be left in the utils directory and the last correct one which you will find on the server.

If there is no upload fail, I do not need anything.

I just disabled my soilmoisture sensor and will see what I get (must be similar to you)
I did point to the additional problem of the extra sensor logging you have a pretty complex data log so it is difficult for me to see if something like that happens to you. Are you able to simplify the sensor situation by running a second instance and a beta test site to reduce the complexity?

We will now try to reduce the number of parameters for the error so do not change other sensors or intervals or whatever. Just one sensor disable and we'll see what it will bring us.

The most surprising thing is that when you disable the sensor the first datafile is OK but the second fails. I do not understand that.

[HansR]

Don't worry: I have a failed send as well, I'll work from there.

I Found the cause, now to understand it is another issue.
Tomorrow the next steps.

[lse-cumulus]

hi hans,
started a new test run and can reproduce the issue.

disabled soilmoisture sensor in ecowitt app (really easy way), restarted CMX, had null values in ExtraLog,
started the userAskedData runs manually to shorten the intervalls.

1st UserAskedData: ok, no problems with null values, upload.php ok, Extra Charts still working.

2nd UserAskedData logfile says: No valid JSON data in the received data; BUT i cannot find any extrasensorsdata.json on utils directory, and extrasensorsdata.json file on my server is small/corrupted, Extra Charts stopped working;

3rd UserAskedData: problem with 'No valid JSON data in the received' data data remains.

I will keep it running (now hourly x:05) in that way until morning before I enable sensors again ...

If you need the logfiles, just let me know, but I understood that in parallel you could also reproduce the problem with your sensors at your site.

[HansR]

[...]
If you need the logfiles, just let me know, but I understood that in parallel you could also reproduce the problem with your sensors at your site.

Hij Leo, OK. I now understand the problem, I will make a fix and with the new beta we start testing again.

[HansR]

See release 7.3.3 beta 5

[HansR]

Can you give me the URL of the extrasensors JSON file so that I can inspect online?

I note that I started inspecting around 14h30 and I saw a line. Then the chart completely disappeared and I could not find the JSON. That should not be possible unless you deleted it.

[HansR]

OK. Everything is back...

[lse-cumulus]

Hi Hans,
thx for your efforts.
I upgraded to CU 7.3.3. beta 5. Now all is running fine (sensors are all enabled again).
[I had some internet connectivity issues with my network provider in between, that effected the CMX PC that could not upload to my weberserver and could not post to the forum before]

[HansR]

Ok. But the moisture sensors have a value 0. What is the status there?

[lse-cumulus]

Soil Moisture Sensors were outdoor in frozen earth. Brought them inside to get proper connection again to Ecowitt GW2000 (but that dropped reported value down to 0). So issue is from sensors and Ecowitt (actually 0 is the value that is reported), and has nothing to do with CMX and CU. Usually takes some time until soilmoisture sensors adapt to frozen outside earth again ...

[HansR]

Ok. Thnx.

So we agree its ok now?

[lse-cumulus]

so far so good; just made UserAskedData and sensor show values >0 in sync with ecowitt.