meteohub.de

Hi everybody,

I have a Meteobridge 5.2 running on a TL-MR3020 and have been observing strange behavior for a few days now.
I can see on the raw sensor data sheet that all sensors are working properly and the data is updating every few seconds.
The upload status for all services is also green, but, and this is my problem, the services are actually no longer being updated.
There is no upload to Meteoplug, Windfinder and also the HTTP and SFTP uploads are not working. When I run a test on the service settings page the test file is uploaded correctly, but running a test of one of the configured services does not work.

Has anyone ever noticed similar behavior? I am thankful for every help.

Thanx, Gunnar

One more thing: when I do a reboot, the services are triggered once, but then never again.

Has anybody already a fix for this?

I have the same or similar issue since a couple of weeks.
I am using a Meteobridge PRO with version 5.2 from. 11.01.2022 build 14655 FW 1.4. My uploads to Windfinder, Windguru and Windy are configured to upload every 5 minutes, skipping any retries but those stopped working properly. After a power cycle, the Meteobridge Pro comes up and transmits weather data for approximately 6 hours. After that no data is transferred to any service anymore. However there is still a green check mark for each service in the web interface although the last upload is already way older than the configured 5 minute interval. In case I increase the repeat interval, the 6 hours are significantly reduced. Moreover from the charts, I can see that there is valid data received by the Station itself, so it has something to do with the uploading service, not being triggered for some reason :-/

Yes sounds like something in the uploading, ISP and/or internet/router connection/status or the recipient server/s have had some changes? Multiple recipient destinations refusing data may tend to indicate more an ISP/router type issue?

If you have defined retries on certain uploads and the target URL is not responding or slow then jobs will queue up. The buffer usage percentage on system info tab will then tell someting about 100%. Is that the case in your setup?

At the moment it looks quite fine and the station is running since the last reboot yesterday evening around 23:00.
At the moment all the upload to the weather services are skipping the retries.
Nevertheless, there seem to be two entries queued up.

I will again set the retries and have a look if the buffer fills up.

A few items in the buffer is fine as new jobs are first added to the queue and then executed from there.

admin wrote: ↑Mon Jan 24, 2022 11:10 am A few items in the buffer is fine as new jobs are first added to the queue and then executed from there.

You were right, the station upload stopped working again, and the buffer is filled up to 100%. Do you think, the internet connection is the source of the problem here? It would just be strange, that I am still able to access the station :-/. Will do a speed test tomorrow. Should the station be able to recover from that?

The strange thing also is, that the upload interval is set to 5min for Windguru and 1hour for the other two networks and without retries. So for 1 hour you should have maximum of 3 + 12 + 3 = 18 Buffer entries caused by the weather upload service. As in the Screenshot below the station is filled with 155 entries after 42minutes already. However, weather data doesn’t upload already somewhere in between where the buffer fills up, not only if it has already been filled up.
Any way to see what’s in these entries, or what causes them? Any way to check the logs of the weather upload service?

I meanwhile downgraded to 5.1, to see if the problem somehow occurs there as well.

On tab "Live Data - Upload Status" you can check the history of events for your services. Clicking on or hoovering over the marks presents a log list.
That way you should be able to identify which service is filling up the buffer. Can also be a weather network fed with invalid credentials or similar.

I meanwhile upgraded to the latest 5.3 release, disabled any upload except for Windguru and I did also check the "Livedata Upload - Upload Status" and it seems, the Buffer is filled up by "Anzeige verwendet".
However, there seems to be an issue with the Upload itself, so one can see, that the "Windguru Upload" already stopped working, before even the buffer has been filled up to 100%, which would then cause a similar error pattern.

We finally figured out what was causing our problems. First of all, it has nothing to do with Meteobridge. It was a network problem.
Elsewhere, network components have been replaced, resulting in certain requests being filtered out.
I guess that's what happens when one hand doesn't know what the other is doing

gunnar_s wrote: ↑Mon Feb 28, 2022 2:24 pm We finally figured out what was causing our problems. First of all, it has nothing to do with Meteobridge. It was a network problem.
Elsewhere, network components have been replaced, resulting in certain requests being filtered out.
I guess that's what happens when one hand doesn't know what the other is doing

Just for anyone that may come across this post having similar issues as I had. In my case the hardware was faulty and has been replaced. Now the station is working fine again

jhausladen wrote: ↑Sun May 29, 2022 11:58 am

gunnar_s wrote: ↑Mon Feb 28, 2022 2:24 pm We finally figured out what was causing our problems. First of all, it has nothing to do with Meteobridge. It was a network problem.
Elsewhere, network components have been replaced, resulting in certain requests being filtered out.
I guess that's what happens when one hand doesn't know what the other is doing

Just for anyone that may come across this post having similar issues as I had. In my case the hardware was faulty and has been replaced. Now the station is working fine again

I think gunnars problems where related differently.

However, when filling buffer up in short term without having excessive upload events defined and with working Internet connection, then there might be a HW issue. I had 2 or 3 cases on that with >1000 units in operation, so it is very rare. When you give me remote access I can check as the HW defect comes with easy to identify stalled IO processes that cannot be killed and by that stall the whole upload processing.