Cumulus connection via TCP/IP

[KoenDV]

Hi all,

I use CumulusMX version 3107 and have it connected via TCP/IP with a Meteobridge Nano SD. Since a few versions Cumulus connects and disconnects 3 times after each other at certain times as you can see in the image below.


When I used version 3074 this didn't happen. Once CumulusMx connected, it stayed connected without disconnecting even one time.

The problem is, when CumulusMx disconnects that much like now, Meteobridge can't handle this after a while and will hang. Is there a reason why it disconnects so much in these more recent versions, and is there a way to prevent it?

Kind regards,
Koen

[mcrossley]

There hasn't been any change to the disconnecting in Cumulus that I can recall. The meteobridge seems to forcibly close the connection if it has other things to do. Have you enabled more upload features on the meteobridge?

Take a look in your MXdiags log file to see if the connection is being closed.

[KoenDV]

Hi Marc,

No I didn't. I tested this with first installing the latest version. The disconnect problem happened, after this, I reinstalled version 3074 which didn't had this reconnect issues. After reinstalling the latest version again, it happened again...

Here you can find the data from the MXdiags log file from the disconnect period of 11:34

2021-02-04 11:33:48.430 SendLoopCommand: Starting - LPS 2 1
2021-02-04 11:33:49.453 SendLoopCommand: Starting - LOOP 50
2021-02-04 11:33:53.489 LOOP: 3 - Expected data not received, expected 99 bytes, got 0
2021-02-04 11:33:57.314 SendLoopCommand: Starting - LPS 2 1
2021-02-04 11:34:00.159 DoLogFile: Writing log entry for 4/02/2021 11:34:00
2021-02-04 11:34:00.168 DoLogFile: Written log entry for 4/02/2021 11:34:00
2021-02-04 11:34:00.168 Writing today.ini, LastUpdateTime = 4/02/2021 11:34:00 raindaystart = 116,2 rain counter = 116,2
2021-02-04 11:34:00.840 InitTCP: Connecting to the station
2021-02-04 11:34:01.212 Realtime[238]: Uploading - realtimegauges.txt
2021-02-04 11:34:01.846 InitTCP: Flushing input stream
2021-02-04 11:34:06.396 InitTCP: Connection confirmed
2021-02-04 11:34:06.396 InitTCP: Connecting to the station
2021-02-04 11:34:07.523 InitTCP: Flushing input stream
2021-02-04 11:34:12.077 InitTCP: Connection confirmed
2021-02-04 11:34:12.077 SendLoopCommand: Error sending LOOP command [LPS 2 1]: Geen toegang tot een verwijderd object.
Objectnaam: System.Net.Sockets.NetworkStream.
2021-02-04 11:34:12.077 InitTCP: Connecting to the station
2021-02-04 11:34:13.085 InitTCP: Flushing input stream
2021-02-04 11:34:16.230 Realtime[239]: Uploading - realtimegauges.txt
2021-02-04 11:34:17.638 InitTCP: Connection confirmed
2021-02-04 11:34:17.638 SendLoopCommand: Starting - LOOP 50
2021-02-04 11:34:31.222 Realtime[240]: Uploading - realtimegauges.txt
2021-02-04 11:34:35.397 LOOP: 10 - Expected data not received, expected 99 bytes, got 1
2021-02-04 11:34:38.401 SendLoopCommand: Starting - LPS 2 1
2021-02-04 11:34:39.421 SendLoopCommand: Starting - LOOP 50
2021-02-04 11:34:46.244 Realtime[241]: Uploading - realtimegauges.txt
2021-02-04 11:35:00.428 DoLogFile: Writing log entry for 4/02/2021 11:35:00

[mcrossley]

Hi, I don't know what more to say, later versions have more error checking and reties added specifically to try and cope with the vagaries of meteobridge.

But to be honest the meteobridge is supposed to emulate Davis IP loggers but they don't do it very well. The Cumulus code works fine with Davis IP loggers, and I'm getting fed up with added more and fixes to try and make them work. I believe WeeWx has similar problems with these devices.

[jlmr731]

I use the meteobridge and see some of that within the logger every few hours, my guess is that it has something to do with what meteobridge does. now Cumulus keeps chugging along with out skipping a beat, if you notice connection closed and on that same second of time its back up. So could be poor tcp stack on the device or within meteobridge firmware.
Honestly I have not had an issue for many years with this setup but to be honest today was the first time i looked at the logs, also it could be something new within the system they changed. Now i just rebooted the meteobridge and see there was a new firmware release so we will see what happens now