I wanted to have a few days worth of data to report the findings. There was indeed an empty variable in one of the short templates that I modified manually, hence the error in the log and consequently the device freezing.
The variable is now fixed and I haven't noticed any related errors or devices frozen neither in the modified nor the original traccar server.
The unmodified version has less devices, therefore less incoming data, so I am troubleshooting mainly on the customized version. Still, after fixing the variable and applying the suggested change in code, I haven't noted any irregularities on the official version either.
I will keep on monitoring and let you know.
Note: The root case in this case turned out to be something particular to the short templates that I modified. However, since other users also report freezing devices, it might be worth checking why the server stops writing in the positions table after such type of errors.
I suspect all other users also have some exceptions. Maybe not the same as yours.
Yes, this is my assumption as well. Since we cannot control the exceptions (or at least not in all cases), maybe we can make sure the positions table data continues to get recorded even after an exception.
This is what I implemented already.
Ahh great. Thanks for confirming.
Hi friend, reviewing everything discussed in this forum. In the last few days it has happened to me with the preliminary version, that the devices freeze and I have left the notification templates short of the official code and they have returned to normal. Which indicates to me that the freezing problem continues to be caused by the Firebase notifications...
I wanted to have a few days worth of data to report the findings. There was indeed an empty variable in one of the short templates that I modified manually, hence the error in the log and consequently the device freezing.
The variable is now fixed and I haven't noticed any related errors or devices frozen neither in the modified nor the original traccar server.
The unmodified version has less devices, therefore less incoming data, so I am troubleshooting mainly on the customized version. Still, after fixing the variable and applying the suggested change in code, I haven't noted any irregularities on the official version either.
I will keep on monitoring and let you know.
Note: The root case in this case turned out to be something particular to the short templates that I modified. However, since other users also report freezing devices, it might be worth checking why the server stops writing in the positions table after such type of errors.