[Bug]: RAK4631 not responding

[cracky22]

Category

Other

Hardware

Rak4631

Firmware Version

2.5.15.79da236

Description

I've already tried restarting, resetting and even re-flashing, but my RAK4631 with nrf52 chip doesn't hold 80 nodes. I can definitely identify the problem in connection with the nodes because I have created a channel where there are only 8 nodes and the device stays online there.

With over 80 nodes the rak crashes from time to time (between 6 and 9 hours) and the green LED lights up continuously.

What can I do??

Relevant log output

No response

[SimbimChimbetov]

This problem only on 2.5.15?

[cracky22]

This problem only on 2.5.15?

No, unfortunately not. I have already tried older fw versions and always encountered this problem. Do you know a working fw? Actually it works fine if I just change the channel Our place but I do not use the ones from the city

[SimbimChimbetov]

I'm not sure but maybe I have the same problem with v 2.5.11
I recently updated a Node on a Mountain with ober 100 nodes in reach, few days ago I lost Signal, last signal showed 88% Battery and the weather was fine.
Because it is really hard to access and I don't had the time to go there, I can't confirm, maybe it is just stolen.
I hope I have the time this Weekend.

[cracky22]

It would be interesting to know if it is the same for other nrf52 devices. I have a t1000-e but it works there. Could it be due to the different storage options such as ram and eeprom?

[markbirss]

It would be interesting to know if it is the same for other nrf52 devices. I have a t1000-e but it works there. Could it be due to the different storage options such as ram and eeprom?

All nrf52 devices have only 28kb of littlefs storage available where preferences, ble parings and nodedb are stored (the littlefs block size will also waste some space)

As a measure to prevent this issue the nodedb (db.proto) size was reduced from 100 to 80 recently

#5346

you can confirm the size of the nodedb file running this test fw file

(https://discord.com/channels/867578229534359593/919642584480112750/1305904252626927729)

use "list-files-s140_nrf52_611_softdevice-1.0.0.4265ae9.uf2" for rak6431
the other for seeed boards with newer SoftDevice

if you are able to share the output this could help understand the issue further

[cracky22]

It would be interesting to know if it is the same for other nrf52 devices. I have a t1000-e but it works there. Could it be due to the different storage options such as ram and eeprom?

All nrf52 devices have only 28kb of littlefs storage available where preferences, ble parings and nodedb are stored (the littlefs block size will also waste some space)

As a measure to prevent this issue the nodedb (db.proto) size was reduced from 100 to 80 recently

#5346

you can confirm the size of the nodedb file running this test fw file

(https://discord.com/channels/867578229534359593/919642584480112750/1305904252626927729)

use "list-files-s140_nrf52_611_softdevice-1.0.0.4265ae9.uf2" for rak6431
the other for seeed boards with newer SoftDevice

if you are able to share the output this could help understand the issue further

I can't even open the uf2, can you put it here?

[cracky22]

It would be interesting to know if it is the same for other nrf52 devices. I have a t1000-e but it works there. Could it be due to the different storage options such as ram and eeprom?

All nrf52 devices have only 28kb of littlefs storage available where preferences, ble parings and nodedb are stored (the littlefs block size will also waste some space)

As a measure to prevent this issue the nodedb (db.proto) size was reduced from 100 to 80 recently

#5346

you can confirm the size of the nodedb file running this test fw file

(https://discord.com/channels/867578229534359593/919642584480112750/1305904252626927729)

use "list-files-s140_nrf52_611_softdevice-1.0.0.4265ae9.uf2" for rak6431
the other for seeed boards with newer SoftDevice

if you are able to share the output this could help understand the issue further

I can't even open the uf2, can you put it here?

[cracky22]

So that means I connect the RAK to my computer and send you the logs?

[cracky22]

So that means I connect the RAK to my computer and send you the logs?

[thebentern]

Please note that this file will not work for the T1000-E, as it uses a different SoftDevice version
list-files-s140_nrf52_611_softdevice-1.0.0.4265ae9(1).uf2.zip

[cracky22]

So that means I connect the RAK to my computer and send you the logs?

.

[thebentern]

I have had rotten luck reproducing this issue so far, but today I am trying an all day run of a RAK board connected to msh/US against MQTT (client proxy) to see if it triggers at all for me. >140 nodes witnessed so far.

[cracky22]

I have had rotten luck reproducing this issue so far, but today I am trying an all day run of a RAK board connected to msh/US against MQTT (client proxy) to see if it triggers at all for me. >140 nodes witnessed so far.

Hi, how do you get this output? Is there a tool that can log and save all the important information?

[cracky22]

Or is it just serial?

[thebentern]

Or is it just serial?

It is just serial logs. I like to use tio (https://github.com/tio/tio) because it will re-attach to the device in the case of a failure or reboot.

I ran the RAK node for about 7 hours yesterday on the msh/US topic and picked up over 600 nodes with no crashes or failures. To rule out any issues with file corruption problems, have you tried a factory reset (or even just nodedb rese)?

[cracky22]

Or is it just serial?

It is just serial logs. I like to use tio (https://github.com/tio/tio) because it will re-attach to the device in the case of a failure or reboot.

I ran the RAK node for about 7 hours yesterday on the msh/US topic and picked up over 600 nodes with no crashes or failures. To rule out any issues with file corruption problems, have you tried a factory reset (or even just nodedb rese)?

How did you get 600 nodes? We're not talking about MQTT, are we?

[garthvh]

Or is it just serial?

It is just serial logs. I like to use tio (https://github.com/tio/tio) because it will re-attach to the device in the case of a failure or reboot.
I ran the RAK node for about 7 hours yesterday on the msh/US topic and picked up over 600 nodes with no crashes or failures. To rule out any issues with file corruption problems, have you tried a factory reset (or even just nodedb rese)?

How did you get 600 nodes? We're not talking about MQTT, are we?

Yes, that is how you get to 600 nodes quickly, the topics are for mqtt

[cracky22]

@thebentern @garthvh @markbirss

[cracky22]

meshtastic-log-2024-12-07T11-01-23.989Z.log

[cracky22]

[cracky22]

I know that this is unnecessary, but it would be possible if there was an option in the firmware to query how much "memory" is available on the board and that could also be displayed in the Android app in a small graphic/text

[markbirss]

I know that this is unnecessary, but it would be possible if there was an option in the firmware to query how much "memory" is available on the board and that could also be displayed in the Android app in a small graphic/text

the littlefs support specific for nrf52 dont currently have function to get free space
you could look at adding a android app feature request on the app for free memory
https://github.com/meshtastic/Meshtastic-Android/issues

[cracky22]

Ok, I can do that. But what about my problem with the RAK? What can I do or how can I debug further

[markbirss]

Ok, I can do that. But what about my problem with the RAK? What can I do or how can I debug further

Ok, are you able to capture log as the reboot/crash occur at all?
Listing of files still show db.proto size after crashed ? (or it this the already provided listing ?)

[cracky22]

This is already the list.
No, I can't log it as it crashes as it takes between 6 and 9 hours