Create 3 Freezes and Reboots Regularly with Firmware G.3.1

[scharbonneau8661]

Hello,

My Create 3 robot freezes and reboots regularly ever since I upgrade to Firmware 3.1.1. The ring light turns red and the robot reboots sometime after. This happens at least a few times a day. When the light ring turns red, I cannot access the robot Web server.

Any help on this would be greatly appreciated.

Thank you and best regards,

Serge

[shamlian]

Could you please attach a log, if you are able to get one? If you aim your web browser at http://[ip address of your robot]/logs-raw , you should be able to get something when this load spike starts, though it may take a few tens of seconds to initiate.

Are there other robots on your network? Are you in a home network or an enterprise one? Is the robot actively in use when this happens, or does it seem random?

[shamlian]

I'm sorry that you've had this trouble, and we would love to understand why the problem is happening; if you are able to get a log after the application restarts, hopefully it would help us to understand the cause of the problem so that we can fix it. Feel free to keep your Python script running if you feel it is the trigger (and if you don't mind, we'd love to see the Python script, too). I do like the idea of running the experiment without the script running to see if the event doesn't happen, as well, but I'd say when debugging it would be best to only change one thing at a time.

[scharbonneau8661]

No worries. We'll figure it out. The Create 3 rocks. I've been dreaming about have a robotic platform like this one for 10 years now, 10 years ago I was hacking my Roomba 550 with a Rootooth device and talking to it via bluetooth.

[shamlian]

Any updates on this? Hoping your robot is behaving itself.

[scharbonneau8661]

My robot is behaving most of the time, but I still have the problem identified above on occasions. I only have this problem when my Python script runs. When the robot is completely idle, I don't get this problem. The script executes a 'ros2 topic echo' command for each topic for 5 seconds (15 seconds for the battery_state topic). It launches the command, waits 5 seconds (or 15), stops the command and collects the results. It does it for each topic one at a time continuously. I notice that I get the following error message running a 'ros2 topic echo' command when the light ring turns red prior to the robot restarting:

2022-10-12 06:55:41.232 [RTPS_TRANSPORT_SHM Error] Failed to create segment 6dfbcded97e123a5: boost::interprocess_exception::library_error -> Function compute_per_allocation_extra_size
2022-10-12 06:55:41.232 [RTPS_MSG_OUT Error] boost::interprocess_exception::library_error -> Function init

I remember reading there is a problem when RMW_IMPLEMENTATION is set to rwm_fastrtsp_cpp which is exactly what I am running. Can this be causing the issue I'm experiencing?

[justinIRBT]

That does look like the RMW layer complaining, what are the memory/CPU printouts in the log around the time this error occurs? They look like:

Sep 27 14:06:41 Alfred user.notice create-platform: [INFO] [1664287601.864329542] [alfred.system_health]: CPU usage: max 72 [%] mean 58 [%] RAM usage: 36/59 [MB]

I am not personally aware of an issue with the usage model you are describing.
Have you tried using cyclone to as the RMW to see if the same error occurs?