Random reboots on headless Jetson xavier AGX

I went through /var/log/syslog and I was able to find this:

Feb 19 20:31:48 ubuntu systemd[1]: Stopping Session 1 of user sagi.
Feb 19 20:31:48 ubuntu gnome-shell[5659]: gnome-shell: Fatal IO error 11 (Resource temporarily unavailable) on X server :0.
Feb 19 20:31:48 ubuntu systemd[1]: Stopping Session c1 of user gdm.
Feb 19 20:31:48 ubuntu systemd[1]: Removed slice system-getty.slice.
Feb 19 20:31:48 ubuntu systemd[1]: Removed slice system-modprobe.slice.
Feb 19 20:31:48 ubuntu systemd[1]: Stopped target Graphical Interface.
Feb 19 20:31:48 ubuntu systemd[1]: Stopped target Multi-User System.
Feb 19 20:31:48 ubuntu systemd[1]: Stopped target RPC Port Mapper.
Feb 19 20:31:48 ubuntu systemd[1]: Stopped target Sound Card.
Feb 19 20:31:48 ubuntu systemd[1]: Stopped target Timers.

I’m not 100% sure that’s the reboot reason, but it sure looks suspicious enough.
I found another reboot around 5:25 am today but this log isn’t there (in fact, there is no suspicious log there at all)

I’ve set journal to be persistent so I hope to have more info by the next time. But any thoughts about this one in the meanwhile?

please use the serial console log to check.

Thank you but I’m not sure how that helps. can you please elaborate on what I should check?

Put your board into the situation that you can reproduce your issue. In the meantime, turn on the serial console and monitoring the log.

UART serial log can know if there is kernel panic happened and it will print it on serial console.

Syslog cannot capture it if the system is dead/panic at that time. That is why you said you don’t see any suspicious log.

Got it. I dont have any way to reproduce the issue, it just happens, sometimes a few times a day sometimes once every few days.
Maybe ill just connect a PC to log all the data to a file or something?

Is this Xavier AGX devkit or some custom board?

AFAIK original DevKit

Ok, just put it there with serial console enabled and see what would happen then.

1 Like

Finally got it all connected and running, during this time there were about 2-3 unexpected reboots (first one is manual by me)
The debug information doesn’t show anything prior to the reboot. Looks like only the initialization data after the startup is shown

minicom_28.2.23_8.50.log (290.2 KB)

Hi,

Could you directly flash your board with sdkmanager and without using nvme drive and see if this behavior keeps happening?

Didn’t get this part. The os is running on the nvme, is there another way to run it?

Every jetson AGX Xavier has an internal 32GB emmc on it. That is the default method we use.

Please remove every peripherals except the usb cable and power cable on it, reflash your board and see if issue still happened under this situation.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.