3.3b1 SBC inexplicable crash during print

oliof

I just had a print stop a good hour in with no discernible reason.

This is a machine with Duet 3 main board, tool distribution board, one toolboard 1LC, and a panel due -- in SBC mode with the Raspberry Pi powered separately.

On the SBC, duetcontrolserver didn't run anymore and could not be restarted due to an error -- I will try to get the log files.

I only saw due to a visual check. on the PanelDue, I had a few scary messages

Lost connection to Linux
Warning: Discarded msg src=10 typ=4510 RID=204 exp 205
Cancelled printing file 0:/gcodes/Cleric_DT_v1.2-S3D.gcode, print time was 0h45m
Error: Failed to switch off remote heater 1: Response timeout CAN addr 10 req type 6013, RID=204

The PanelDue showed temperatures were falling, WebUI was stuck showing print temps, with the machine as "off".
Attempts to pull a report via M122 failed.

The Raspberry Pi was still reachable via ssh
Trying to restart duetcontrolserver failed with an error.

Here are some logs: duet_logs.txt

After rebooting both the printer and the Raspberry Pi, the machine ran again. Here is the first M122 I was able to pull: M122-fail.txt

dc42

I am currently looking into a CAN driver bug in 3.3b1 that might explain the CAN errors. The bug would cause a task that is sending messages over CAN to lock up. Possibly this might cause a problem with SPI comms too, I don't know.