Jetson Xavier AGX Self Reboot(Crash) after 1 or 2 hours while Inference

[akashzade.ai]

We have Jetson Xavier AGX with Jetpack 4.4, While inference or working with PyCharm, AGX starts self rebooting(Crash) sometimes without opening any software. I am unable to find the exact reason behind it. So for the same can you help. You can ask for more info if required.

Temperature: 34-39C
Power Mode: 30W All
RAM Consumption: Approx 8GB
GPU Consumption: 20 - 99% (While Inference)
CPU Consumption: 20 - 65%(8 cores)
Free Space: 5.6 GB

Note: Power never fluctuate, AGX connected with ZED2 camera through USB Hub(with Power adopter)

[dkreutz]

Maybe your problem is related to the one I had with AGX freeze in MAXN mode…?

[AastaLLL]

Hi,

Could you share the logs in dmesg and /var/log/kern.log with us first?
Thanks.

[akashzade.ai]

Latest Info:
dmesg.txt (78.4 KB)
kern.log (1.5 MB)

For live monitoring (tegrastats & dmesg --follow) purpose I had taken ssh of AGX on another system, so below is terminal output stored in txt files

agx crash dmesg.txt (74.8 KB)
agx crash tegrastats.txt (149.5 KB)

[AastaLLL]

Hi,

Thanks for sharing the log.

It seems GPU and some CPU is highly used right before the reboot.
Could you share the steps to reproduce this issue with us?

Thanks.

[akashzade.ai]

I am just start ZED2 camera using ROS and instance segmentation inference

[AastaLLL]

Hi,

Could you share more details about how to reproduce this issue in our environment?

For example, what kind of instance segmentation inference do you use?
Could you share a runnable with us?

More, could you test if this issue also happens with a non-stereo camera and JetPack BSP?

Thanks.