Jetson TX1 power issues after jetpack 3.1 update

I flashed TX1 with jetpack 3.1 from jetpack 3.0. After this TX1 shuts down frequently without my knowledge. I could not do my work properly. I had no issues with it when I was using jetpack 3.0. Should I flash it down to jetpack 3.0?. Thanks in advance.

3.1 (which uses L4T R28.1) is generally more reliable, not less…though R27.1 wasn’t too bad. If you were to monitor “dmesg --follow” via serial console, or by ssh (if your terminal is set up with lots of scrollback buffer), then at shutdown your terminal would still see the last messages. Is it possible you can follow “dmesg --follow” via ssh or serial console and see those last messages?

Note: Serial console is superior…at whatever point networking shuts down ssh log monitoring will go away, but serial console continues until all hardware is actually down.

Do make sure you sell all “ok” from:

sha1sum -c /etc/nv_tegra_release

Yes everything is ok for

sha1sum -c /etc/nv_tegra_release

some red outputs from dmesg
[ 0.194185] Error: Driver ‘tegra-mc’ is already registered, aborting…
[ 0.357542] regulator_get() failed for (dvfs,vdd-cpu), -19
[ 0.492150] tegra-pcie 1003000.pcie-controller: failed to get PHY: -517
[ 0.492177] tegra-pcie 1003000.pcie-controller: failed to get PHYs: -517
[ 0.530496] tegra_gpu_edp gpu_edp: unable to get max GPU freq
[ 2.572627] gk20a 57000000.gpu: platform probe failed
[ 4.660388] tegra_gpu_edp gpu_edp: unable to get max GPU freq
[ 4.726641] vi 54080000.vi: all channel register failed

All of those messages are actually normal. There are many things general hardware setup will attempt to test and will show such messages for. You probably need a serial console connected to see what goes on at the moment of shutdown.

One slight possibility I can think of…is there any chance you can test running on an UPS with brownout support? Meaning that if line voltage drops low the UPS still supplies power…some cheaper UPS support only off/on, and do not support momentary sagging of power line voltage. I am thinking power delivery issues could cause a reboot.