Installing JetPack 4.2 on the TX2 is failing to flash the Jetson OS. This is using a PC with a fresh install of Ubuntu 18.04. sdkmanager_0.9.11-3405_amd64.deb was installed and the instructions in the user documentation where followed. Recovery mode was started manually.
There are several error messages in the log. The most concerning seem to be the two ‘lost connection’ messages. For example:
2019-03-27 07:01:59.481 - info: Using GenericInstaller to Install NV_L4T_FLASH_TX2_WITH_OS_IMAGE_COMP
2019-03-27 07:01:59.482 - info: enter working directory: /home/roger/nvidia/nvidia_sdk/JetPack_4.2_Linux_P3310/Linux_for_Tegra
2019-03-27 07:01:59.483 - info: [ Component Install Started ]
2019-03-27 07:01:59.483 - info: current working directory is /home/roger/nvidia/nvidia_sdk/JetPack_4.2_Linux_P3310/Linux_for_Tegra
2019-03-27 07:01:59.483 - info: exec_command [target]:
2019-03-27 07:01:59.483 - info: **********************
2019-03-27 07:01:59.484 - info: #!/bin/bash
2019-03-27 07:01:59.484 - info: set -e
2019-03-27 07:01:59.484 - info: export LC_ALL=en_US.UTF-8 LANG=en_US.UTF-8 LANGUAGE=en_US.UTF-8
2019-03-27 07:01:59.484 - info: sudo reboot forced-recovery
2019-03-27 07:01:59.484 - info: **********************
2019-03-27 07:01:59.484 - info: exec_command: scp -q -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no /tmp/tmp_NV_L4T_FLASH_TX2_WITH_OS_IMAGE_COMP.sh nvidia@192.168.55.1:~; ssh -t -q -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no nvidia@192.168.55.1 “~/tmp_NV_L4T_FLASH_TX2_WITH_OS_IMAGE_COMP.sh && rm -f ~/tmp_NV_L4T_FLASH_TX2_WITH_OS_IMAGE_COMP.sh”
2019-03-27 07:04:09.193 - error: lost connection
2019-03-27 07:06:20.264 - info: exit status 255
2019-03-27 07:06:20.265 - warn: command scp -q -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no /tmp/tmp_NV_L4T_FLASH_TX2_WITH_OS_IMAGE_COMP.sh nvidia@192.168.55.1:~; ssh -t -q -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no nvidia@192.168.55.1 “~/tmp_NV_L4T_FLASH_TX2_WITH_OS_IMAGE_COMP.sh && rm -f ~/tmp_NV_L4T_FLASH_TX2_WITH_OS_IMAGE_COMP.sh” finished with error, but ignored
Only the base system is needed at this time. If it helps, the loading of cuda, opencv, etc. can be skipped.
Note that the office’s security processes interfered with sdkmanager being able access the internet (even using a proxy), so this install was run on my home network (192.168.1.*). I haven’t checked the iptables commands yet to see how the 192.168.55.1 is being used. I’ll look into that now.
Does this look familiar to anyone? Any suggestions on where to start?
A zip of ~/.nvsdkm/ is attached.
Thanks,
Roger
nvsdkm.zip (214 KB)