Upgrading to AODT 1.2.0

Accelerated Computing Aerial Forum
1

Hi,

I have been trying to upgrade to v1.2.0 today on a machine where I have previously been running v1.1.0.

I have completed the installation, although to complete, I had to run two non-listed steps:

  1. When hit with WARNING: An NVIDIA kernel module ‘nvidia-uvm’ appears to be already loaded in your kernel, I ran: sudo lsof /dev/nvidia* and killed the processes. This enabled install.sh to run further.
  2. I then needed to run docker login nvcr.io -u ‘$oauthtoken’ -p “$NGC_CLI_API_KEY” when the containers would not pull (as shown in image). This enabled install.sh to finish.

image
image912×436 113 KB

Following this, when I re-accessed the GUI, the omniverse launcher only showed v1.1.0 so I hit uninstall on this incase there was an issue due to the old version being present.

I now want to run the GUI, but when running ~/aodt_1.2.0/frontend/start.sh I get:

image

I can run both ~/aodt_1.2.0/backend/start.sh or docker compose up -d from within the backend directory seemingly successfully, but at no point have I been able to see the v1.2.0 UI.

I look forward to hearing back.

Thanks

2

@oed The installation process generates logs. Can you please share those? Do the following,
cd ~/aodt_1.2.0
./install.sh

You will find a file named

  • ./aodt_1.2.0/install_YYYYMMDD_HHmmSS.log
  • ./aodt_1.2.0/frontend/OV_GUI_install.log
    Please share those via DM
3

@kpasad1, fixed and now running AODT 1.2.0 after reinstalling fresh OS.

1 Like