"Censored": "the output has been hidden due to the fact that 'no_log: true' was specified for this result", "changed": false

[eybae]

Hi, I’m using AutoML which upgraded latest on tao toolkit v4.0.
The problem is, when I setup the TAO Toolkit API service on a bare-metal server, I got this issue

at the progress with deployment.

Here is my PC’s spec
ubuntu 20.04
nvidia-driver 510.108.03
tao-toolkit 4.0
python 3.9.12
ansible -v 2.14.1
jinja version = 3.1.2

[Morganh]

@eybae

“Censored”: “the output has been hidden due to the fact that ‘no_log: true’ was specified for this result”, “changed”: false

Seems that it is common failure info from the websites searching result.
May I know more detailed when you set up the server? Please refer to Setup — TAO Toolkit 4.0 documentation

[eybae]

I followed Setup — TAO Toolkit 4.0 documentation that guide,
and here is the config of this part

vi tao-toolkit-api-ansible-values.yml

when I check ,

vi tao-toolkit-api-helm-values.yml

The file exists, but empty which means there is no any values.

[Morganh]

Could you check if the blog https://developer.nvidia.com/blog/training-like-an-ai-pro-using-tao-automl/ can help you?
Please note that the “one-click deploy tar” file is actually as below. It is already available after you run "ngc registry resource download-version “nvidia/tao/tao-getting-started:4.0.0"” .
$ ls tao-toolkit-api-bare-metal
check-inventory.yml cnc hosts install-tao-toolkit-api.yml requirements.txt setup.sh tao-toolkit-api-ansible-values.yml tao-toolkit-api-helm-values.yml uninstall-nvidia-drivers.yml

[Morganh]

There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks

Hi,
Below error comes from “check sufficient memory” task in “check-inventory.yml” .

Could you set “no_log” to false in order to get the full log?

More, this task is to check if “ansible_memtotal_mb >= 32000” (See the setting in check-inventory.yml) .
So, the error log may due to

1. The ansible_memtotal_mb is not accessible.
2. Or there is no 32000 Mb memory.

So the solution can be:

1. If the ansible_memtotal_mb is not accessible, then, please find ansible.cfg (should be in ~/ or in /etc/ansible/ansible.cfg). Check if the param gather_subset contains !hardware. If there is, then delete it. Refer to Ansible memtotal_mb fact is undefined - Stack Overflow
2. If the there is no 32000 Mb mem, then decrease the number of mem in the check-inventory.yml