AutoML “401 UNAUTHORIZED” problem

[jerrywang8472]

[Morganh]

Could you disable firewall and retry?

ufw disable

More, please make sure the ngc key is correct.

[jerrywang8472]

I disable firewall and generate new ngc key and follow
“NVIDIA NGC” to setup, but still get the same error.

[Morganh]

Could you run similar command as below to check if it can work?
$ curl -k https://127.0.1.1:32080/api/v1/login/bDNyourkey

[jerrywang8472]

I try both https and http.

[Morganh]

Please try the new key since your key is already updated.

[jerrywang8472]

I follow “NVIDIA NGC” to setup and change the “ngc_api_key” in “quickstart_api_bare_metal/tao-toolkit-api-ansible-values.yml”.
Still get the same error.

[Morganh]

How about
$ curl -k https://127.0.1.1:32080/api/v1/login/yournewkey

[jerrywang8472]

[Morganh]

Could you replace “127.0.1.1” with the ip address of your master node and retry the curl command?

[jerrywang8472]

Get the same error.

[Morganh]

Please help collect some logs.
$ kubectl get pods

then
$ kubectl logs tao-toolkit-api-app-pod-5858d657f9-5zqzr | sed “/health/d”

Where the exact pod name can be found with “kubectl get pods”

[jerrywang8472]

[Morganh]

Please run below test.py to share the result.
$ vim test.py

import requests

key = “yourkey” # FIXME: use your latest NGC API KEY
r = requests.get(‘https://authn.nvidia.com/token’, headers={‘Accept’: ‘application/json’, ‘Authorization’: 'ApiKey ’ + key})
print(r.status_code)

$ export PYTHONPATH=/usr/lib/python3/dist-packages:$PYTHONPATH
$ python test.py

Beside test.py, you can also use curl as below.
$ curl --header ‘Accept: application/json’ --header ‘Authorization: ApiKey yourkey’ https://authn.nvidia.com/token

[jerrywang8472]

[Morganh]

So, the key seems to be correct.
Could you run with below test_1.py ?
$ vim test_1.py

import requests
endpoint = “http://127.0.1.1:32080/api/v1/login/yourkey”
response = requests.get(endpoint)
user_id = response.json()[“user_id”]
print("user_id: ", user_id)

$ python test_1.py

That is the cell content of the notebook.
If not working, try to modify line3 as below and re-run.
response = requests.get(endpoint, verify=False)

[jerrywang8472]

I try both “requests.get(endpoint)” and “requests.get(endpoint, verify=False)”.
Both get <Response [401]>.

[Morganh]

Suggest you to follow the method in 401 unauthorized: why?how to fix it? to narrow down.

[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

More, when you fix the issue in your another topic AutoML installation problem [Waiting for the Cluster to become available] - #8 by Morganh , could you share the result of below commands?
I am asking this in order to make sure the tao-api setup is successful.
$ kubectl get pods
$ kubectl logs tao-toolkit-api-app-pod-5858d657f9-5zqzr (Where the exact pod name can be found with “kubectl get pods”)
$ kubectl describe pods tao-toolkit-api-app-pod-5858d657f9-5zqzr (Where the exact pod name can be found with “kubectl get pods”)