Riva Deployment Error following tutorial on AWS with EKS

AI & Data Science Deep Learning (Training & Inference) Riva
,
1

Hi I just followed the tutorial for Deploy Riva at Scale on AWS with EKS.

And when it reaches installing riva-api on helm by “helm install riva-api riva-api …” and then the triton0 pod will just failed to init, showing Error: Invalid apikey

The log is as follow:

Screenshot 2024-09-19 at 11.49.09 PM
Screenshot 2024-09-19 at 11.49.09 PM1263×438 31.7 KB

I’ve set all the keys and credential, not sure if it’s the same issue as this

Does someone know why this happens and how to fix it, or any alternatives? Thanks in advance. (I really wan to test and use Riva!!)

2

I guess the error would be on riva-api/templates/trition.yaml:
line 88: download_and_deploy_ngc_models $model_full_path 2>&1 | tee /dev/shm/{{ $group_name }}.log
for container https://helm.ngc.nvidia.com/nvidia/riva/charts/riva-api-2.16.0.tgz

My API key should be correct, so the error is in command download_and_deploy_ngc_models, which I hope technician from Nvidia might need to take a look at this asap.