Core dump when using S3 Model repository

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

Please provide the following information when requesting support.

Hardware - GPU: T4
Hardware - CPU: Unknown
Operating System: Amazon Linux2 (Docker container)
Riva Version: 1.8.0

I recently changed my model repository to use S3 instead of a local repo.
I run the riva-build and riva-deploy containers and upload all the files and folders to S3, however, when I go to run the Triton server, I get the following error:

terminate called after throwing an instance of 'std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >'

I am running Triton using

tritonserver --log-verbose=1 --log-info=true --log-warning=true --log-error=true --strict-model-config=false --model-control-mode=explicit "$LOAD_MODEL_STR" --model-repository "$MODEL_REPOSITORY"

And Riva container using

docker run -it --rm --gpus all --ulimit memlock=-1 --ulimit stack=67108864 --shm-size=64g \
  -p 8000-8002:8000-8002 -p 50051:50051 \
  riva-api:$RIVA_VERSION

Here’s a pastebin of the full logs: ============================= Riva Speech Skills =========================== - Pastebin.com

Any insight into why this might be happening?

#3

Hi, did you find a solution?

Mini Militia App Lock

#4

I did not and reverted back to using our old method which is just zipping the files, uploading them, then unzipping at deploy time :/

#5

Hi @pineapple9011

Thanks for your interest in Riva,

Apologies for the delay,

I will check regarding this issue/error further with the team and will provide an update soon,

Also my apologies, the Pastebin log attached is not accessible, will it be possible to attach as zip file in this thread

#6

riva-logs.txt (32.4 KB)

Please see the attached file.

#7

Hi @pineapple9011

Thanks for your interest in Riva,

I have checked with the team,

Apologies, Currently as of today, we do not support S3