Issues with AWS Lambda Docker Container (permission denied) - python-3.x

Question

I am deploying a Docker container with AWS Lambda which executes a machine learning model. But whenever I attempt to load the saved checkpoints, I am getting a permission denied message when attempting to read the local filesystem or a no directory found based on where I am storing the model...?
Here is the relevant Dockerfile:
FROM public.ecr.aws/lambda/python:3.8
# copy requirements.txt file to the container
COPY requirements.txt ./
# upgrade pip and install the python requirements from requirements.txt
RUN python3.8 -m pip install \
--upgrade pip
RUN python3.8 -m pip install \
-r requirements.txt
# Copy function code
COPY app.py ./
# Install the runtime interface client
RUN python3.8 -m pip install \
awslambdaric
# clean up image for small container
RUN find . -type d -name "tests" -exec rm -rf {} +
RUN find . -type d -name "__pycache__" -exec rm -rf {} +
RUN find . -type d -name "include" -exec rm -rf {} +
RUN rm -rf ./{caffe2,wheel,wheel-*,pkg_resources,boto*,aws*,pip,pip-*,pipenv,setuptools}
RUN rm -rf ./{*.egg-info,*.dist-info}
RUN find . -name \*.pyc -delete
RUN find . -type d -name "test" -exec rm -rf {} +
RUN ls -R -al
# update linux libraries
RUN yum update -y
# install python3 and unzip
RUN yum install -y python3 unzip
# pull model files
RUN mkdir ./model
RUN curl https://somewhere.com/model.zip -o ./model/model.zip
RUN unzip ./model/model.zip -d ./model
RUN chmod 644 ./model
RUN chmod -R 644 ./model/*
RUN rm ./model/model.zip
WORKDIR ./
ENTRYPOINT [ "python3", "-m", "awslambdaric" ]
CMD [ "app.lambda_handler" ]
And the app.py:
from __future__ import print_function
import json, time
import urllib.request
from jose import jwk, jwt
from jose.utils import base64url_decode
from sentence_transformers import SentenceTransformer, util
model = ModelFunction('/model')
def lambda_handler(event, context):
body = json.loads(event['body'])
token = body['jwttoken']
utterance = body['utterance']
comparestring = body['comparestring']
# generate embeddings for each phrase
embeddings1 = model.encode(utterance, convert_to_tensor=True)
embeddings2 = model.encode(comparestring, convert_to_tensor=True)
# compute score
score = util.embeddings(embeddings1, embeddings2)
# output score
print("Score:")
print(score.item())
return {
"statusCode": 200,
"score": json.dums(score.item())
}
# the following is useful to make this script executable in both
# AWS Lambda and any other local environments
if __name__ == '__main__':
event = {
'token': '',
'email': 'somewhere#somewhere.com'}
lambda_handler(event, None)
A couple of the relevant error messages:
[ERROR] PermissionError: [Errno 13] Permission denied: './model/modules.json'
Traceback (most recent call last):
File "/var/lang/lib/python3.8/imp.py", line 234, in load_module
return load_source(name, filename, file)
File "/var/lang/lib/python3.8/imp.py", line 171, in load_source
module = _load(spec)
File "<frozen importlib._bootstrap>", line 702, in _load
File "<frozen importlib._bootstrap>", line 671, in _load_unlocked
File "<frozen importlib._bootstrap_external>", line 848, in exec_module
File "<frozen importlib._bootstrap>", line 219, in _call_with_frames_removed
File "/var/task/app.py", line 8, in <module>
model = model('./model')
File "/var/lang/lib/python3.8/site-packages/model/model.py", line 115, in __init__
with open(os.path.join(model_path, 'modules.json')) as fIn:
and
[ERROR] FileNotFoundError: [Errno 2] No such file or directory: '/home/sbx_user1051/.cache/model'
Traceback (most recent call last):
File "/var/lang/lib/python3.8/imp.py", line 234, in load_module
return load_source(name, filename, file)
File "/var/lang/lib/python3.8/imp.py", line 171, in load_source
module = _load(spec)
File "<frozen importlib._bootstrap>", line 702, in _load
File "<frozen importlib._bootstrap>", line 671, in _load_unlocked
File "<frozen importlib._bootstrap_external>", line 848, in exec_module
File "<frozen importlib._bootstrap>", line 219, in _call_with_frames_removed
File "/var/task/app.py", line 8, in <module>
model = model('/model')
File "/var/lang/lib/python3.8/site-packages/model/model.py", line 101, in __init__
shutil.rmtree(model_path)
File "/var/lang/lib/python3.8/shutil.py", line 709, in rmtree
onerror(os.lstat, path, sys.exc_info())
File "/var/lang/lib/python3.8/shutil.py", line 707, in rmtree
orig_st = os.lstat(path)
Any ideas? I think I am not storing the model checkpoints in the correct directory? The above logs have been somewhat scrubbed due to NDA etc.

Answer 1

Try running:
chmod 644 $(find . -type f)
chmod 755 $(find . -type d)
on the files you're trying to deploy.
As the AWS Docs say, there may be a permissions issue if the files aren't set to be executable by any user.