Can't deploy container image to lambda function

Question

I try to deploy container image to lambda function, but this error message appear

The image manifest or layer media type for the source image <image_source> is not supported.

here is my Dockerfile, i believe i have use the proper setup

FROM public.ecr.aws/lambda/python:3.8

# Install dependencies
COPY requirements.txt ./
RUN pip install -r requirements.txt

# Copy function code
COPY app/* ./

# Set the CMD to your handler
CMD [ "lambda_function.lambda_handler" ]

When creating lambda function, after specifying the container image, A red box with the message appear @Marcin — Bramanta
– Bramanta, Commented Jan 7, 2021 at 8:59
@Bramanta take a look at this official blog post which has an exact example of the python docker image packing, it worked for me. — samtoddler
– samtoddler, Commented Jan 7, 2021 at 9:01

SharpThunder · Accepted Answer · 2023-01-17 16:22:40Z

58

If you are using buildx >= 0.10 specifying target platform does not work since it also creates multi-platform index by default.

To fix this problem set --provenance=false to docker build.

For more details please see: https://github.com/docker/buildx/issues/1509#issuecomment-1378538197

edited Jan 17, 2023 at 16:22

answered Jan 17, 2023 at 16:22

SharpThunder

6715 silver badges3 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Clivest Over a year ago

Thanks! This fixed our issue. We use buildkit to build docker images for AWS Lambda hosted on AWS ECR. Our deployments started failing recently with error: "Resource handler returned message: "The image manifest or layer media type for the source image *.dkr.ecr.*.amazonaws.com/*:* is not supported. (Service: Lambda, Status Code: 400...)"

safex Sep 24 at 11:54

solved it for me

HemChe · Accepted Answer · 2024-11-07 04:05:14Z

14

I encountered the same problem, building a linux/amd64 image following the official AWS tutorial.

I finally succeeded to create a (working) lambda function using the one of the 3 images published in the ECR repro : not the one with the tag mentionned in the "docker push" command, not the one with size 0 but the 3rd one using its hash tag.

I don't know why the "docker push" command generate 3 images in the ECR repo ! and why the image tagged in the "docker push" command is not working...

edited Nov 7, 2024 at 4:05

HemChe

2,3271 gold badge22 silver badges34 bronze badges

answered Nov 1, 2024 at 16:37

Jean-Luc Guigné

1411 silver badge3 bronze badges

6 Comments

Community Over a year ago

As it’s currently written, your answer is unclear. Please edit to add additional details that will help others understand how this addresses the question asked. You can find more information on how to write good answers in the help center.

eAndy Over a year ago

its been almost 4 years and this is the correct answer. Why can't AWS update their documentation. This is the 3rd AWS resource I've run into some weird problem where the "hello world" examples don't work. this breaks customer trust and stops adoption. Its almost like no one is actually testing the documentation that they write (or that the people testing are too knowledgeable and aren't actually following the written instructions).

getup8 Over a year ago

This worked for me as well. I went into the console, went to Elastic Container Registry, clicked the repository of interest, and then clicked "Copy URI" next to the non-latest one, with type "Image" instead of "Image Index". Pasting this URI into the command worked as intended.

gphilip Jan 16 at 15:43

Worked for me. It's very strange.

Aeryes Jun 11 at 12:51

This worked for me.

|

y0m1g · Accepted Answer · 2022-01-20 18:58:40Z

11

Try by specifying the target platform of the image you build as amd64:

docker build --platform linux/amd64 . -t my_image.

I get the same error while trying to deploy a lambda based on an image that supports both linux/amd64 and linux/arm64/v8 (Apple Silicon) architectures.

answered Jan 20, 2022 at 18:58

y0m1g

1011 silver badge4 bronze badges

3 Comments

Florent Over a year ago

Yes, ECR supports multi-arch images but lambda don't seem to handle them.

Jaime Bellmyer Over a year ago

I believe lambda does handle them, but you need to make that decision when you create/configure the lambda. AWS will expect the build to target the CPU architecture you told it.

Ifarontimi Oct 4 at 22:08

Thanks for this comment my issue was the platform flag. My old Mac was using --platform linux/arm64 while silicon Mac needed that to be linux/arm64/v8. I didn't know that until this comment. But still don't know why the tagged image doesn't work but the unwanted extra image uploaded works

Pat Myron · Accepted Answer · 2025-02-05 18:58:08Z

6

Ran into the same issue on an Apple Silicon MacBook. Pass both --platform linux/amd64 and --provenance=false.

edited Feb 5 at 18:58

Pat Myron

4,6764 gold badges30 silver badges52 bronze badges

answered Feb 5 at 3:01

faisal

611 silver badge3 bronze badges

Comments

Taylor · Accepted Answer · 2025-01-19 22:28:04Z

3

I fought with this for hours. Finally pieced together a couple of key elements from other threads... On my Silicon process Macbook, I was ONLY able to get this to work by passing BOTH the --platform and --provenance switches to the docker build command.

You'll need to replace "lambda-images" in this command with the name of your repository and be sure to replace "us-east-2" with the region you are using.

Example command:

docker build -f Dockerfile \
     --platform linux/amd64 \
     --provenance=false \
    -t 1234567890.dkr.ecr.us-east-2.amazonaws.com/lambda-images:latest \
    .

answered Jan 19 at 22:28

Taylor

7011 gold badge8 silver badges15 bronze badges

Collectives™ on Stack Overflow

Can't deploy container image to lambda function

5 Answers 5

2 Comments

6 Comments

3 Comments

Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

5 Answers 5

2 Comments

6 Comments

3 Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related