I'm trying to launch a lambda function for local development inside a docker container using AWS SAM. The gotcha here is I'm running on a Macbook using a M1 chipset and I have a base container which must be executed on x86_64 chips (Oracle thick client requirement).
I was able to get the container to build using colima, and I have a default colima VM running as below:
PROFILE STATUS ARCH CPUS MEMORY DISK RUNTIME ADDRESS
default Running x86_64 4 12GiB 60GiB docker
This looks good and I can startup and execute other containers as expected.
The problem I"m facing is when running a container through AWS SAM its failing with:
[ERROR] (rapid) Init failed error=Runtime exited with error: signal: illegal instruction (core dumped) InvokeID=
Which looks to be related to not enough memory.
I'm launching this using a simple configure template.yml file of
AWSTemplateFormatVersion: '2010-09-09'
Transform: AWS::Serverless-2016-10-31
Globals:
Function:
Timeout: 360
Tracing: Active
Resources:
Function:
Type: AWS::Serverless::Function
Properties:
PackageType: Image
Architectures:
- x86_64
Metadata:
Dockerfile: Dockerfile.local
DockerContext: .
I'm launching this through the Pycharm AWS Toolkit plugin. The question is how do I provide more memory to the container which is running the function?
I should note that this works flawlessly when running on a native x86_64 chip, and only uses about 1GB of memory.
I'm sure I've seen at one point in the AWS Documentation a Metadata.DockerArgs or something similar but I can no longer find reference to it in the documentation.
