MSSQL (SQL Server) ODBC AWS Lambda layer

This is a small project for building and deploying the mssql-odbc drivers as AWS Lambda layer. You can leverage this layer to build AWS Lambda python serverless applications to connect with a MSSQL (SQL Server) database.

Prerequisites

You will need to have the following running on your local environment in order to successfully perform the steps:

• A POSIX compatible shell (e.g. sh, zsh, ksh or bash)
• Docker installed
• The zip utility installed and available on the PATH
• The aws CLI installed and available on the PATH
• Also, you have set up the aws CLI so you can make calls to the target AWS environment

Usage

You need to perform two steps in order to install this AWS Lambda layer:

1. Building (not required if you just want to use the precompiled layer.zip in this repository)
2. Deploying

Building

You will need to build this AWS Lambda layer by invoking sh build.sh or when chmod-ed with the executable flag (like 0750) simply by invoking ./build.sh.

This build script will make use of the lambci/lambda Docker container so the environment looks almost identical to the AWS Lambda running environment.

After the build.sh script finishes, you will end up with a zip archive called: layer.zip. This zip archive contains:

• The compiled version of unixODBC (with the right paths e.g. /opt in place)
• The msodbcsql17 driver which has been obtained from the official Microsoft YUM repository
• A odbcinst.ini file ready for use

Deploying

The prerequisite for this step is that you have the artifact layer.zip ready to be shipped to AWS.

You have basically two options to have this artifact in place:

1. Run the Build yourself
2. Download the precompiled layer.zip that has been uploaded to this repository

Also, before running the actual deployment script, you might want to export the environment variables with the AWS environment you want to deploy to. This, for instance, can be done quite easiliy with the awsume script from trek10inc.

When you are ready to deploy this AWS Lambda layer, you will need to invoke sh deploy.sh or otherwise if the executable flag has been set, simply by invoking ./deploy.sh.

The deployment script will perform the following:

• Upload the layer.zip to AWS
• Create a AWS Lambda layer named mssql-odbc and making it compatible for runtime python3.7

Done: you are now able to use the AWS Lambda layer from within your AWS Lambda serverless application / script!