aws-lambda-rdbms-integration | Integrating AWS Lambda with EC2 hosted Relational Databases | Infrastructure Automation library

Oracle supports the use of Java methods for UDFs. The Java class below uses the AWS SDK to invoke a named Lambda function (fn_name) in either synchronous (RequestResponse) or asynchronous (Event) mode, passing parameters in the form of a JSON string (fn_argsJson):.
Install git and maven on the database server instance, if they're not already installed.
As the oracle user, download and build the aws-lambda-rdbms-integration project from github. Steps: sudo su - oracle git clone https://github.com/awslabs/aws-lambda-rdbms-integration.git cd aws-lambda-rdbms-integration/oracle mvn clean package This builds a self contained .jar file containing the LambdaInvoke java class and all its dependencies, including the AWS SDK class files.
Verify that the EC2 instance role is correct and that you can connect to the AWS Lambda service and successfully call our function, using the main method included in the class: java -cp target/aws-rdbmsinteg-1.0.jar com.amazonaws.rdbmsinteg.LambdaInvoke If all is well, the following output will be displayed: {"Status": "OK", "name": "bob"}
Load the .jar file into the Oracle database: loadjava -user system/<password> target/aws-rdbmsinteg-1.0.jar
Create the awslambda_fn() and awslambda_fn_async() functions using the script provided: cat sql/awslambda_ddl.sql | sqlplus system/<password>
Grant required permissions to the SYSTEM user using the script provided: cat sql/permissions.sql | sqlplus system/<password>
Oracle's Java keystore must trust the certificate authority (CA) used by the AWS service; by default it is empty. An easy way to fix this problem is to replace the default Oracle Java keystore with a populated keystore from a standard Java installation: cp $ORACLE_HOME/javavm/lib/security/{cacerts,cacerts.orig} cat /usr/lib/jvm/jre-1.7.0-openjdk.x86_64/lib/security/cacerts > $ORACLE_HOME/javavm/lib/security/cacerts
Log into the database, and test the awslambda_fn function, passing the name of your Lambda function and a JSON input parameter string: sqlplus system/<password> SQL> SELECT awslambda_fn('lambdaTest','{"name":"bob"}') AS lambdatest FROM DUAL; LAMBDATEST -------------------------------------------------------------------------------- {"Status": "OK", "name": "bob"}
For PostgreSQL, use the PL/Python language to create your UDFs, leveraging the AWS Python SDK to launch Lambda functions and retrieve the results.
Make sure your database EC2 instance has Python and the two AWS SDK modules, boto and boto3, installed: sudo pip install boto boto3 python -c "import boto; import boto3; print 'AWS SDK verified OK'"
Download the aws-lambda-rdbms-integration project as the postgres user: sudo su - postgres git clone https://github.com/awslabs/aws-lambda-rdbms-integration.git cd aws-lambda-rdbms-integration/postgresql
Create the awslambda_fn() and awslambda_fn_async() functions using the script provided: psql -U postgres -f sql/awslambda_ddl.sql
Log into the database, and call the awslambda_fn function, passing the name of the test Lambda function and a JSON input parameter string: psql -U postgres postgres=# SELECT awslambda_fn('lambdaTest','{"name":"bob"}') AS lambdatest ; lambdatest ---------------------------------------------------------- {"Status": "OK", "name": "bob"} (1 row)