At the moment, when i'm developing AWS cloudformation and EC2 userdata scripts I make source code edits, destroy any existing stack and launch the new one. This is a slow process. Every time I have to wait for a stack to destroy itself and the new one to deploy. If my userdata script dosnt work I have to load up the log and debug the issue.
What development workflow do you use, is there a better way of doing this?
If you update your CloudFormation (CFN) stack, the UserData should be updated accordingly - and be available on the machine.
You can then retrieve the metadata with the following command:
cfn-get-metadata --stack $STACKNAME --resource $RESOURCE --region $REGION
where $STACKNAME is your stack, $RESOURCE is your Configuration. and $REGION is your region. An example would be:
cfn-get-metadata --stack WebserverStaging-EYMMJRF0BU01 --resource myLaunchConfiguration --region eu-west-1
If you like to try the new configuration, you can execute it with:
sudo cfn-init -v --stack $STACKNAME --resource $LAUNCHCONFIG --region $REGION
If you are on a ec2 instance, you can most often retrieve the command directly from /var/log/cloud-init-output.log
