create-or-update-stack.sh

#!/usr/bin/env bash

usage="Usage: $(basename "$0") region stack-name [aws-cli-opts]

where:
  region       - the AWS region
  stack-name   - the stack name
  aws-cli-opts - extra options passed directly to create-stack/update-stack
"

if [ "$1" == "-h" ] || [ "$1" == "--help" ] || [ "$1" == "help" ] || [ "$1" == "usage" ] ; then
  echo "$usage"
  exit -1
fi

if [ -z "$1" ] || [ -z "$2" ] || [ -z "$3" ] ; then
  echo "$usage"
  exit -1
fi

shopt -s failglob
set -eu -o pipefail

echo "Checking if stack exists ..."

if ! aws cloudformation describe-stacks --region $1 --stack-name $2 ; then

  echo -e "\nStack does not exist, creating ..."
  aws cloudformation create-stack \
    --region $1 \
    --stack-name $2 \
    ${@:3}

  echo "Waiting for stack to be created ..."
  aws cloudformation wait stack-create-complete \
    --region $1 \
    --stack-name $2 \

else

  echo -e "\nStack exists, attempting update ..."

  set +e
  update_output=$( aws cloudformation update-stack \
    --region $1 \
    --stack-name $2 \
    ${@:3}  2>&1)
  status=$?
  set -e

  echo "$update_output"

  if [ $status -ne 0 ] ; then

    # Don't fail for no-op update
    if [[ $update_output == *"ValidationError"* && $update_output == *"No updates"* ]] ; then
      echo -e "\nFinished create/update - no updates to be performed"
      exit 0
    else
      exit $status
    fi

  fi

  echo "Waiting for stack update to complete ..."
  aws cloudformation wait stack-update-complete \
    --region $1 \
    --stack-name $2 \

fi

echo "Finished create/update successfully!"

also deploy does not support rollback configuration

Also useful as a workaround to serverless/serverless#2233 and serverless/serverless#2831.

The logic between lines 53 and 63 can be removed by simply adding --no-fail-on-empty-changeset to the aws cloudformation update-stack ... command.

Changed my mind - you actually do need separate logic for handling the case of "deploy to an existing stack, but there are no updates to apply". Consider a stack in CREATE_COMPLETE status, and we go to perform another deploy as an update. If we have no changes to apply, the stack will remain in CREATE_COMPLETE status, and we must not attempt to use aws cloudformation wait stack-update-complete because this command would hang forever. So I think the existing snippet is correct, and you should not use --no-fail-on-empty-changeset

Am getting the below error while trying to update the existing stack. What do I need to change in the script?
An error occurred (AlreadyExistsException) when calling the CreateStack operation: Stack [Test-stack] already exists

@viveksamaga I've not tested this since 2016 sorry, it's possible the AWS CLI has introduced some breaking changes