How do I design a Bootup script for setting up an apache web server running on a brand new AWS EC2 Ubuntu instance? - apache

Question

I want to configure an EC2 instance so that it installs, configures and starts an Apache web server without my (human) intervention.
To this end, I am taking advantage of the "User Data" section and I have written the following script:
#!/bin/bash
sudo apt-get upgrade
sudo apt-get install -y apache2
sudo apt-get install -y awscli
while [ ! -e /tmp/index.html ]; do aws s3 cp s3://vietnhi-s-bucket/index.html /var/www/html; done
sudo systemctl restart apache2
Description of the Functionality of the Bootup script:
The script forces an update of the Ubuntu instance from whatever the date of the AMI image was when the image was created to today, when the EC2 instance is created from the image.
The script installs the Apache 2 server.
The script installs the AWS CLI interface. Because the aws s3 cp command on the next line is not going to work without the AWS CLI interface.
The script copies the sample index.html file from the vietnhi-s-bucket S3 bucket to the /var/www/html directory of the Apache web server and overwrites its default index.html file.
The script restarts the Apache web server. I could have used "Start" but I chose to use "restart".
Explanatory Notes:
The script assumes that I have created an IAM role that permits AWS to copy the file index.html from an S3 bucket called "vietnhi-s-bucket". I have given the name "S3" to the IAM role and assigned the "S3ReadAccess" policy to that role.
The script assumes that I have created an S3 bucket called "vietnhi-s-bucket" where I have stashed a sample index.html file.
For reference, here are the contents of the sample index.html file:
[html]
[body]
This is a test
[/body]
[/html]
Does the bootup script work as intended?

Answer 1

The script works as-is.
To arrive at that script, I had to overcome three challenges:
Create an appropriate IAM role. The minimum viable role MUST include the "S3ReadAccess" policy. This role is absolutely necessary for AWS to be able to use the public and private access keys for your AWS account and that are loaded in your environment. Copying the index.html file from the Vietnhi-s-bucket S3 bucket is not feasible if AWS does not have access to your AWS account keys.
Install the AWS CLI interface (awscli). For whatever reason, I never saw that line included in any of the AWS official documentation or any of the support offered on the web including the AWS forums. You can't run the AWS CLI s3 cp command if you don't install the AWS CLI interface.
I originally used "aws s3 cp s3://vietnhi-s-bucket/index.html /var/www/html" as my original copy-from-S3 instruction. Bad call. https://forums.aws.amazon.com/thread.jspa?threadID=220187
The link above refers to a timing issue that AWS hasn't resolved and the only workaround is to wrap retries around the aws s3 cp command.