CodeDeploy Timeout Error and Corrupted EC2 - shell

Question

I have an issue where my most recent CodeDeploy failed due to a timeout and am now not able to connect or load my EC2 instance. It didn't rollback to the previous working version and since I can't connect I can't take a deeper look at the CodeDeploy logs. When looking at the log tail there isn't anything glaring about the error log and I don't think it is related to the timeout error, but I'm not sure what in my shell script could be creating the timeout. Any ideas of what might be wrong in my setup? Should I do a check for a node_modules folder before running sudo npm install?
Error:
Error Code: ScriptTimedOut
Script Name:scripts/npm-install.sh
Message: Script at specified location: scripts/npm-install.sh failed to complete in 300 seconds
Log Tail:
LifecycleEvent - AfterInstall
Script - scripts/npm-install.sh
[stderr]npm WARN deprecated lodash-node#3.10.2: This package is discontinued. Use lodash#^4.0.0.
[stderr]npm WARN deprecated sendgrid#4.10.0: Please see v6.X+ at https://www.npmjs.com/org/sendgrid
[stderr]npm WARN deprecated nodemailer#2.7.2: All versions below 4.0.1 of Nodemailer are deprecated. See https://nodemailer.com/status/
[stderr]npm WARN deprecated mailcomposer#4.0.1: This project is unmaintained
[stderr]npm WARN deprecated socks#1.1.9: If using 2.x branch, please upgrade to at least 2.1.6 to avoid a serious bug with socket data flow and an import issue introduced in 2.1.0
[stderr]npm WARN deprecated buildmail#4.0.1: This project is unmaintained
[stderr]npm WARN deprecated sendgrid#1.9.2: Please see v6.X+ at https://www.npmjs.com/org/sendgrid
[stderr]npm WARN deprecated mailparser#0.6.2: This project is unmaintained
[stderr]npm WARN deprecated mimelib#0.3.1: This project is unmaintained
[stderr]
Here is my appspec.yml:
version: 0.0
os: linux
files:
- source: /
destination: /var/www/app/
hooks:
AfterInstall:
- location: scripts/npm-install.sh
runas: ec2-user
timeout: 300
ApplicationStart:
- location: scripts/npm-start.sh
runas: ec2-user
timeout: 60
npm-install.sh:
#!/bin/bash
source /home/ec2-user/.bash_profile
cd /var/www/app
sudo npm install
npm-start.sh:
#!/bin/bash
source /home/ec2-user/.bash_profile
cd /var/www/app
npm start

Answer 1

As you probably know, the 300 second timeout is configured in your appspec.yml file. I'm assuming that npm install should finish in less than 300 seconds.
With the information provided, it's impossible to know what went wrong. Per your concern about the node_modules directory not existing, it won't matter because npm install will create that directory if it doesn't already exist. If possible, I would strongly suggest upgrading the deprecated packages. It's possible NPM is not happy about that.
As for your instance, there is nothing in the scripts that you've provided that should kill your instance. If you can't connect to your EC2 instance, you'll probably want to recycle it and get a new one. If you have something actually running on the EC2 instance you can't connect to, you could create a new CodeDeploy deployment group with the new instance, deploy the code and see if the instance remains healthy. This could be an intermittent oddity, and it would be more valuable to troubleshoot if the issue becomes repeatable.