Cron for restarting a service when it fails on Amazon EC2 Ubuntu instance

Question

I have made a service which contains some redis operations and node scripts on a ubuntu Amazon Ec2 instance. However when there are any errors the service stops with a pid.

I need a cron job to do the following :

1. detect when service has failed
2. delete the pid
3. restart the service

I have seen that cron can be created for time interval but have not come across if this can be done for service fail scenarios as well . Any hints and answers on this will be very helpful .

Answer 1

You can use Forever . From their page:

A simple CLI tool for ensuring that a given script runs continuously (ie forever).

The usage is quite simple:

forever start your_app.js

You can read the documentation in their page: Forever - Github . You have several examples here: Forever examples .

There are other alternatives:

1. Monit: http://mmonit.com/monit/
2. Upstart: http://upstart.ubuntu.com/
3. Daemontools: http://cr.yp.to/daemontools.html
4. Launchtool: http://people.debian.org/~enrico/launchtool.html

Answer 2

Other alternatives I like (which they're not written in Node.js) are:

• Supervisor (Python)
• lockrun (C - really lightweight)