Let’s say, I have an application that is written in NodeJs, that process the latest blocks of the steem blockchain and write data into a database. I have three ways of running it on my server.

Using Screen Command

As you probably know, when you are disconnected from a SSH session, the programs that are launched in the session will be terminated. There is any easy way to fix this. You can start a session by using command screen -S name before you run any long-running program e.g. system updates or a program that needs to run forever.

In case the SSH session is disconnected for some reasons e.g. network connectivity issues, the programs will be still running. And you can use command screen -r name to re-connect to the session. You can screen -ls to list the current sessions, and screen -d name to detach a session. You can exit the current screen session using the short-cut Ctrl+A,D which will set the session status to detached.

The advantages of using this method are:

• you can easily monitor the application (by seeing its output)
• you can easily terminate or restart the application e.g. Ctrl + C
• the application runs continuously, meaning that you can keep up to the latest blocks and provide a less than 3 second sync time.

Some disadvantages:

• If the application exists abnormaly, it is not restarted automatically – thus you would need to add the error-handling in your code.
• Long-running code may tend to have more problems: e.g. memory-leaks, crash at some point.

Putting it in Crontab

Another way is to run your program in crontab. However, the finest granularity is every minute i.e. you can specify your application to run every minute.

Advantages are:

• You probably don’t need to do error-handling. In case it fails, it fails. You can just wait for next minute (restart automatically).
• Not to worry about memory leaks. It is less likely to go wrong due to memory issue.

Disadvantages:

• Lose the ability to sync real-time with the blocks. For example, you will be 60 seconds behind the real-time blocks.
• Not easy to monitor the output (you can redirect the errors to files though)

Using a Process Manager

You can combine the both advantages using a Process Manager. For example, if it is a NodeJs application, you can use pm2 utility.

Some Advantages of using pm2 are:

• automatically restart your application in case it crashes or some conditions are met e.g. memory constraints, age.
• easy to monitor the output of each application e.g. you can pm2 show app
• easy to terminate pm2 delete app or restart pm2 restart app
• the application runs continuously
• error-handling is optional – but considered a good practise not to let pm2 restart your app.

–EOF (The Ultimate Computing & Technology Blog) —

推荐阅读：
Win the “Integral” WordPress Theme from Themely　 7 Habits of Highly Productive Bloggers　 How a Business Owner Builds Relationships With Influential Blogg　 5 Promising Tools to Help Boost Your Online Startup　 How to For-Loop and do Math/Arithmetic Operations in Windows/NT 　 How to Find the Closest Sum of Three in an Array using Two Point　 The Lazy Singleton Design Pattern in Java　 How to Find out the Most Frequent Subtree Sum using Depth First 　 A Microsoft Coding Interview Screening for Position Principal So　 The Brace Expansion Algorithms using Breadth First Search or Dep