How to Run a Process as a Background Task

To my disappointment, much of my experience in wrangling DNA sequences involves short bacterial genomes. My workstation, which features a 4.0-GHz 8-core processor and 32 Gb RAM, had been sufficient for my purposes, but only until I encountered the problem of aligning 147 whole-genome sequences prior to tree building. Even though MAFFT allows multithreading, my workstation would have taken weeks to complete the process. For fear of not graduating in time, using our project-based server suddenly became compulsory.

In my initial attempts of using our server, my processes either ended up being in a sleep (requires some resource that is currently not available) or zombie state (dead process that hasn’t been cleaned up properly). Here’s how to keep a process running as a background task in a Linux server:

nohup command >command.log 2>command.err &

1. nohup: Intercepts hangup (HUP) signal that is normally sent to a process to inform that the user has logged off, allowing the process to continue running
2. command: User-provided command (e.g. mafft in > out)
3. >command.log: File that captures the standard output
4. 2>command.err: File that captures the standard error. To redirect the standard error to the same output file which contains the standard output, use 2>&1.
5. &: Runs the command as a background task

pgrep can be used to monitor the process even after the original terminal that accepted the command was closed. pgrep lists the process IDs that match the selection criteria (pattern) to standard output. For example:

pgrep -u root mafft

will only list the processes that are called mafft and owned by root.

Using the process identifier (PID) listed by pgrep, keep tabs on the parent and child processes of the command using the ps command:

ps PID

A quick snapshot of the absolute paths of the processes running in the server can be obtained with:

top -c