Monitor disk usage on your server


You might consider monitoring your disk usage, because this might reveal trends that help you to plan for the future. Here, we’ll use a cron job to periodically scan a directory with the duc program by Ico Doornekamp.

Repeating jobs each month#

Let’s create two jobs that will run periodically:

  1. Create an index of a folder.

  2. Scan the index and email a text report.

We can create new cron jobs by running:

crontab -e

A text editor should open. Paste the following text into the editor. Make sure you change the paths to match your filesystem. Notice that we write the full path to the duc executable file and the full path to the script.

# +---------------- minute (0 - 59)
# |  +------------- hour (0 - 23)
# |  |  +---------- day of month (1 - 31)
# |  |  |  +------- month (1 - 12)
# |  |  |  |  +---- day of week (0 - 6) (Sunday=0 or 7)
# |  |  |  |  |
# *  *  *  *  *  command to be executed
# At 2am every day, re-index disk usage of /data/srlab with duc
0 2 * * * /PHShome/USERNAME/.local/bin/duc index /data/srlab 2>/dev/null
# 26th of each month send a report
5 8 26 * * /PHShome/USERNAME/ 2>/dev/null

Email reports#

Here is the bash script that we will call on the 26th of each month to email a report to multiple recipients:

#!/usr/bin/env bash

# Recipients of the email.

# Full path to the duc executable.

# The directory of interest.
mkdir -p "${dir}/disk-usage"

# The date for today.
day=$($duc info | grep "$dir" | cut -f1 -d' ')

# Size of each subdirectory.
  paste <(
    $duc info | head -n1 | tr ' ' '\n' | grep '\w'
  ) <(
    $duc info | tail -n+2 | head -n1 | tr ' ' '\n' | grep '\w'

# Build a complete message.
  echo "$info" && echo && $duc ls "$dir"

# Write the message to a text file on disk.
echo "$message" > "${dir}/disk-usage/${day}.txt"

# Send the message in an email.
echo "$message" | mail -s "Partners disk usage: '$dir' $day" "$rec"

The body of the email looks like this:

Date    2018-12-26
Time    02:00:01
Files   10.3M
Dirs    1.1M
Size    118.7T
Path    /data/srlab

 16.4T abcdefg
 14.8T abcdef
 14.3T ghijkl
 13.7T abcdefghijklm
 12.3T nopqr
 11.9T abcde
  7.6T fghijklm
  4.2T abcd

In practice#

I actually used this script to email myself and monitor the disk usage in Soumya Raychaudhuri’s lab at Harvard. After we let the job repeat for a few months, we can see how quickly the disk usage is increasing over time:

Disk usage on a Partners server

We can also see how each subdirectory has changed over time. Below, you might notice that the green qox subdirectory has grown from 1 TB to over 20 TB in less than a year.

A few subdirectories like demyt or xrcwq have not changed over the entire time course.

Disk usage on a Partners server

The real names of the subdirectories are masked for privacy.


Here is the R script for reading the text files and creating the plots in this note: