A previous version of this tutorial was written by Shaun Lewis.
Cron is a time-based job scheduling daemon found in Unix-like operating systems, including Linux distributions. Cron runs in the background and tasks scheduled with cron, referred to as “cron jobs,” are executed automatically, making cron useful for automating maintenance-related tasks.
This guide provides an overview of how to schedule tasks using cron’s special syntax. It also goes over a few shortcuts one can use to make job schedules easier to write and understand.
To complete this guide, you’ll need access to a computer running CentOS 8. This could be your local machine, a virtual machine, or a virtual private server.
Regardless of what kind of computer you use to follow this guide, it should have a non-root user with administrative privileges configured. To set this up, follow our Initial Server Setup guide for CentOS 8.
Almost every Linux distribution has some form of cron installed by default. However, if you’re using a CentOS machine on which cron isn’t installed, you can install it using
Before installing cron on a CentOS machine, update the computer’s local package index:
- sudo dnf update
Then install the cron daemon with the following command:
- sudo dnf install crontabs
This command will prompt you to confirm that you want to install the
crontabs package and its dependencies. Do so by pressing
This will install cron on your system, but you’ll need to start the daemon manually. You’ll also need to make sure it’s set to run whenever the server boots. You can perform both of these actions with the
To start the cron daemon, run the following command:
- sudo systemctl start crond.service
To set cron to run whenever the server starts up, type:
- sudo systemctl enable crond.service
Following that, cron will be installed on your system and ready for you to start scheduling jobs.
Understanding How Cron Works
Cron jobs are recorded and managed in a special file known as a
crontab. Each user profile on the system can have their own crontab where they can schedule jobs, which is stored under
To schedule a job, you just need to open up your crontab for editing and add a task written in the form of a cron expression. The syntax for cron expressions can be broken down into two elements: the schedule and the command to run.
The command can be virtually any command you would normally run on the command line. The schedule component of the syntax is broken down into 5 different fields, which are written in the following order:
|Day of the month||
|Day of the week||
Together, tasks scheduled in a crontab are structured like this:
minute hour day_of_month month day_of_week command_to_run
Here’s a functional example of a cron expression. This expression runs the command
curl http://www.google.com every Tuesday at 5:30 PM:
30 17 * * 2 curl http://www.google.com
There are also a few special characters you can include in the schedule component of a cron expression to make scheduling easier:
*: In cron expressions, an asterisk is a wildcard variable that represents “all.” Thus, a task scheduled with
* * * * * ...will run every minute of every hour of every day of every month.
,: Commas break up scheduling values to form a list. If you want to have a task run at the beginning and middle of every hour, rather than writing out two separate tasks (e.g.,
0 * * * * ...and
30 * * * * ...), you could achieve the same functionality with one (
0,30 * * * * ...).
-: A hyphen represents a range of values in the schedule field. Instead of having 30 separate scheduled tasks for a command you want to run for the first 30 minutes of every hour (as in
0 * * * * ...,
1 * * * * ...,
2 * * * * ..., and so on), you could just schedule it as
0-29 * * * * ....
/: You can use a forward slash with an asterisk to express a step value. For example, instead of writing out eight separate separate cron tasks to run a command every three hours (as in,
0 0 * * * ...,
0 3 * * * ...,
0 6 * * * ..., and so on), you could schedule it to run like this:
0 */3 * * * ....
Note: You cannot express step values arbitrarily; you can only use integers that divide evenly into the range allowed by the field in question. For instance, in the “hours” field you could only follow a forward slash with
Here are some more examples of how to use cron’s scheduling component:
* * * * *– Run the command every minute.
12 * * * *– Run the command 12 minutes after every hour.
0,15,30,45 * * * *– Run the command every 15 minutes.
*/15 * * * *– Run the command every 15 minutes.
0 4 * * *– Run the command every day at 4:00 AM.
0 4 * * 2-4– Run the command every Tuesday, Wednesday, and Thursday at 4:00 AM.
20,40 */8 * 7-12 *– Run the command on the 20th and 40th minute of every 8th hour every day of the last 6 months of the year.
If you find any of this confusing or if you’d like help writing schedules for your own cron tasks, Cronitor provides a handy cron schedule expression editor named “Crontab Guru” which you can use to check whether your cron schedules are valid.
Once you’ve settled on a schedule and you know the job you want to run, you’ll need to put it somewhere your daemon will be able to read it.
As mentioned previously, a crontab is a special file that holds the schedule of jobs cron will run. However, these are not intended to be edited directly. Instead, it’s recommended that you use the
crontab command. This allows you to edit your user profile’s crontab without changing your privileges with
crontab command will also let you know if you have syntax errors in the crontab, while editing it directly will not.
You can edit your crontab with the following command:
- crontab -e
This will open up your crontab in your user profile’s default text editor.
Note: On new CentOS 8 servers, the
crontab -e command will open up your user’s crontab with
vi by default.
vi is an extremely powerful and flexible text editor, but it can feel somewhat obtuse for users who lack experience with it.
If you’d like to use a more approachable text editor as your default crontab editor, you could install and configure
nano as such.
To do so, install
- sudo dnf install nano
When prompted, press
y and then
ENTER to confirm that you want to install
nano as your user profile’s default visual editor, open up the
.bash_profile file for editing. Now that you’ve installed it, you can do so with
- nano ~/.bash_profile
At the bottom of the file, add the following line:
. . . export VISUAL="nano"
This sets the
VISUAL environment variable to
VISUAL is a Unix environment variable that many programs — including crontab — invoke to edit a file. After adding this line, save and close the file by pressing
CTRL + X,
.bash_profile so the shell picks up the new change:
- . ~/.bash_profile
Once in the editor, you can input your schedule with each job on a new line. Otherwise, you can save and close the crontab for now. If you opened your crontab with
vi, the default CentOS 8 text editor, you can do so by pressing
ESC to make sure you’re in
vi’s command mode, then type
:x and press
Please note that, on Linux systems, there is another crontab stored under the
/etc/ directory. This is a system-wide crontab that has an additional field for which user profile each cron job should be run under. This tutorial focuses on user-specific crontabs, but if you wanted to edit the system-wide crontab, you could do so with the following command:
- sudo nano /etc/crontab
If you’d like to view the contents of your crontab, but not edit it, you can use the following command:
- crontab -l
You can erase your crontab with the following command:
Warning: The following command will not ask you to confirm that you want to erase your crontab. Only run it if you are certain that you want to erase it.
- crontab -r
This command will delete the user’s crontab immediately. However, you can include the
-i flag to have the command prompt you to confirm that you actually want to delete the user’s crontab:
- crontab -r -i
Outputcrontab: really delete sammy's crontab?
When prompted, you must enter
y to delete the crontab or
n to cancel the deletion.
Managing Cron Job Output
Because cron jobs are executed in the background, it isn’t always apparent that they’ve run successfully. Now that you know how to use the
crontab command and how to schedule a cron job, you can start experimenting with some different ways of redirecting the output of cron jobs to help you track that they’ve been executed successfully.
If you have a mail transfer agent — such as Sendmail — installed and properly configured on your server, you can send the output of cron tasks to the email address associated with your Linux user profile. You can also manually specify an email address by providing a
MAILTO setting at the top of the crontab.
For example, you could add the following lines to a crontab. These include a
MAILTO statement followed by an example email address, a
SHELL directive that indicates the shell to run (
bash in this example), a
HOME directive pointing to the path in which to search for the cron binary, and a single cron task:
. . . MAILTO="[email protected]" SHELL=/bin/bash HOME=/ * * * * * echo ‘Run this command every minute’
This particular job will return “Run this command every minute,” and that output will get emailed every minute to the email address specified after the
You can also redirect a cron task’s output into a log file or into an empty location to prevent getting an email with the output.
To append a scheduled command’s output to a log file, add
>> to the end of the command followed by the name and location of a log file of your choosing, like this:
* * * * * echo ‘Run this command every minute’ >> /directory/path/file.log
Let’s say you want to use cron to run a script but keep it running in the background. To do so, you could redirect the script’s output to an empty location, like
/dev/null which immediately deletes any data written to it. For example, the following cron job executes a PHP script and runs it in the background:
* * * * * /usr/bin/php /var/www/domain.com/backup.php > /dev/null 2>&1
This cron job also redirects standard error — represented by
2 — to standard output (
>&1). Because standard output is already being redirected to
/dev/null, this essentially allows the script to run silently. Even if the crontab contains a
MAILTO statement, the command’s output won’t be sent to the specified email address.
You can manage which users are allowed to use the
crontab command with the
cron.deny files, both of which are stored in the
/etc/ directory. If the
cron.deny file exists, any user listed in it will be barred from editing their crontab. If
cron.allow exists, only users listed in it will be able to edit their crontabs. If both files exist and the same user is listed in each, the
cron.allow file will override
cron.deny and the user will be able to edit their crontab.
For example, to deny access to all users and then give access to the user ishmael, you could use the following command sequence:
- sudo echo ALL >>/etc/cron.deny
- sudo echo ishmael >>/etc/cron.allow
First, we lock out all users by appending
ALL to the
cron.deny file. Then, by appending the username to the
cron.allow file, we give the ishmael user profile access to execute cron jobs.
Note that if a user has
sudo privileges, they can edit another user’s crontab with the following command:
- sudo crontab -u user -e
cron.deny exists and user is listed in it and they aren’t listed in
cron.allow, you’ll receive the following error after running the previous command:
OutputThe user user cannot use this program (crontab)
By default, most cron daemons will assume all users have access to cron unless either
There are also several shorthand commands you can use in your crontab file to help streamline job scheduling. They are essentially shortcuts for the equivalent numeric schedule specified:
Note: Not all cron daemons can parse this syntax (particularly older versions), so double-check it works before you rely on it.
@reboot shorthand will run whatever command follows it any time the server starts up:
@reboot echo "System start up"
Using these shortcuts whenever possible can help make it easier to interpret the schedule of tasks in your crontab.
Cron is a flexible and powerful utility that can reduce the burden of many tasks associated with system administration. When combined with shell scripts, you can automate tasks that are normally tedious or complicated.