cron

From ArchWiki
(Redirected from Crontab)
Jump to: navigation, search

From Wikipedia:

cron is the time-based job scheduler in Unix-like computer operating systems. cron enables users to schedule jobs (commands or shell scripts) to run periodically at certain times or dates. It is commonly used to automate system maintenance or administration.

Installation

There are many cron implementations, but none of them are installed by default as the base system uses systemd/Timers instead. See the Gentoo's cron guide, which offers comparisons.

Packages available:

Configuration

Activation and autostart

After installation, the daemon will not be enabled by default. The installed package likely provides a service, which can be controlled by systemctl. For example, cronie uses cronie.service.

Check /etc/cron.daily/ and similar directories to see which jobs are present. Activating cron service will trigger all of them.

Note: cronie provides the 0anacron hourly job, which allows for delayed runs of other jobs e.g. if the computer was switched off at the moment of standard execution.

Handling errors of jobs

cron registers the output from stdout and stderr and attempts to send it as email to the user's spools via the sendmail command. Cronie disables mail output if /usr/bin/sendmail is not found. In order for mail to be written to a user's spool, there must be an smtp daemon running on the system, e.g. opensmtpd. Otherwise, you can install a package that provides the sendmail command, and configure it to send mail to a remote mail exchanger. You can also log the messages by using the -m option and writing a custom script.

  1. Edit the cronie.service unit.
  2. Install esmtp, msmtp, opensmtpd or write a custom script.

Example with msmtp

Tango-view-fullscreen.pngThis article or section needs expansion.Tango-view-fullscreen.png

Reason: How exactly do I make sure it detects the new sendmail command? (Discuss in Talk:Cron#)

Install msmtp-mta which creates a symbolic link from /usr/bin/sendmail to /usr/bin/msmtp. Restart cronie to make sure it detects the new sendmail command. You must then provide a way for msmtp to convert your username into an email address.

Then either add MAILTO line to your crontab, like so:

MAILTO=your@email.com

or create /etc/msmtprc and append this line:

aliases /etc/aliases

and create /etc/aliases:

your_username: your@email.com
# Optional:
default: your@email.com

Then modify the configuration of cronie daemon by replacing the ExecStart command with:

ExecStart=/usr/bin/crond -n -m '/usr/bin/msmtp -t'

Example with esmtp

Install esmtp and procmail.

After installation configure the routing:

/etc/esmtprc
identity myself@myisp.com
       hostname mail.myisp.com:25
       username "myself"
       password "secret"
       starttls enabled
       default
mda "/usr/bin/procmail -d %T"

Procmail needs root privileges to work in delivery mode but it is not an issue if you are running the cronjobs as root anyway.

To test that everything works correctly, create a file message.txt with "test message" in it.

From the same directory run:

$ sendmail user_name < message.txt 

then:

$ cat /var/spool/mail/user_name

You should now see the test message and the time and date it was sent.

The error output of all jobs will now be redirected to /var/spool/mail/user_name.

Due to the privileged issue, it is hard to create and send emails to root (e.g. su -c ""). You can ask esmtp to forward all root's email to an ordinary user with:

/etc/esmtprc
force_mda="user-name"
Note: If the above test didn't work, you may try creating a local configuration in ~/.esmtprc with the same content.

Run the following command to make sure it has the correct permission:

$ chmod 710 ~/.esmtprc
Then repeat the test with message.txt exactly as before.

Example with opensmtpd

Install opensmtpd.

Edit /etc/smtpd/smtpd.conf. The following configuration allows for local delivery:

listen on localhost
accept for local deliver to mbox

You can proceed to test it. First start smtpd.service. Then do:

$ echo test | sendmail user

user can check his/her mail in with any reader able to handle mbox format, or just have a look at the file /var/spool/mail/user. If everything goes as expected, you can enable opensmtpd for future boots.

This approach has the advantage of not sending local cron notifications to a remote server. On the downside, you need a new daemon running.

Note:
  • At the moment of writing the Arch opensmtpd package does not create all needed directories under /var/spool/smtpd, but the daemon will warn about that specifying the required ownerships and permissions. Just create them as suggested.
  • Even though the suggested configuration does not accept remote connections, it's a healthy precaution to add an additional layer of security blocking port 25 with iptables or similar.

Long cron job

Suppose this program is invoked by cron :

#!/bin/sh
echo "I had a recoverable error!"
sleep 1h

What happens is this:

  1. cron runs the script
  2. as soon as cron sees some output, it runs your MTA, and provides it with the headers. It leaves the pipe open, because the job hasn't finished and there might be more output.
  3. the MTA opens the connection to postfix and leaves that connection open while it waits for the rest of the body.
  4. postfix closes the idle connection after less than an hour and you get an error like this :
smtpmsg='421 … Error: timeout exceeded' errormsg='the server did not accept the mail'

To solve this problem you can use the command chronic or sponge from moreutils. From their respective man page:

chronic
chronic runs a command, and arranges for its standard out and standard error to only be displayed if the command fails (exits nonzero or crashes). If the command succeeds, any extraneous output will be hidden.
sponge
sponge reads standard input and writes it out to the specified file. Unlike a shell redirect, sponge soaks up all its input before opening the output file… If no output file is specified, sponge outputs to stdout.

Chronic too buffers the command output before opening its standard output.

Crontab format

The basic format for a crontab is:

minute hour day_of_month month day_of_week command
  • minute values can be from 0 to 59.
  • hour values can be from 0 to 23.
  • day_of_month values can be from 1 to 31.
  • month values can be from 1 to 12.
  • day_of_week values can be from 0 to 6, with 0 denoting Sunday.

Multiple times may be specified with a comma, a range can be given with a hyphen, and the asterisk symbol is a wildcard character. Spaces are used to separate fields. For example, the line:

*/5 9-16 * 1-5,9-12 1-5 ~/bin/i_love_cron.sh

Will execute the script i_love_cron.sh at five minute intervals from 9 AM to 4:55 PM on weekdays except during the summer months (June, July, and August). More examples and advanced configuration techniques can be found below.

Besides, crontab has some special keywords:

@reboot at startup 
@yearly once a year
@annually ( == @yearly)
@monthly once a month
@weekly once a week
@daily once a day
@midnight ( == @daily)
@hourly once an hour

For example:

@reboot ~/bin/i_love_cron.sh

Will execute the script i_love_cron.sh at startup.

See more at: http://www.linuxandlife.com/2013/01/how-to-use-crontab-to-schedule-tasks.html#sthash.AXvOdsAz.dpuf

Basic commands

Crontabs should never be edited directly; instead, users should use the crontab program to work with their crontabs. To be granted access to this command, user must be a member of the users group (see the gpasswd command).

To view their crontabs, users should issue the command:

$ crontab -l

To edit their crontabs, they may use:

$ crontab -e
Note: By default the crontab command uses the vi editor. To change it, export EDITOR or VISUAL, or specify the editor directly: EDITOR=vim crontab -e.

To remove their crontabs, they should use:

$ crontab -r

If a user has a saved crontab and would like to completely overwrite their old crontab, he or she should use:

$ crontab saved_crontab_filename

To overwrite a crontab from the command line (Wikipedia:stdin), use

$ crontab - 

To edit somebody else's crontab, issue the following command as root:

# crontab -u username -e

This same format (appending -u username to a command) works for listing and deleting crontabs as well.

Examples

The entry:

01 * * * * /bin/echo Hello, world!

runs the command /bin/echo Hello, world! on the first minute of every hour of every day of every month (i.e. at 12:01, 1:01, 2:01, etc.).

Similarly:

*/5 * * jan mon-fri /bin/echo Hello, world!

runs the same job every five minutes on weekdays during the month of January (i.e. at 12:00, 12:05, 12:10, etc.).

The line (as noted in "man 5 crontab"):

*0,*5 9-16 * 1-5,9-12 1-5 /home/user/bin/i_love_cron.sh

will execute the script i_love_cron.sh at five minute intervals from 9 AM to 5 PM (excluding 5 PM itself) every weekday (Mon-Fri) of every month except during the summer (June, July, and August).

Periodical settings can also be entered as in this crontab template:

# Chronological table of program loadings                                       
# Edit with "crontab" for proper functionality, "man 5 crontab" for formatting
# User: johndoe

# mm  hh  DD  MM  W /path/progam [--option]...  ( W = weekday: 0-6 [Sun=0] )
  21  01  *   *   * /usr/bin/systemctl hibernate
  @weekly           $HOME/.local/bin/trash-empty

Default editor

To use an alternate default editor, define the EDITOR environment variable in a shell initialization script as described in Environment variables.

As a regular user, su will need to be used instead of sudo for the environment variable to be pulled correctly:

$ su -c "crontab -e"

To have an alias to this printf is required to carry the arbitrary string because su launches in a new shell:

alias scron="su -c $(printf "%q " "crontab -e")"

run-parts issue

cronie uses run-parts to carry out script in cron.daily/cron.weekly/cron.monthly. Be careful that the script name in these won't include a dot (.), e.g. backup.sh, since run-parts without options will ignore them (see: man run-parts).

Running X.org server-based applications

Cron does not run under the X.org server therefore it cannot know the environmental variable necessary to be able to start an X.org server application so they will have to be defined. One can use a program like xuserrunAUR[broken link: archived in aur-mirror] to do it:

17 02 * ... /usr/bin/xuserrun /usr/bin/xclock

Or then can be defined manually (echo $DISPLAY will give the current DISPLAY value):

17 02 * ... env DISPLAY=:0 /usr/bin/xclock

If done through say SSH, permission will need be given:

# xhost +si:localuser:$(whoami)

Asynchronous job processing

If you regularly turn off your computer but do not want to miss jobs, there are some solutions available (easiest to hardest):

Cronie

cronie comes with anacron included. The project homepage reads:

Cronie contains the standard UNIX daemon crond that runs specified programs at scheduled times and related tools. It is based on the original cron and has security and configuration enhancements like the ability to use pam and SELinux.

Dcron

Vanilla dcronAUR supports asynchronous job processing. Just put it with @hourly, @daily, @weekly or @monthly with a jobname, like this:

@hourly         ID=greatest_ever_job      echo This job is very useful.

Cronwhip

cronwhipAUR is a script to automatically run missed cron jobs; it works with the former default cron implementation, dcron. See also the forum thread.

Anacron

Anacron is a full replacement for dcron which processes jobs asynchronously.

It is provided by cronie. The configuration file is /etc/anacrontab. Information on the format can be found in the anacrontab(5) man page. Running anacron -T will test /etc/anacrontab for validity.

Fcron

Like anacron, fcron assumes the computer is not always running and, unlike anacron, it can schedule events at intervals shorter than a single day which may be useful for systems which suspend/hibernate regularly (such as a laptop). Like cronwhip, fcron can run jobs that should have been run during the computer's downtime.

When replacing cronie with fcron be aware the spool directory is /var/spool/fcron and the fcrontab command is used instead of crontab to edit the user crontabs. These crontabs are stored in a binary format with the text version next to them as foo.orig in the spool directory. Any scripts which manually edit user crontabs may need to be adjusted due to this difference in behavior.

A quick scriptlet which may aide in converting traditional user crontabs to fcron format:

cd /var/spool/cron && (
 for ctab in *; do
  fcrontab ${ctab} -u ${ctab}
 done
)

See also the forum thread.

Ensuring exclusivity

If you run potentially long-running jobs (e.g., a backup might all of a sudden run for a long time, because of many changes or a particular slow network connection), then flock (util-linux) can ensure that the cron job won't start a second time.

  5,35 * * * * /usr/bin/flock -n /tmp/lock.backup /root/make-backup.sh

cronie

Long time users of vixie-cron (traditional cron) will be confused by how cronie is set up. Here is the relevant file hierarchy:

   /etc
     |----- anacrontab
     |----- cron.d
              | ----- 0hourly
     |----- cron.daily
     |----- cron.deny
     |----- cron.hourly
              | ----- 0anacron
     |----- cron.monthly
     |----- cron.weekly
     |----- crontab


Note that the crontab file is not created by default, but jobs added here will be run if you wish to use this file. Cronie provides both cron and anacron functionality. The difference is that cron will run jobs at particular time intervals (down to a granularity of one minute) if the machine is on at the particular time specified, while anacron runs jobs (with a minimum daily granularity) without assuming that the machine is turned on all the time. When the machine is on, anacron will check to see if there are any jobs that should have been run and will run them accordingly. The /etc/cron.d and /etc/cron.hourly directories are associated with cron functionality, while the /etc/anacrontab file and /etc/cron.daily, /etc/cron.weekly, and /etc/cron.monthly directories are associated with anacron functionality. The /etc/cron.deny file is there so that any user who is not specifically prohibited can create their own cron jobs.

To implement a system-wide cron job, create a crontab-like file for it and place it in the /etc/cron.d directory or add the job to /etc/crontab. Any executable (these are almost always shell scripts) in /etc/cron.hourly will be run every hour.

Anacron functionality is implemented similarly, however using the /etc/cron.daily, /etc/cron.weekly, or /etc/cron.monthly directories, depending on how frequently you want the job to be run. The anacron job files are also executables; i.e. not in crontab-format. Anacron is triggered at the beginning of every hour by the crontab file /etc/cron.d/0hourly which runs the executables in /etc/cron.hourly including the file /etc/cron.hourly/0anacron - deleting these will prevent anacron running any daily, weekly or monthly tasks.

Note: the output of systemctl status cronie might show a message such as crond[<PID>]: (root) CAN'T OPEN (/etc/crontab): No such file or directory. However, this is not an error as of cronie 1.4.8. See this discussion.

Dcron

The cron daemon parses a configuration file known as crontab. Each user on the system can maintain a separate crontab file to schedule commands individually. The root user's crontab is used to schedule system-wide tasks (though users may opt to use /etc/crontab or the /etc/cron.d directory, depending on which cron implementation they choose).

/var/spool/cron/root
# Run command at a scheduled time
# Edit this 'crontab -e' for error checking, man 1 crontab for acceptable format

# <@freq>                       <tags and command>
@hourly         ID=sys-hourly   /usr/sbin/run-cron /etc/cron.hourly
@daily          ID=sys-daily    /usr/sbin/run-cron /etc/cron.daily
@weekly         ID=sys-weekly   /usr/sbin/run-cron /etc/cron.weekly
@monthly        ID=sys-monthly  /usr/sbin/run-cron /etc/cron.monthly

# mm  hh  DD  MM  W /path/command (or tags) # W = week: 0-6, Sun=0
  21  01  *   *   * /usr/bin/systemctl suspend

These lines exemplify one of the formats that crontab entries can have, namely whitespace-separated fields specifying:

  1. @period
  2. ID=jobname (this tag is specific to dcron)
  3. command

The other standard format for crontab entries is:

  1. minute
  2. hour
  3. day
  4. month
  5. day of week
  6. command

The crontab files themselves are usually stored as /var/spool/cron/username. For example, root's crontab is found at /var/spool/cron/root

See the crontab man page for further information and configuration examples.

See also