[2024-feb-29] Sad news: Eric Layton aka Nocturnal Slacker aka vtel57 passed away on Feb 26th, shortly after hospitalization. He was one of our Wiki's most prominent admins. He will be missed.

Welcome to the Slackware Documentation Project

Planification de Tâches dans Linux

TRADUCTION en cours — Cedric M. 2015/09/11 01:52


Cet article traite de certains outils utilisés dans un système Linux pour planifier des tâches à exécuter automatiquement à des intervalles de temps spécifiques ou à un moment donné. Cet article ne détaillera pas les commandes en profondeur; il est juste une brève introduction à l'utilisation de ces commandes. Lisez les HOWTOS individuels de chaques commandes pour plus de précisions sur leurs options.

Quelques daemons Linux/UNIX de planification de tâches :

NdT: Les liens suivants font référence aux versions originales.

  • at – planification de tâches unique
  • cron – le planificateur périodique le plus utilisé
  • anacron – anachronistic cron; a periodic scheduler that doesn't rely on the system being left on 24×7

Utiliser at

La commande at permet à un utilisateur d'exécuter des commandes ou des scripts à une date (obligatoire) et heure (facultatif) spécifiques. La commande peut être passé sur l'entrée standard, en redirection ou dans un fichier.

darkstar:~% at

at en mode Interractif

Utiliser la commande at avec l'entrée standard (avec le clavier) est un petit peu plus compliqué que de taper une ligne de commande au prompt. La commande utilise un “sub-shell” pour rassembler les information demandées. Une fois que l'entrée de la commande d'information est complète, Ctrl+D (EOT) signifiera la fin de l'entrée. L'argument -m spécifie qu'un message mèl sera envoyé à l'utilisateur lorsque le job sera terminé, à moins qu'une sortie ait été créée.

darkstar:~% at 12:01 -m
warning: commands will be executed using (in order) a) $SHELL b) login shell c) /bin/sh
at> ./my_script.sh
at> <EOT>
job 4 at 2015-06-22 12:01 

File-driven at

Commands can also be contained within a file and run by at:

darkstar:~% at 12:32 -m -f /usr/local/bin/my_script.sh
warning: commands will be executed using (in order) a) $SHELL b) login shell c) /bin/sh
job 8 at 2015-06-22 12:10

The -m flag will email the user after completion of the command; the -f flag specifies the command will read the job from a file, not from standard input. After the command is typed in (and the appropriate warning is displayed), the at job number1) is displayed.

at Internal Scheduling

The job numbers provided after a command is typed in, or when a file is read, allow the user to know which internal job will be run in sequential order. If a user wants to delete a specific task, all that needs to be known is this internal job number. To remove the job, the command atrm (at remove) is used:

darkstar:~%  at -l
7      2015-06-22 12:10 p tux
8      2015-06-22 12:15 p root

The command atq (at queue) is the same as at -l:

darkstar:~%  atq
7      2015-06-22 12:10 p tux
8      2015-06-22 12:15 p root

To remove the user job, use atrm with the job number:

darkstar:~%  atrm 7

Using cron

cron is a daemon that runs tasks in the background at specific times. For example, if you want to automate downloads of patches on a specific day (Monday), date (2 July), or time (1300), cron will allow you to set this up in a variety of ways. The flexibility inherent in cron can allow administrators and power users to automate repetitive tasks, such as creating backups and system maintenance.

cron is usually configured using a crontab file. The following command will open your user account crontab file:

darkstar:~% crontab -e

To edit the system-level crontab, first log into the root account:

darkstar:~# crontab -e

If your system has sudo installed, type in:

darkstar:~% sudo crontab -e 

The crontab file syntax is:

 # * * * * *  command to execute
 # │ │ │ │ │
 # │ │ │ │ │
 # │ │ │ │ └───── day of week (0 - 6) (Sun(0) /Mon (1)/Tue (2)/Wed (3)/Thu (4)/Fri (5)/Sat (6))
 # │ │ │ └────────── month (1 - 12)
 # │ │ └─────────────── day of month (1 - 31)
 # │ └──────────────────── hour (0 - 23)
 # └───────────────────────── min (0 - 59)

Using an asterisk in any placeholder location, will match any value. For example, the following will run example_script.sh at noon (1200) everyday during the first three months of the year:

#For more information see the manual pages of crontab(5) and cron(8)
# min hr day month weekday command
0 11 * 1-3 * /home/user/example_script.sh

Using anacron

anacron is not installed in Slackware by default.2)

anacron is unique from cron in the respect that it does not expect the operating system to be running continuously like a 24×7 server. If the time of execution passes while the system is turned off, anacron executes the command automatically when the machine is turned back on. The reverse is not true for cron - if the computer is turned off during the time of scheduled execution, cron will not execute the job. Another key difference between anacron and cron is the minimum chronological “granularity” - anacron can only execute jobs by day, versus the ability of cron to execute by the minute. Finally, anacron can only be used by root, while cron can be used by root and normal users.


As distinct from a process ID (PID) known to the operating system
See Slackbuilds.org for more information on anacron on Slackware
 fr:howtos:general_admin:task_scheduling ()