Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

My crontab looks something like this;

1 * * * * /var/www/cron/site1.sh > /dev/null 2>&1

0 * * * * /var/www/cron/site2.sh > /dev/null 2>&1

3 * * * * /var/www/cron/site3.sh > /dev/null 2>&1

This works great and lets me place all the nasty little script calls into one place, rather than making crontab harder to read than it already is.

But, this fails massively when site2.sh needs one script to run once a day, another to run once a week and another to run every 5 minutes. And of course it gets worse as new scripts get added with different timings.

Is there a better way?

EDIT

By better I mean making it more manageable, having a large crontab is not manageable, but neither is having scripts all over the place.

Not a GUI necessarily.

share|improve this question

migrated from stackoverflow.com Nov 8 '12 at 4:23

This question came from our site for professional and enthusiast programmers.

    
Better in what sense? a GUI? Welcome to the dark world of crontab -e –  nickhar Nov 7 '12 at 21:57
    
Nope, not a GUI. I will edit the question. –  Jake N Nov 7 '12 at 22:25
add comment

2 Answers 2

up vote 5 down vote accepted

Making crontab more manageable, something I've definitely dreamed of many times. When it comes down to it though, if you need 20 scripts to be scheduled to run - well - you need 20 scripts to be scheduled to run.

The primary issue, at least for me, has always been creating a new entry in /etc/crontab (or in a user's crontab -e). Especially when these new entries are set to run at the same time as an existing entry (or even a list of entries).

Say, for instance, you needed 5 different scripts to run on-the-hour every-hour. You could create 5 different entries in crontab for this - or - you can take advantage of run-parts. run-parts allows you to specify a crontab entry just like any other entry, but the "command" is actually a directory. When it's the command's time to execute, it will execute every script in the given directory.

Here's a list that are common to linux systems:

01 * * * * root run-parts /etc/cron.hourly
02 4 * * * root run-parts /etc/cron.daily
22 4 * * 0 root run-parts /etc/cron.weekly
42 4 1 * * root run-parts /etc/cron.monthly

In the first line, /etc/cron.hourly states that every hour at 1-minute-past-the-hour (01 * * * *), all of the scripts that are located in the /etc/cron.hourly directory will be executed. The same concept is applied to the other lines as well. You're not limited to these times and directories either - you can fully customize them to suit your needs as a standard crontab entry.

Other than batching-scripts by directories with run-parts, you can make things more manageable by separating them by "users" and putting them into each user's crontab via crontab -u username -e. Say, for instance, you have a handful of scripts specific to reporting, you could create a "reportRunner" user and assign all report-related crons to that user. Doing this, you'll easily have list-separation and it would be easy to manage the different categories of scripts/schedules (in my opinion). It won't make the total list of crons shorter, but it should help make any-given list you're looking at shorter (and categorized).

share|improve this answer
    
Thanks! Just for clarification, the folder /etc/cron.hourly already exists, does this mean I still need to add 01 * * * * root run-parts /etc/cron.hourly to my crontab? Or does it's existence mean that crontab is already aware of it and it'll be run anyway? –  Jake N Nov 9 '12 at 12:39
    
Hmmm, I have set it up and the only output I am getting via email is /bin/sh: root: not found –  Jake N Nov 9 '12 at 12:56
    
This seems to be a user that the cron is trying to run as, but my cron is run as www-data and not root. Changing root to www-data still fails with /bin/sh: www-data: not found –  Jake N Nov 9 '12 at 13:16
    
@jakenoble If you're using run-parts inside of a crontab -e editor for a user, you don't need to specify the username. Your line would look just like 01 * * * * run-parts /etc/cron.hourly. You'd only need to specify the username when you're directly editing /etc/crontab –  newfurniturey Nov 9 '12 at 13:43
    
this still fails, or rather does nothing, I am using 38 * * * * run-parts /var/www/cron/cron.hourly then in /var/www/cron/cron.hourly I have a file called test which just does touch /var/www/cron.txt. Any ideas? –  Jake N Nov 13 '12 at 18:00
show 1 more comment

I've been party to some fairly horrific crontabs and there's no getting around them. However, I've seen a few good ground rules:

  1. Create structure
  2. Group scripts by site/name/function
  3. Annotate everything - that way, others can see what is going on.
  4. Group crons by site (if you have many) eg. /var/www/site/cron/

-----8<-----

#### Site A ####

  #--- tasks ---#

    # cleanup sessions (ev. 20mins)
      */20 * * * * /var/www/cron/cleanup.script > /dev/null 2>&1

    # Garbage collection (ev. 24 hrs @ 0000)
      0 0 * * * /var/www/cron/gcoll.script > /dev/null 2>&1

    # De-dupe id's from x table (ev. 18 mins)
      */18 * * * * /var/www/cron/dedupe.script > /dev/null 2>&1


  #--- Reporting ---#

    # Generate mod reports (ev. 24 hrs @ 0400)
      0 4 * * * /var/www/cron/daily_report.script > /dev/null 2>&1

#### /Site A ####
share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.