It is always a struggle configuring a cronjob to run certain scripts; there is always a mis configuration, a PATH, an environment variable that is different from when you use it in interactive mode, etc… Do you have any tricks to avoid or minimize those issues?

cc @[email protected]

  • skilltheamps@feddit.de
    link
    fedilink
    arrow-up
    10
    ·
    1 year ago

    Use a systemd-service + systemd-timer. You can then run “systemctl start myjob.service” to check that it runs as you expect. If it works “systemctl enable --now myjob.timer” to kick it off as scheduled

  • nitrolife@rekabu.ru
    link
    fedilink
    arrow-up
    6
    ·
    1 year ago

    You can use crontab -e from user. That set user environment to usual state. Or you can use systems timers instead crontab.

    Or you can set all environment variables inside cron file.

  • johntash@eviltoast.org
    link
    fedilink
    English
    arrow-up
    2
    ·
    1 year ago

    Depending on what your setup is like, you could use something other than cron.

    Kubernetes and Nomad both have cron-like schedulers. It’s convenient to keep your cronjob definitions near the app code.

    I’ve also been using windmill.dev lately for some simple cronjob-like tasks and it’s working pretty great. It also has a UI if that helps.

  • SokathHisEyesOpen@lemmy.ml
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    Put all the logic in a php file and just configure crontab to run the file at the given interval and output the results to a log file. Use crontab.guru to help with scheduling.

  • Coelacanthus@lemmy.kde.social
    link
    fedilink
    arrow-up
    1
    ·
    1 year ago

    Always use absolute path. If you write some shell snippet which may be run in different environment, avoid use just command name for uncommon one.

  • Kimusan@feddit.dk
    link
    fedilink
    arrow-up
    1
    ·
    1 year ago

    If Cron is too hard for you (it is really easy so I don’t really understand your problems) then check out runat instead. Simple and easy to use with all the features I would suspect you use in cron

  • thisisawayoflife@lemmy.world
    link
    fedilink
    arrow-up
    1
    ·
    1 year ago

    Wrap all the cron job tasks in a bash script and make sure it works. Call the script directly in the cron configuration. Dump the output of the cronjob to a logfile. Set it up to run every minute so you can ensure it works via the cron environment.

  • Lupec@lemm.ee
    link
    fedilink
    arrow-up
    1
    ·
    edit-2
    1 year ago

    I basically combine a few mechanisms to mitigate that:

    • Having cron call scripts deployed to a fixed location I can manually test or use on demand as needed
    • Piping each script’s output/exit code to healtchecks.io so I can get feedback notifications when/if they fail
    • Using Ansible to deploy everything so it’s all repeatable and version controlled