A command works in my terminal but fails or does nothing when run from cron (crontab). Why, and how do I debug cron jobs?

Cause: cron runs jobs with a minimal environment, not your login shell's. PATH is usually just /usr/bin:/bin, your shell profile (.bashrc, .profile) isn't loaded, the shell is /bin/sh unless set, the working directory is your home directory, and there's no terminal. So commands found through your PATH (node, python from a venv, tools in /usr/local/bin or ~/.local/bin), environment variables (API keys, virtualenvs, nvm), and relative paths all break. Output and errors go to local mail (often not set up), so failures look silent. Fix: 1. Capture output first: append >> /tmp/myjob.log 2>&1 to the crontab line, or set MAILTO to an address that works. Read the log after the next run. 2. Use absolute paths for the program and files: /usr/local/bin/node /home/me/app/job.js, /home/me/app/.venv/bin/python script.py. 3. Set what the job needs at the top of the crontab (cron supports simple VAR=value lines, as documented in crontab(5)): PATH=/usr/local/bin:/usr/bin:/bin, SHELL=/bin/bash. Or put everything in a wrapper script that cds to the right directory and sets the environment, and call that. 4. Escape percent signs: in a crontab line % means newline unless escaped as \% (common with date +%Y-%m-%d). 5. Check the schedule and that cron is running: the five time fields, the user's crontab (crontab -l) vs /etc/cron.d files (which need a user field), the system timezone, and systemctl status cron (or crond). The system log (journalctl -u cron, or /var/log/syslog) shows whether the job started. 6. Don't store secrets in the crontab itself; read them from a protected file or a secrets manager in the wrapper script. Verify: set the job to run every minute temporarily, watch the log file, then restore the real schedule. Sources: https://man7.org/linux/man-pages/man5/crontab.5.html https://man7.org/linux/man-pages/man8/cron.8.html