read

I introduced sleep and gtimeout in my post on some nice bash commands available in macOS.

The advanced way of scheduling job is to run cron jobs.

Crontab sets up cron jobs and runs them at specific time, and repeatedly.

One frequent use case I have is to schedule my computer to download live youtube streams at wee hours, using streamlink.

Basic Cron Job

To edit the cron jobs, run:

env EDITOR=nano crontab -e

We use nano text editor as that is easier. You need to save and exit the editor every time you finish editing it.

Enter a line of text, representing a cron job, like this:

59 23 * * * cd ~/Documents && /usr/local/bin/streamlink https://www.youtube.com/watch\?v\=y3oW28Draso best >/tmp/stdout.log 2>/tmp/stderr.log

This long line:

• runs the command on 23:59 every day
• cd to a directory
• then run the streamlink command, with full path to the command because cron is run without knowing the env var PATH
• writes the output to /tmp/stdout.log and error to /tmp/stderr.log

Use a script

To keep things tidy in crontab, I will often write the commands in a script eg “job.sh” in ~/Documents.

cd ~/Documents
touch job.sh
chmod +x job.sh

The content of “job.sh”:

#!/bin/sh

PATH=/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin:/usr/local/go/bin

cd ~/Documents

now=`date "+%Y-%m-%d %H%M"`

streamlink -o "$now.ts" https://www.youtube.com/watch\?v\=9kVxPmh8eSQ best

• set the PATH env var, so that all common executables are available
• use date for the file name eg. “2018-07-07 0810.ts”
• calls streamlink to download live youtube stream

Then in crontab, each job is simplified to just running the script.

59 23 * * * ~/Documents/job.sh >/tmp/stdout.log 2>/tmp/stderr.log

Other Tips

# List the processes with name streamlink
ps aux | grep streamlink

# Kill the process with the id eg. 26959
kill -9 26959

# Kill all streamlink processes in 1 command:
ps aux | grep streamlink | awk '{print $2}' | xargs kill -9