As I was building my slackbot, I was always running it locally. The problem was that I had two computers where I could be deving on it, and occasionally I would have two instances running. This was a pain as it would make two post in the slack room where it was being used, or if both computers were off, no one could use it. My friend was nice enough to give me ssh access to his raspberry pi so I could run my bot there and not have to buy server space or leave a dedicated computer running at all times. And while it was a great solution for where to run the bot to prevent the issue of too many instances or no instances, this presented a new problem, deploying. Having to ssh in and go through the deploy process every time I knew would eventually get too annoying. I knew I needed some kind of continuous integration and auto deploy process, but I didn’t want to pay for any services, nor for any servers to get something like Jenkins set up. So I decided to build my own solution using cron and a simple script.

Shell Scripting

Like any good developer, as I was preparing to write the shell script I had to lay out the steps needed to make it work. I knew I needed to figure out if there were any new commits, if so, deploy. Naturally, the next step was deciding what a deploy consisted of. Deploying was as simple as killing the current running instance, gitting the new code, building, and running again in the background. The script below fetches the repo and greps the status for the word “behind”, then simply checks if it was found, and if so runs the commands needed to deploy.

cd ~/starbot
git fetch
git=$(git status -uno | grep "behind")
if [[ ! $git == "" ]]; then
        ps -ef | grep "starbot" | awk '{print $2}' | xargs kill -9
        git pull
        mvn clean package
        java -jar target/*-dependencies.jar > /tmp/starbot.log &

If you wanted to use this command yourself, just be sure to change the ‘cd’ to the correct directory if it’s different from where the script lives, otherwise delete the line. After, change the commands in the if block to be specific to your application’s deploy process and then set cron to run the script.


Cron jobs run a process at a specified interval. In this case, I want it to check for changes and redploy if they are there once a minute. To add a new cron job first run crontab -e. It will most likely ask you to choose an editor to continue. After you have the file opened simply add the following line */1 * * * * bash /home/scott/

The */1 * * * * specifies an interval to run at, in our case one minute. you can also specify to run every hour on the 15 minute mark, or over the course of weeks, days, or months. I found this tutorial to be quite useful when I was trying to figure the intervals out as well as cron in general.


Once you have your cron job set to run, you’re set, just push to the repo and it will poll at the specified interval and redeploy for you! I use this method to auto deploy my blog now as well in combination with the deploy script I wrote about in another blog post. Thanks for reading.