During this month I have been practicing some common DevOps activities like setting up a repository, opening a server and updating it automatically with cron. Some of them were easy tasks because I’ve already done similar activities, but there were other task that I had no idea how to do and it took me a while to figure it out.
The first part was easy, most of the step were things I already had like the Linux distribution(Ubuntu 16) and all the development environment. I had to struggle a little bit with cron because I din’t know how to make sure it was running, but after reading a couple of tutorials I found that I had to send the output to a file with the «>>» symbol. This operator appends anything I pass to the file without eliminating what was there before. This is an example of how my cron file looks:
* * * * * [path]/bashscript >> [path]/log.out
For the second part I also had experience using SSH keys for GitHub, I use that all the time now that I have 2FA because for the terminal it won’t let me use my normal password. To code the server I used NodeJS and ExpressJS because those were tools I was familiar with for this kind of things, later I had trouble with that but for this part it did what it had to do, the script moves to the git folder and pulled from master. Just to keep track of what was happening I kept the log file. During a period of time the script created a file and after doing all the stuff it had to do deleted it, that was surrounded with an if conditional that checked if the file was created. I deleted that on part 3 to simplify my tests because while I was working on it I had no idea what was giving me errors.
Finally, the part 3 was the most difficult for me. I quickly created a Maven project and in 5 minutes I could run the test without problems. The tricky part for me was that I could not find the way to read a file with the tests results and pass that to the web page. I tried a lot of ways and after a lot of failures and attempts to debug my code I learned that ExpressJS was not made to host HTML files. My HTML contained the JS that reads the tests results, that JS script was run in an environment without the proper modules so it didn’t work. My solution was to format the tests results files to have the proper HTML tags and then put that into a static HTML. This is how it looks:
It’s not pretty, but is useful…I guess. The code of the practices is available here
Overall, I enjoyed learning and using things that I don’t normally use, like cron or bash scripts. I think these tools are really helpful and I would like to learn more because I could find more interesting uses to it. After doing a little bit of research I found tools that could’ve helped me to do this easier. Jenkins is a tool that automates the whole development process, it’s open source and free. Kubernetes is another tool for automation but this one works for containers, from what I’ve heard is harder to learn and it’s not free. Both of them use Docker, a tool to containerize apps that I’ve hear a lot but haven’t had the chance to use it and learn why is useful. I think that all these tools are really powerful but sometimes complex, I need to learn how to use them well before implementing anything with them.
Now I know that doing all these tasks can be a lot of work and requires a lot of knowledge and experience to do some high quality automation. Being a DevOp also requires a lot of creativity, each environment is different in hardware, software and people, the DevOp has to have a different set of abilities to solve each problem, whether it is about automation scripting or people management.