What I am doing this week
- Continuing to complete my Appeals database App at work (This will be ongoing until completion)
- Software Engineering Update
Appeals Database
I spent most of my time starting to explore deployment methods. I have spent a great time with Docker and wanted to containerize my appeals application so I can deploy anywhere that supports containers. This can be in your own datacenter with an instance of Kubernetes or Docker Swarm for orchestration. Or it can be in the cloud provider of your choosing, AWS/GAE/ACS. First, I ran through a bunch of videos on Pluralsight getting me up to speed with using Docker in depth and Kubernetes. I learned that Docker for community has a edge edition that allows you to run an instance of Kubernetes instead of swarm on your laptop for development.
Once the edge version was installed I quickly enabled the feature inside of the Docker preferences to swap between the orchestration methods. I was off and running my Kubernetes instance but realized that I couldn't get access to the dashboard like I could with Minikube. Instead I found the dashboard GitHub page which had some Kubernetes YAML files to start up a node running the dashboard. It is as easy as downloading the file or pointing to the raw GitHub dump of the text file and telling Kubernetes to create. In minutes the dashboard is up and running and with no programming used at all. Now I can really see what is running in my environment.
Utilizing a DockerFile to build my project I tell Docker to take a Linux instance with Go already installed. This was provided by the official Go Docker image on Docker Hub. I then copy all my source code to the containers $GOPATH\src directory and as a development feature I instruct the container to build the app and run it when it starts. I think once I get the application production ready I will instead just copy over the templates and binary to the Docker images that way the container doesn't have to build the app every single time it is started.
Then I needed to make some configuration files for Kubernetes to understand. These are built a little different than a Dockerfile. The Dockerfile is more imperative which explains step-by-step how to build the container. In Kubernetes, we tell the engine what to do declaratively. Declarative means that you tell Kubernetes what you expect, and Kubernetes handles all the busy work. This busy work includes all the networking, scaling up an app across nodes, creating services, exposing ports, handling all the IP addressing and DNS between apps. In a YAML file that was less than 50 lines.
I also had to update my appeal importer program to import into the MSSQL database I launched into my Kubernetes instance. I made sure to expose a node port, so I could access it from my laptop. Then all I had to do is point my configuration file to localhost and the exposed MSSQL port and it could start inserting rows right into the database. I ran into a bunch of issues where the appeals web server nodes were starting up before the MSSQL database. I didn't want to host my Docker images in the cloud, so in the Kubernetes Config I needed to tell it when I told which image to load to never fetch it from the cloud which forced it to look at my computer only. Later I think I would just push my images to a private repository in the cloud like AWS or Google Cloud and point Kubernetes there.
Software
I checked with my group to see what was due for the week. Kind of confused with the deliveries, but I found out it is just a book report. Powered through Don't Make Me Think and finished it in 3 days. Enjoyed it heavily and will try to get his other books and chew through them as well.
Next Week
- Finish up some notices and forms
- I think my group wants to finish the sub-features this week, so guess we will be coding
Bibliography
Poulton, N. (2017, May 22). Getting started with kubernetes [Video course]. Retrieved from
https://www.pluralsight.com/courses/getting-started-kubernetes
Poulton, N. (2015, December 22). Docker and containers: The big picture [Video course]. Retrieved from
https://www.pluralsight.com/courses/docker-containers-big-picture
Poulton, N. (2016, August 1). Getting started with docker [Video course]. Retrieved from
https://www.pluralsight.com/courses/docker-getting-started
Poulton, N. (2018, January 4). Docker deep dive [Video course]. Retrieved from
https://www.pluralsight.com/courses/docker-deep-dive-update
Poulton, N. (2015, September 22). Integrating docker with devops automated workflows [Video course]. Retrieved from
https://www.pluralsight.com/courses/integrating-docker-with-devops-automated-workflows
Florek, S. (2018, February 13). Kubernetes dashboard [Code repository]. Retrieved from
https://github.com/kubernetes/dashboard/releases/tag/v1.8.3
