Docker and the Path to a Better Staging Environment - a 1 1/2 armed #LiveBlog from #STPCON Spring 2018

Wow, time flies. I'm in the last session of the conference, and while I've had to split my time and focus on work stuff today, I've still been able to participate more than I thought I would. I have a couple of additional entries I'll make later (so they don't really count as "Live Blogs" but I still want to get them down).

My current work environment uses Docker extensively for a number of things. Perhaps most specifically we use Docker for massive parallelization of our automated tests and for our CI?CD system. Docker has been around since 2013 and it was at that time that we implemented our parallelization strategy. A fair amount of Docker was still experimental at that time, so we implemented some creative methods to get Docker top do what we wanted it to. We accomplished our goal, but Docker has matured a lot in five years. THus I wanted to get a better feel for the present state of Docker. Gil Tayar is providing that opportunity :).

To be frank, there is no way I am going to be able to do justice to what Gil is covering here as he is typing fast and furious. I am, however, getting a better appreciation for how Docker actually does what it does. I have wondered what the port mapping actually did and why it was so important. Seeing it live is pretty cool. It's transactionally very fast.

Little things I am learning that are quick hits:

- Kubernetes is a large scale docker coverage tool. It means "captain of the ship" in Greek.
- #K8s is short for kubernetes. Now I finally understand what that means :).
- There's a lot of stuff we can do with Docker that my current implementation is not doing. I need to do some digging in the dirt for this.

And with that, I'm going to say "adieu" to old friends, new friends, and to those who I haven't met yet... there's still time :). Please say "hello" before you leave. Just look for the guy with the elbow and wrist brace. I'm pretty sure I'm the only STP-CON attendee that has that distinction :).

Scaling QA With Docker, Jenkins and Mesos - Live from #AgileTD

Adé Mochtar & Maarten van den Ende put together a workshop based around the idea that we don't have to spend so much time manually setting up Selenium Grids. What's more, each of us were given a USB stick with the files and exercises to work with. The goal of the sessions is to be able to set up scalable and solid Selenium Grids, as well as writing maintainable and scalable Selenium / WebDriver tests.

At the moment, we have been setting up our environments so that we can load the environment and get it up and running. If you haven't seen much of me the past hour, well, that's why ;).

There are a bunch of components that are part of this package, including vagrant, selenium, selenium grid, jenkins, docker, a few virtualbox machines and scheduler called Mesos. For this process of scaling Selenium Grid, even in this small example, Mesos gives the flexibility that will allow as many CPUs and RAM that you can throw at the system. In our example, the CPU's are limited to the number of virtual machines we have created and are spinning up.

One additional cool aspect of Mesos is that it's a generic scheduler. It can be used to schedule anything, so any test and configuration can be run using it.

Very interesting model, I've just barely scratched the surface, but I am curious to play with it some more, seeing as I have it set up and all that :).