The Preface

In my time with Oracle, the company shifted it’s course from OCI-Classic to onboarding all new customers, and migrating current customers to OCI. This would provide some challenges for current customers as they would require an exercise in taking inventory of their current cloud resources. This would help in capacity planning for their future in Oracle’s cloud. Some customers kept up to date spreadsheets of all their current services in OCI-C, while others were seeking help with a more automated approach at mapping their current OCI-C resources.

The Situation

I was assigned to help a customer with their capacity planning for migration. They were in need of some information that Oracle couldn’t provide (as it was beyond a security boundary) and other information that could be provided programmatically using the OPC API. I was determined to give this customer the most information that could be provided in an automated fashion. They were specifically interested in their IaaS compute instances and collecting metadata information about them. I figured a simple CLI application could do the trick.

The Solution

I was able to cobble together a NodeJS application that could be configured with endpoint information, check the user’s credentials and pull all the necessary information down from the API. Once that was done, we could introspect it for the specific information we were looking for and write it to a simple CSV file.

If you’re on OCI-C and need to collect your own IaaS instance information, hopefully this tool can help you too!

Click to check the Repo out on GitHub

The Preface

I bought a new computer a while back. I do a lot of file moving/manipulation/cleanup on my home PC because I’m an avid Plex user. For these tasks I simply use .bat or .cmd files and the Windows task scheduler. It’s quite a simple solution and generally runs like clock work. Until it doesn’t and you can’t understand why.

First Attempt

After some internet searches I stumbled across the Tweaking.com windows repair program that offered fixes and had quite a few recommendations. With my fingers crossed I installed the app and tried it out. It did not fix my Windows task scheduler problem. But I did end up buying a license for it because it’s an amazing piece of software that does a lot of optimizations.

Coding Your Own

Time to think about rolling my own. What do I need this thing to do?

1. Always be running
2. Run .bat or .cmd files on an interval

So for #1 something always running, I immediately think Windows service. I’ve long used TopShelf for this sort of thing. It’s highly configurable and easy to use. For #2 I considered using Quartz.NET, but I thought it would be overkill. I just rolled my own interval and tracking code. I specifically needed to run .bat and .cmd files, for this I found a project called MedallionShell. It runs commands in threads within your application. You could use it for parallelism, but I decided not to introduce that complexity into my simple use case.

Learning

This project brought some fun learning aspects. I was able to implement builds using Cake builds for C# and also automate my builds using AppVeyor. With all of this in place it will build my code and put a release on GitHub for me. Pretty nifty! Even got some badges working on the project home page.

Conclusion

I learned a lot doing this simple little project and had a lot of fun doing it. I wish I knew why the Windows task scheduler doesn’t run my tasks. I guess as long as it runs all of the other important ones, I can be good with that.

Click to check the Repo out on GitHub

I have been doing a lot of research into the MEAN stack as of late in hopes of possibly using in on my next project. First I wanted to see if using TypeScript was viable, because I come from a .NET background and strong types help me find problems at compile time. Then I wanted to see if I could get some sort of debugging experience. I remember the good ol’ days of Classic ASP scripting and having to debug write everything so I could get a picture of what was happening.

Editor(s)

So many to choose from in this department. Personally I chose WebStorm because it afforded me that debugging experience I was looking for in the IDE. SublimeText and VS Code are excellent options as well, but like I said… Debugging.

Getting Lessons

My first round of searches brought me to an excellent video tutorial by Brad Traversy on MEAN Stack Front To Back. It’s a series in which you’ll use ES6 to create a MEAN stack application that does some simple authentication. It’s a great series, I highly suggest watching it and checking out the code here. It uses ES6 and like I said, I’m looking for TypeScript.

My next round of searching brought me to Brian Love’s blog where he has an article TypeScript2 + Express + Mongoose + Mocha + Chai. Whew! That’s a mouth full, and a lot of tech to learn. It’s a fantastic article about combining all those technologies to get a running MEAN stack application with testing. Check out his article here and certainly check out the code here.

Combo

My mission was to combine the two projects to make one MEAN application with a TypeScript and test driven backend and an Angular2 front end. I was able to take the lessons learned about TypeScript, Grunt, Mocha and Chai and apply it to the meanauthapp created in Brad’s videos. It makes for very easy and type-safe development with NodeJS. Big thanks to both Brian and Brad for their work!

Click to check the Repo out on GitHub

Isn’t the whole point of durable, reliable messaging that we can re-queue a message for processing?
– Me

The Background

I recently got involved in a project using Azure Service Bus. The premise was simple, run service bus queue messages through a WebJob durably and reliably. For this project I ended up using the Topics feature for a publish/subscribe model. I got the project up and running fairly quickly thanks to the WebJobs SDK. By default the message will be attempted 10 times and then automatically sent to the DeadLetter queue. So to test this, I throw an exception during message processing and sure enough it goes as expected. I now have a message in the DeadLetter queue. In the Azure Management Portal I can see the message exists where it’s supposed to. As of the writing of this article the Azure Management Portal only supports changing the properties of Topics and not managing the messages in them.

Supplemental Software

What I need is a piece of software that will let me manage the messages in the queue. I could write it myself, but this seems like a task there would be a tool for.

My first round of searches dug up Service Bus Explorer. It’s a free Service Bus management tool. It looks like a great piece of well thought out management software, but I was unable to get the DeadLetter Repair and Resubmit Message to work properly for me.

Update: After a bit of digging about BrokeredMessage serialization I was able to find this article on StackOverflow explaining how to dictate how Azure Service Bus handles my message content. After implementing that Service Bus Explorer is working like a charm.

My next round of digging brought me to Azure Management Studio. This is a paid piece of management software that helps manage more than just your Service Bus. This tool was able to properly copy and re-submit the queue message back to it’s Topic of origin. Success!

Side Note

If Domain Driven Design and Message Queues are highly important to your system, you may want to check out NServiceBus. I just needed some bolt-on queues for middleware which is why I decided to just tackle queueing on my own.

Happy coding!

Every day in the ever changing world of technology and development, I try to do some reading to keep up with the world around me. I have worked in places with small teams and big teams, but my current situation puts me as the only person with development skills in my current company. It’s really important in this situation, not to lose touch with the world of development around you. This feed has kept me up to date with new tech and also taught me even more about tech I all ready thought I knew.

I’ve been curating this list of blog feeds since I became a developer a long time ago, so many of them may be out-dated. Still, there are always great articles popping up for me to read and dabble in the tech if I have the time. Personally I use Feedly for my RSS reader, but the file should easily import into your favorite RSS reader.

Developer Feed File - You may need to Right Click -> Save As

Anyhow, hope you enjoy!

I was talking to a friend the other day about what you need to know to be an effective developer. We touched on topics like source control, people skills, presentation skills and a ton of other skills that are inherent things that are need to know. Other things could include frameworks, good blog sites, or how to get what you’re asking for from your boss. After a while we got into the code aspect of being a developer. Where do you start? What is a waste of time to learn? When should I know what? Well, I certainly don’t know the best way but I’ll share the order that worked for me.

Let’s just start basic. The first thing you should know about developing for whatever it is you want to write code for is the syntax of the language you’ll be using. Start simple, think “Hello World”.

If you’re learning an Object Oriented language (like C#), your next stop should be The four pillars of Object Oriented Design

• Abstraction
• Polymorphism
• Inheritance
• Encapsulation

Once you have an understanding of the 4 pillars in my opinion your next stop should be learning the SOLID principles

• Single responsibility principle
• Open/closed principle
• Liskov substitution principle
• Interface substitution principle

Once you’ve gotten to principles you can start to make more maintainable code in larger projects.

Finally, once you’ve gotten those down it’d be best to look into the Gang of Four Design Patterns. These are proven patterns for enterprise scale production systems. If you have a good understanding of the problem you’re trying to solve, there’s a good chance a combination of these patterns will help you solve it.

All of this being said, there are problems you come across time and again as a developer. One of the biggest ones I face often is that of cross cutting concerns. I have this thing that does one thing, but now needs to do another. For example, you have data going into a database and you’d like to audit who put the record in with some sort of logging. In this case there are many solutions, and you could even write your own using all those spiffy patterns. Or in my case I found a great library called MediatR that implements a nice architecture solving that problem for me. Sometimes with the right amount of digging you’ll find a good library that solves the recurring problems.

Update

It is almost imperative now to know about cloud architecture for hightly available and scalable applications. I had the pleasure of reconnecting with a former colleague of mine who so happens to be a cloud architect at Microsoft. He pointed me toward his GitHub for some guidance on Microsoft Azure cloud architecture. You can see the repository on his GitHub here.

Happy coding!