How an AppDomain can help you isolate your method calls

In one of my latest pet-projects (which with a bit of luck is about to become a real project, knock on wood!) I have been in need of executing code in total isolation form the containing application. 

Without giving too many details, the system allows plugins to be uploaded ( as standard .NET assemblies ), stored and eventually executed. However, in order to avoid a rogue plugin to jeopardise the entire application with nasty stuff like infinite loops or random exceptions, I needed a way to isolate their execution.

Being this a new project, I started writing it using .NET Core but had to revert very quickly to the ol’ reliable Framework. Seems that as of now Assemblies can be loaded in memory but there’s no way to unload them.
Just for the sake of documentation, the class responsible of loading is AssemblyLoadContext . Unfortunately loaded assemblies will be kept in memory till the containing process gets closed (more details here).

So what’s the alternative? AppDomains ! As often happens, the MS documentation does a pretty good job explaining what they are and how should be used so I’ll move straight to the point.

I wrote a small example and pushed it to GitHub, you can find it here.

The core is the Isolator class: as you can see from the code it can be instantiated passing the Type of the class you want to isolate and the name of the method you want to run on it. Something like this:

using (var isolatedFoo = new Isolator(typeof(Foo), “Bar”))
     isolatedFoo.Run();

Under the hood the Isolator will:

  • create a new AppDomain
  • load the assembly containing typeof(Foo) 
  • create an instance of Foo
  • execute the “Bar” method on said instance

Also, being a very polite class, it will unload the AppDomain during the Dispose() .

More under the hood, the real magic happens inside the IsolatorRunner class. As you can see from the code it inherits from MarshalByRefObject, this will allow the instance to act as an intermediary between the two AppDomains, with a proxy that will be automatically generated to intercept the calls on the other side.

Being this a quick example, I didn’t focus on the quality of the interfaces. So for example being forced to use a string to denote the method to execute is a little…disappointing. 

I’ll work a little more in the next days (weeks?) and try to come up with a nicer API, maybe using Expressions and Funcs πŸ™‚

How to reset the entities state on a Entity Framework Db Context

I had two bad days. Those days wasted chasing a stupid bug. I had a test class with 4 test cases on my infrastructure layer. If executed one by one, they pass. If the whole suite was executed, only the first one was passing.

At the end I found out that it was due to the Entity Framework Core db Context tracking the state of the entities. More or less. 
In a nutshell, every time a call to SaveChanges() fails,  the subsequent call on the same instance of the db context will retry the operations. 

So let’s say your code is making an INSERT with bad data and fails. Maybe you catch that and then you do another write operation reusing the db context instance.

Well that will fail too. Miserably.

Maybe it’s more correct to say that the second call will look for changes on the entities and will try to commit them. Which is basically the standard and expected behaviour.

Since usually db context instances are created for every request this might not be a big issue.

However, in case you are writing tests using XUnit Fixtures, the instance is created once per test class and reused for all the tests in that class. So in this case it might affect test results.

A potential solution is to reset the state of the changed entities, something like this:

Another option is to avoid entirely reusing the db context and generating a new one from scratch.
In my code the db context was registered on the DI container and injected as dependency. I changed the consumer classes to use a Factory instead and that fixed the tests too πŸ™‚

How I used Travis CI to deploy Barfer on Azure

Ok this time I want to talk a little bit about how I used Travis CI to deploy Barfer on Azure. Last time I mentioned how much helped having a Continuous Delivery system up and running so I thought it would be worth expanding a little bit the concept.

Some of you may say: “Azure already has a continuous delivery facility, why using another service?”. Well there’s a bunch of reasons why:

  • when I started writing Barfer I had no idea I was going to deploy on Azure
  • in case I move to another hosting service, I’ll just have to change a little bit the deployment configuration
  • CD on Azure is clunky and to be honest I still don’t know if I like Kudu.

Let me spend a couple of words on the last point. Initially I wanted to deploy all the services on a Linux web-app. Keep in mind that I wrote Barfer using Visual Studio Code on a Macbook Pro. So I linked the GitHub repo to the Azure Web App and watched my commits being transformed into Releases.

Well, turns out that Kudu on Linux is not exactly friendly. Also, after the first couple of commits it was not able to delete some hidden files in the node_modules folder. Don’t ask me why.

I spent almost a week banging my head on that then at the end I did two things:

  1. moved to a Windows Web-App
  2. dropped Azure CD and moved to Travis CI

Consider also that I had to deploy 2 APIs, 1 web client and 1 background service (the RabbitMQ subscriber) and to be honest I have absolutely no desire of learning how to write a complex deployment script. I want tools to help me doing the job, not blocking my way.

The Travis CI interface is  very straightforward: once logged in, link the account to your GitHub one and pick the repository you want to deploy. Then all you have to do is create a .yml script with all the steps you want to perform and you’re done.

Mine is quite simple: since I am currently deploying all the services together (even though each one has its own independent destination), the configuration I wrote makes use of 5 Build Stages. The first one runs all the unit and integration tests then for every project there’s a custom script that

  1. downloads the node packages (or fetches them from the cache)
  2. builds the sources
  3. deletes the unnecessary files
  4. pushes all the changes to Azure

The whole process takes approx 10 minutes to run due to the fact that for every commit all the projects will be deployed, regardless where the actual changes are. I have to dig deeper into some features like tags or special git branches, but I will probably just split the repo, one per project. I just have to find the right way to manage the shared code. 

#hashtags just landed on #Barfer!

Yeah I know, I blogged yesterday. I probably have too much spare time these days (my wife is abroad for a while) and Barfer has become some kind of obsession.

You don’t know what Barfer is? Well go immediately check my last article. Don’t worry, I’ll wait here.

So the new shiny things are #hashtags! Yeah, exactly: now you can barf adding your favourite #hashes and you can even use them to filter #barfs!

The implementation is for now very simple, just a string array containing the tags, as you can see from the Barf interface defined here.

The command handle responsible for storing the Barf uses a regex to parse the text and extract all the tags (apart from checking for xss but that’s another story).

On the UI then before getting rendered every Barf goes through the same regex but this time the #tag is replaced with a link to the archive.

Quick&dirty.

Next thing in line would be adding some analytics to them but that would require a definitely bigger community πŸ˜€

I also went through some small refactoring and cleaning of the frontend code, although I will probably move to an SPA sooner or later. Thing is, I’m still not sure if using React, Angular or Vue so in the meantime I’m focusing on the backend.

There are so many features I would like to add that to be honest I prefer to not focus on the frontend for now. Maybe I’ll start looking for somebody who helps me on that.

One thing I’m quite happy for now but I plan to rework is CI/CD. Well for now I’m working alone on this so probably I can’t really talk about integration. But whatever.
As I wrote already, I’m using Travis CI and I’m very happy with the platform. Even though I’m still on the free tier, the features are awesome and flexibility is huge.  I’ll probably write a blog post on this in the next few days.

In the meanwhile, #happy #barfing! 

I’m becoming a Barfer!

More than a month! My last post on this blog was more than one month ago. I should write more often. No wait let me rephrase that: I should write on this blog more often.

Why? How I spent my last month? Barfing, here’s how!

Ok, let’s add more details. A while ago I decided it was a good idea starting to move a little bit away from the .NET world and explore what’s around. And  NodeJs arrived, of course with Typescript: I am 100% sure it’s basically impossible to write a semi-reliable system without some kind of type checking (along with 10000 other things). 

Then I said: “I don’t want to just read a manual, what can I write with it?”. For some crazy reason I opted for a Twitter clone. Yeah, I was really bored.

Early in the analysis phase RabbitMQ and MongoDb joined the party. I was not exactly familiar with RabbitMQ so I thought was a good opportunity to learn something new.

In order to speedup the development and to obtain certain features (eg. authentication ) I have used a bunch of third party services

The system uses a microservices architecture with a front-end that act as api-gateway. For each service I’ve taken the CQRS path along with Publish/Subscribe. An Azure WebJob is scheduled to run continuously and listen to the various events/messages on the queues. I’ll blog more about the architecture but that’s it more or less.

What am I expecting from this? Well it’s simple: nothing. I mean, I’m not expecting people to use it (even though would be very nice), I am just exploring new possibilities. Nothing more.

Oh yeah, before I forgot: https://barfer.azurewebsites.net/ . Enjoy!

© 2018 Davide Guida

Theme by Anders NorenUp ↑