it is quite common to see advice about API url design, where URLs are on the "/api/v1/resource" format, and then when the API changes, we could change to /api/v2 etc.
Now, this must be implemented somehow. There are a number of options:
deploy the project at the root of the web server, and let the routing rules take care of handling the /api/v1 part
deploy the project in a /api/ subfolder (virtual directory), routing rules take care of the /v1, /v2 etc. parts, but are not aware of the /api/ portion of the URLs.
deploy the project in a /api/v1 subfolder (virtual directory). A new version of the API is a new project altogether, deployed separately. The project deals strictly with the resources as a root concept, but is generally not aware of the /api/vX part.
So, which method would you pick, and why?
Regards,
Daníel
I have deployed my project at the root of the sub-folder and let my routing rules handle the versioning routes. I prefer to have my implementation with as little dependence on my hosting environment as possible in case I want to deploy to an environment that does not support my method of implementation.
SDammann.WebApi.Versioning is the solution I used to achieve this in one of my applications.
However ASP.NET and Web Tools for Visual Studio 2013 Release will make this even easier.
Related
I have stateless .net core 2 Fabric Service Web Application creating using one of the templates that comes with Service Fabric SDK. It is a real pain to develop since I have to do a full deploy before I can see any changes to code/html/script. In my case that operation takes more than 5 minutes.
I have looked at this article that states how it can be done by running the web app from the commandline.
That article is based on Net Core RC2. Does anyone has an updated example on how to do this?
https://dzone.com/articles/aspnet-core-with-kestrel-and-service-fabric
Together with Azure Developer Support i found a solution to speed up the development process
I Fabric Explorer you need to find the node where you Web Application is running. I my case that is _Node_0
By SF SDK design, local SF published file is under C:\SfDevCluster\Data_App\ this folder. In my environment, the website file path is C:\SfDevCluster\Data_App_Node_0\Application1Type_App1\Web1Pkg.Code.1.0.0\wwwroot\lib\bootstrap\dist
So you can also find your HTML, CSS, JS and other static resources under below path:
C:\SfDevCluster\Data_App[node_id][application_type_and_instance_name][service_type_and_version]\
You can just modify the files in this folder, then the change will immediately apply to your local test web browser. Please notice if your service is hosted by micro-service running in several nodes, you may need to modify all nodes files because load balancer may access any folder files randomly.
I'm developing a Service Fabric-based trading platform that will host hundreds of different long-running trading algorithms, all of which conform to a common interface and share a good deal of common code but can be vastly different in their internal specifics. I could model each of the different algos as an application type (which I'd dynamically load) but given the large number of different algos I have to wonder if in makes more sense to create a single Plugin Runner application type then implement the algos as plugins.
In a related question, I understand how to implement a plugin architecture, in general, but I'm not quite sure where one would place the actual plugins in order to be discoverable by an instance running on Service Fabric.
Anyway, thanks for your help....
Both approaches can work I think. Using lots of Application Types adds the (significant) overhead of running lots of processes, but allows you to use and upgrade multiple versions of the same algorithm running simultaneously.
Using the plugin approach requires you to deal with versioning yourself.
Using the Application approach probably requires some kind of request router, while the
plugin service could make it's own decisions (if it's stateless).
You can create a Stateful service that acts as the plugin repository, or mount a file share, or use a database, no restrictions from the platform here. You can use naming conventions to locate the proper plugin.
The following approach could work if an application upgrade is acceptable to you when changing the set of plugins needed for a given application instance.
Recall that Service Fabric apps must be packaged before deployment or upgrade. Using either msbuild tasks or Powershell, you could copy your plugin dlls to the plugin runner service's code package as a post-packaging step prior to the app upgrade. Then your plugin dlls would be available to the service at startup using Assembly.Load and the code package's path, available in your service implementation's Context.CodePackageActivationContext.GetCodePackageObject("Your-Code-Package-Name").Path property. The code package's name is defined in ServiceManifest.xml, and is named Code by default.
I am trying to change site url from http://localhost/yiiwebsite/backend/web/index.php url to http://localhost/yiiwebsite/admin and http://localhost/yiiwebsite/frontend/web/index.php url to http://localhost/yiiwebsite/.
Can anyone help me to do this.
It's described in official docs here.
Here is some basic info:
The application installed according to the above instructions should
work out of box with either an Apache HTTP server or an Nginx HTTP
server, on Windows, Mac OS X, or Linux running PHP 5.4 or higher. Yii
2.0 is also compatible with facebook's HHVM. However, there are some edge cases where HHVM behaves different than native PHP, so you have
to take some extra care when using HHVM.
On a production server, you may want to configure your Web server so
that the application can be accessed via the URL
http://www.example.com/index.php instead of
http://www.example.com/basic/web/index.php. Such configuration
requires pointing the document root of your Web server to the
basic/web folder. You may also want to hide index.php from the URL, as
described in the Routing and URL Creation section. In this subsection,
you'll learn how to configure your Apache or Nginx server to achieve
these goals.
By setting basic/web as the document root, you also prevent end users
from accessing your private application code and sensitive data files
that are stored in the sibling directories of basic/web. Denying
access to those other folders is a security improvement.
If your application will run in a shared hosting environment where you
do not have permission to modify its Web server configuration, you may
still adjust the structure of your application for better security.
Further configuration depends on chosen web server (Nginx / Apache), which is not even mentioned in the questoin. But both options are covered in official docs by the given link.
For shared hosting environment there is special section too.
And by the way this was asked before many times here on SO, just do a better research.
Where I work, we release bug fixes in to the system every night when we know our clients are not using the system.
Trying to take a step towards better service I'd like to deploy to IIS while the application is running.
A solution that comes to mind is to setup two different IIS applications and switch them over after deploy using a script. But I'm not going to try this out as I don't want any complications during our busy hours.
Does anyone have experience in this area of deployment?
Thanks
Regardless of whether you're using PHP, ASP, ASP.NET etc there is no native support for transactional deployment on IIS.
The simplest approach would be to have two physical folders and (optionally two web sites - one production, one test) on your web server, for example:
c:\websites\myapp\dep1
c:\websites\myapp\dep2
Initially your site would have its physical path pointing to c:\websites\myapp\dep1.
When you deploy your latest build you'd deploy into c:\websites\myapp\dep2. Once you're done just switch the physical path of the production site over to this folder. This means you still have the original site and can fall back to it if the new code fails for whatever reason.
The next time you do a deployment you'd deploy into c:\websites\myapp\dep1 and once you're done switch the production site to that folder.
You could optionally have a test site that points to the folder you're deploying to so you can make sure the site works before switching your production site over.
This could all be scripted.
Here's some related reading that may be of interest:
Publishing/uploading new DLL to IIS: website goes down whilst uploading
Is smooth deployment possible with componentized ASP.NET MVC apps?
Rob Conery also had an excellent blog post about the lack of a decent deployment story for ASP.NET application. You should take a trawl through the comments some of which are quite insightful:
ASP.NET Deployment Needs To Be Fixed
Getting Constructive On ASP.NET Deployment
We have recently added some functionality to our web site that requires adding a service reference to an api in an external domain. Adding the reference to a VS2008 project createed a sub-folder in the "Service References" folder and added 18 files to that folder that appear to represent the classes in the api. The api provider also had me add custom binding and client references to system.serviceModel in my web.config file.
Do I have to deploy that entire folder with its 18 files to my production site to use the web service? Does some of it get compiled into my project dll? I can find all kinds of references to deploying a service, but not what is necessary to deploy a service consumption module.
Sorry for short, but I don't have the answer for you. But you can get this answer yourself. This is the only way Iwould get the answer. Deploy it without them, see if it works. If it doesn't deploy with. If it works, then you know.
The good news is that I have to deploy absolutely no files from the project have to be deployed to the production site. Everything the web service consumer application needs gets compiled into the application assembly.