Microsoft cloud engineer - SharePoint, Office 365, Azure, DotNet, Angular, JavaScript.
Microsoft cloud engineer - SharePoint, Office 365, Azure, DotNet, Angular, JavaScript.

Development

Windows Azure – publishing from Visual Studio 2010 (step-by-step with screenshots)

Today I deployed my first Azure application and documented the step-by-step process with detailed screenshots.

First I downloaded the Azure SDK and prepared Visual Studio 2010.   Azure is a cool idea because of the support for highly scalable custom code.  One frequent objection to cloud hosting is the basic functionally.   Cloud providers tend to offer what’s easy and convenient … while avoiding custom code.    Hybrid models may be the resolution.  Architects who can connect multiple vendors seamlessly or blend on-premise with in-cloud servers will be in demand.

  • Download the Azure SDK
  • Locate the first Hands On Lab (HOL) “Introduction to Windows Azure”
  • Code and build in Visual Studio 2010
  • Verify local Azure Emulator is working (files / SQL database / Compute)
  • Deploy to Azure in the cloud  (NOTE – this is  lot of work the first time,  must get connected and trusted)

 

image

 

We begin with the "Guest Book” sample application which allows you to post messages on a common wall with a picture attached.  Simple enough.

image
image

 

Here you can see the tray Azure Emulator running the website locally on my laptop.   For development you want to run things locally, test, and deploy to the cloud when stable.

image
image

 

Right clicking on the tray icon allows you to view Emulator status, logs, and details.   You can restart and control instances here.

image

 

Out of curiosity I opened SQL Management Studio  and viewed the “SQLEXPRESS” instance (which the default for the Azure SDK examples).  Here you can see the databases and tables this sample created.  The schema appears very abstracted.   From Visual Studio 2010 all I saw were strongly typed Classes and Objects for data.   Somehow that is being flattened to a database table using the Azure binaries, pretty cool.

image

 

Here is the Visual Studio 2010 Solution Explorer showing the 3 projects and related files inside of the “Guest Book” sample application.

image

 

OK, let’s get to the fun part.   Open a web browser, navigate to https://manage.windowsazure.com/ and login to Windows Azure Platform Management.  There is some new vocabulary which be confusing at first.   First, create a new “Hosted Service.   This can take 5-10 minutes to provision so be patient.

image

 

Once the “Hosted Service” shows “Created” then we must add a “Storage Account” for the ability to hold data in the cloud.

image

 

Certificate trusting the local laptop’s Visual Studio 2010 to publish.   This is a LOT of work, but only a one time operation.  Let’s get started.   Right click on your Azure project within the solution explorer and click Publish.

image

 

The first time here we need to Add Credentials.

image

 

To pair the local Visual Studio with the Azure cloud, we need a certificate.   Since this is our first time, create a new one.  Name it whatever you like.  I chose to name mine after the development laptop “JEFF-PC”

image
image

 

Click “copy the full path” to populate the clipboard with the local file path.

image
image

 

We’re going to leave Visual Studio 2010 for a minute, but keep it open in the background since we’ll be back.  In the web browser go back to Windows Azure Platform Management.  Click “add certificate” and paste the local .CER file path from the previous setup.

image
image

 

Click on the “Subscription” line item to view the “Subscript ID” number.  Copy this to the clipboard.  You will need to give that to Visual Studio 2010 locally for publishing.

image

 

Back in Visual Studio 2010 we need to finish the Add dialog by giving it two values:   Subscription ID and Name.

image
image
image
image
image

Congratulations!!  

Hot

  You’ve successfully connected your local Visual Studio 2010 to the Azure Subscription.   You should now see three cascading drop down menus with the Azure cloud deployment destinations.

STSADM CMD with email output

Waiting and watching STSADM run on the CMD line is an exercise in patience.  Lately I’ve been incorporating two techniques to email myself updates for the freedom to walk away from the keyboard.

A quick command line EXE project allows me to send email by giving parameters (number 0-5).   If 5 is blank, no big deal just send a plain email with subject and body.   If 5 exists then attempt to open that text file, read content, and send as email body.

That’s great, but … how does this help me?  Great question!!

1) Send email at key script milestones

The freedom to walk away and receive alerts as progress continues allows you to be stay confident and know progress is moving forward.  Just add “SendMail.exe” inline with your existing CMD files to send an email at that step in the process.

 

2) Send email with script outcome

The detailed knowledge of what return value came from the CMD when it finished is helpful.   Imagine running a site move (STSADM backup/delete/restore) and heading out for dinner.   It’s very nice to get an email showing “Operation completed successfully” three times. 

Smile

   Just use “SendMail.exe” with parameter 5 being the text file full path and name.

image

image

 

Unlock ASMX for easy manual invoke (developers)

For SharePoint development I like to use the .ASMX web services for reading data.  Out of the box default security prohibits manual remote HTTP queries.  A quick edit to “C:Program FilesCommon FilesMicrosoft SharedWeb Server Extensions12ISAPIweb.config” for changing <protocols> from “remove” to “add” and then you’ll be able to manually Invoke.

The raw XML returned by a manual invoke is incredibly valuable for developers. 

Open-mouthed

Before (out of the box)

image

After (unlocked for development)

image

Always use CMD files

I rarely use absolutes but CMD files are amazing and you really should always use them.  For anything, no matter how small.

Why bother?  What’s in it for me?

  1. Easy to repeat
  2. Easy to test
  3. Can modify while running

The only cost is time.  It takes a little longer to build a CMD file and plan it out.  However, measure twice and cut once pays off with quality.  Let’s explore the benefits one at a time:

  1. Reusing a CMD file is very easy.  Search replace in Notepad++ can switch scope easily.  Also, input parameters like %1, %2, etc. can make the tool modular for using in a slightly different way quickly next time.
  2. Copy a CMD file to lower environments for practice is quick, easy, and helps mitigate risk.  Before you typo against production, try copying the file to a different server first.
  3. This is awesome and very little known!   While a CMD file is running (after you hit enter) you can still open Notepad.exe and modify the file.  Adding an extra line, commenting out with REM, or removing a line can all be done.  As long as CMD has not made it down to that line already you’re edits will be picked up as it steps down.   Great for real-time adjustments.

Thanks for reading! 

Open-mouthed
image

© Copyright 2016
@ SPJeff

Return to Top ▲Return to Top ▲