Hands-On Lab
Embracing Continuous Delivery with Release
Management for Visual Studio 2013
CONTENTS
OVERVIEW ... 3
EXERCISE 1: RELEASE MANAGEMENT OVERVIEW ... 4
EXERCISE 2: CONFIGURING RELEASE MANAGEMENT ... 10
EXERCISE 3: DEFINING A RELEASE PATH AND TEMPLATE ... 19
Overview
In this lab, you will learn about Release Management for Visual Studio 2013 and its suite of release and deployment tools that automate the deployment of applications across the desktop, server, and the cloud. Release Management for Visual Studio 2013 helps development and operations teams integrate with Team Foundation Server 2013 to configure and automate complex deployments of their automated builds to target environments more easily. Development teams can also model their release processes and track approvals, sign-offs, and visualize their release status.
Note: If you want to skip the overview and configuration exercises and go straight to a live
demonstration in action, you can start with Exercise 4. If you are new to Release Management, however, it is recommended that you at least read the first few exercises for some background information.
Prerequisites
In order to complete this lab you will need the Visual Studio 2013 virtual machine provided by Microsoft. For more information on acquiring and using this virtual machine, please see this blog post.
About the Fabrikam Fiber Scenario
This set of hands-on-labs uses a fictional company, Fabrikam Fiber, as a backdrop to the scenarios you are learning about. Fabrikam Fiber provides cable television and related services to the United States. They are growing rapidly and have embraced Windows Azure to scale their customer-facing web site directly to end-users to allow them to self-service tickets and track technicians. They also use an on-premises ASP.NET MVC application for their customer service representatives to administer customer orders.
In this set of hands-on labs, you will take part in a number of scenarios that involve the development and testing team at Fabrikam Fiber. The team, which consists of 8-10 people, has decided to use Visual Studio application lifecycle management tools to manage their source code, run their builds, test their web sites, and plan and track the project.
Exercises
This hands-on lab includes the following exercises: 1. Release Management Overview
3. Defining a Release Path and Template 4. Release Automation Example
Estimated time to complete this lab: 60 minutes.
Exercise 1: Release Management
Overview
In this exercise, you will learn about the Release Management product and how it interfaces with Team Foundation Server to provide an automated deployment solution. You will also see what an example release workflow looks like from the Release Management web client.
1. Log in as Brian Keller (VSALM\Brian). All user passwords are P2ssw0rd.
2. The goal of Release Management’s architecture is to provide a mechanism where application components can be deployed automatically to various target servers in different environments. The components may require different configurations on the various servers but we still want to deploy the same package to all of them. Another key goal is to keep all the configuration
information centralized and manage the deployments as part of a business driven release workflow that involves multiple roles in the organization.
3. In order to accomplish these goals, Release Management is based on four main components.
Figure 1
Release Management Server. The server component is the heart of the solution and consists of
the database, the workflow controller, and a dispatcher that synchronizes activities with the target servers.
Release Management Client. The client comes in two flavors, a Windows Presentation
Foundation (WPF) based client that exposes all the functionalities of the application and a web-client designed for testers, user acceptance approvers and managers.
Release Management Deployer. The deployer is a service that is installed on the target servers
where the application components need to be deployed. It is important to note that the Release Management Server does not require access to the target servers as all operations are based on a pull mechanism from the Deployer.
Deployer Tools. The Release Management solution provides various installation tools that assist
different deployment scenarios such as uninstalling/installing components, deploying reports to Microsoft SQL Reporting Services, and moving files to specific locations.
Note: This self-contained virtual machine has all Release Management components installed
on it for demonstration purposes, including the deployment service.
4. It is important to note that the Release Manage product, after having been acquired by Microsoft, will continue to be transformed and integrated into existing tools. The release management authoring components will be included in Visual Studio Test Professional, Visual Studio Premium, and Visual Studio Ultimate. Everything needed to participate in a release process will be included in the Team Foundation Server CAL. Server components will be
integrated into Team Foundation Server 2013. The deployers (which are required for each node you deploy on) will continue to be licensed separately.
Figure 2
Release Management web application
6. Select the Previously Approved link and note that there was a release triggered by a build of the Fabrikam Call Center.
Figure 3
Viewing previous approvals
7. Select the link displaying the number of deployed components (just underneath the name Brian Keller).
Location of components link
8. This shows us the components that were deployed during each stage of the release path. In this particular case, it was just a specific build of a web site.
Figure 5
Viewing components deployed
9. Press the Escape key to close the dialog window.
10. Note that we can see at a glance that the deployment is currently in the production stage. Select the stages link to view the historical workflow and approvals that occurred to get to that point.
Figure 6
Location of stages link
Figure 7
“Prod” stage historical workflow
12. Select the Previous Stage link.
Location of Previous Stage link
13. This stage was setup to be for QA. Note that this stage has more workflow automation in place – it automatically accepts, installs, and validates the application and then waits for a QA team member to approve it.
Figure 9
“QA” stage historical workflow
Note: The setup used for the stages seen in this lab are for demonstration purposes only. In
normal scenarios, the QA stage would not automate the acceptance step. It would usually be setup for an owner of that stage to decide when to deploy a new version.
Figure 10
“Dev” stage historical workflow
15. Press the Escape key to close the dialog window.
16. The paths are composed on the various servers grouped in environments on which the testing for the stage is performed. Once an application needs to be deployed to a new environment, the server will queue deployment requests to all the required target servers for each
component of the application. This allows an atomic deployment of all the components. 17. The Release Management Deployer running on each target server monitors the Release Management server continually (at a configurable interval) and will pick the installation requests for the one or more components it needs to install locally.
18. The Deployer will then find and download the release package, provided by the Release Management Server that calculates the location using the TFS API - if built by TFS - or using a predefined UNC path - if not.
19. Finally, the Deployer downloads any additional executable (batch file, PowerShell script, .exe) to be ran as part of the installation. These are additional deployment activities beyond the
installation itself; creating test data or triggering automated tests are common scenarios here.
In this exercise, you will learn about the main configuration points that are needed to connect Release Management to Team Foundation Server, various settings for Release Management including those that apply to the Deployer services, the configuration of groups and users, and finally the configuration of servers and environments.
1. Launch the Release Management client from the taskbar.
Figure 11
Release Management client splash screen
2. By default, Release Management will load the Traffic Overview tab which shows deployments moving through all release paths and stages. This shows us that the Fabrikam Call Center application has already had a deployment go through each stage of deployment without any failures.
Figure 12
3. Let’s take a quick look at some of the main configuration tasks that need to be addressed when installing and configuring Release Management. Select the Administration tab followed by the
Manage TFS link.
Figure 13
TFS connection configuration
4. Double-click on the TFS connection that has already been setup.
Figure 14
Loading configuration for TFS connection
5. This connection was setup in the virtual machine ahead of time, but it is important to note that the user account used by Release Management needs to have the “Make requests on behalf of
others” permission within TFS.
Figure 15
Configuring account to connect to TFS
6. Other important settings can be configured in Administration | System Settings. The System
System Settings tab (default) shows various timeouts, version information, SMTP configuration,
Figure 16
System settings
7. Select the Deployer Settings tab to view the configuration options for all deployer services. For example, you can set how often you want the deployer services to poll the Release
Management server for packages to deploy.
Figure 17
Deployer settings
8. Let’s take a quick look at the setup of users and groups for Release Management. Navigate to
Figure 18
User configuration
9. Double-click on the Brian Keller user.
Figure 19
Viewing user configuration details
10. The Brian Keller user is associated with the windows account VSALM\Brian, is designated a
Release Manager, and is an active member of a few different teams.
Figure 20
Viewing user configuration details
Figure 21
Group configuration
12. Note that you can create new groups from scratch or you can import them from Active
Directory or Team Foundation Server. Groups that are imported from AD or TFS in this way are linked by default, and will therefore remain synchronized.
Figure 22
Importing groups from AD and TFS
Note: Synchronization is manual (using the Refresh button) unless the setting AD/TFS-Based Group Refresh Interval is setup to something other than 0 minutes (which is the default).
13. Double-click on the QA Team.
Figure 23
14. The Members tab shows the individual users that are part of the QA Team. You can add more users here if desired (since the group is not linked to AD or TFS).
Figure 24
Viewing group members
15. Select the Security tab.
Figure 25
Location of Security tab
16. The Security tab allows you to specify what Release Management permissions that the group has. For the purposes of this virtual machine, the team members have full control.
Figure 26
Group security settings
Figure 27
Configuring stage types
18. Servers to be used for deployment must have the Release Management Deployer service installed and configured to connect to the Release Management Server over HTTP or HTTPS. In addition, these servers must be explicitly added to Release Management Server. Navigate to
Configure Paths | Servers and note that the Deployer service has already been setup and
configured for this virtual machine.
Figure 28
Configuring servers
Note: Although we won’t do so here, the recommended way to add additional deployment
servers is to select the drop-down arrow next to the New button and then select Scan for
New.
Figure 29
Configuring servers
20. There are many options shown here for the selected deployment server, but the general takeaway is that you want to configure servers that you add to be uniquely identifiable so that Release Management Sever can target them. It is possible to used cloned servers, configure the address type to be a gateway, and to have the server use HTTP(S) to grab the deployment bits from the drop location (if it a UNC path is not an option).
Figure 30
Configuring servers
21. Navigate to Configure Paths | Environments.
Figure 31
22. Servers are grouped into environments so that servers are decoupled from release path definitions and so that various stages of the release path can be restricted to certain environments. Double-click on the first environment named “Int-Dev”.
Figure 32
Viewing environment configuration
23. As the description states, this environment is meant to define the group of servers used for a development environment. If you wanted to restrict the use of this environment to specific stages, you could do so in the Stage Type Security tab.
Figure 33
Viewing environment configuration
Exercise 3: Defining a Release Path and
Template
In this exercise, you will learn how a release path and release template are created and configured. You will also see how to use the actions and tools provided to deploy an application to the correct
environment.
1. Now let’s look at a release path definition. Navigate to Configure Paths | Release Paths and
Figure 34
Viewing release path
2. This release path defines a three-stage path through Dev -> QA -> Prod using the selected environments. Most steps for the first two stages are automated, so the assigned user or group does not intervene. Both the Dev and QA stages required approval before the next stage could begin.
Figure 35
Viewing release path
3. Now let’s look at how the Fabrikam Fiber team defined the process used to deploy their web application. Navigate to Configure Apps | Release Templates and then double-click on the
Figure 36
Configuring release template
4. The release template designer has a toolbox with control flow building blocks, servers, custom components, and a bunch of other actions and tools to help with deployment. Select the
Properties link.
Figure 37
Viewing release template properties
Figure 38
Viewing release template properties
6. Press the Escape key to close the Properties window.
7. The first stage should be selected with its deployment sequence shown below. For the purposes of this lab, the deployment sequence is relatively simple to help illustrate the concept. Select the Collapse All button so that we can dig into the example deployment sequence starting at a high level of detail.
Figure 39
Location of Collapse All button
This toolbox node shows all servers available to the environment configured for the currently selected stage.
Figure 40
Deployment sequence showing server
9. Expand the VSALM node. In summary of the details to follow, the general deployment sequence involves removing the existing web site from IIS, backing up the current bits, xcopy deploying the new bits from the build, re-creating the web site in IIS, and finally rolling back if there are failures.
Deployment sequence
10. Expand the Remove Web Site node. This action was dragged and dropped onto the deployment sequence from the IIS toolbox node. It is configured to remove the “FabrikamDev” site from IIS.
Figure 42
Remove Web Site node
11. Expand the Copy File or Folder node. This action is from the Windows OS toolbox node and is configured to back up the current web site location to a backup folder.
Figure 43
Copy File or Folder node
12. Expand the Call Center Site node. Note the puzzle piece icon on the top-left, which indicates that it is an instance of a custom component.
Figure 44
Call Center Site node
13. Let’s look at this component by navigating to Configure Apps | Components and then
Figure 45
Viewing custom component configuration
14. In the Source tab, note that the “Builds with application” option is selected. This means that the component will inherit the team project and build definition from the release template. The path to the package to deploy is currently set to [Build Drop
Location]\_PublishedWebsites\FabrikamFiber.Web.
Figure 46
Viewing custom component configuration
Figure 47
Location of Deployment tab
16. This component uses the XCopy Deployer tool, which is backed by a script named irxcopy.cmd. The Arguments property is setup to copy all deployment source files to the Installation Path parameter (which is exposed on the release template design surface).
Figure 48
Viewing custom component configuration
17. Return to the Fabrikam Call Center release template and expand the Create Web Site node. This creates the specified site in IIS.
Figure 49
18. Finally, look at the Rollback sequence, which restores the backup and re-creates the original web site in IIS (if needed).
Figure 50
Rollback sequence
Figure 51
Viewing QA stage
Note: In the likely event that your stages have a similar structure, you can copy and paste
elements from one stage to another. You can even copy the entire deployment sequence from one stage to another. In the event that some servers are not available on the target stage, you will be prompted to replace those servers with available ones. Copying an entire deployment sequence can be accomplished by right-clicking on a stage node in the Release Template and then selecting the option to copy.
Figure 52
Location of Tools tab
21. The current set of configured tools provides the ability to execute command line statements, manipulate files and processes, deploy databases and websites, install applications, manage Azure virtual machines, and even run automated tests defined in Microsoft Test Manager. Some of the tools are backed by scripts, while others are backed by executables. You can easily add in your own tools if needed.
22. Select the Actions tab.
Figure 53
Location of Actions tab
Figure 54
Viewing available actions
Exercise 4: Release Automation Example
In this exercise, you will configure a Team Foundation Server build for continuous integration, ensure that it automatically triggers a release, and then execute/follow that release all the way through the development, QA, and production stages.
1. Log in as Brian Keller (VSALM\Brian). All user passwords are P2ssw0rd.
Figure 55
Team Explorer – Home view
3. Select the Builds tile.
Figure 56
Location of Builds tile
4. Right-click on the “Nightly Fabrikam (Dev)” build definition and select the Edit Build Definition option.
Figure 57
Editing build definition
Figure 58
Location of Trigger tab
6. The name of the build implies that the Fabrikam Fiber application is built each night, even though it is currently set to be manually triggered. Let’s say that the team has decided to go with the Continuous Integration option, to build on each check-in, so select that option.
Figure 59
Selecting Continuous Integration option
Figure 60
Location of Process tab
8. As we pointed out in the previous exercise, a custom build process template needs to be used in order for builds to be handed off to Release Management. Note that the
ReleaseDefaultTemplate.11.1.xaml template is selected.
Figure 61
Custom build process template
Note: The Release Management build process template can be found in the \Microsoft Visual Studio 12.0\Release Management\bin folder.
9. As a quick aside, the custom build process template also contains the logic to tokenize your configuration files. This logic assumes that in your solution, you have two versions of your configuration files. One version is your normal configuration file used during local development, and the other is a corresponding file that has the same content, except that instead of having local values for your variables, tokens have been put there. The build activity will swap those two files before doing the build, so that you end up with the tokenized version of the
configuration files in the drop location.
web.config.token. Your web.config file will stay the way it is now (and that will be used when you run the app locally). The web.config.token file will contain tokens instead of values.
Figure 62
Example of token file
11. Back to our build configuration, scroll down to the Release Management section of the build process parameters and note that the Release Build parameter is set to ‘True’. Both the build definition and Release Management need to be configured in order to allow a build to trigger a release.
Figure 63
Location of Release Build option
Note: In the case of a nightly build, it may make sense to set the Release Target Stage to be
something other than production, perhaps a development or QA stage, but for demonstration purposes, we will take the release all the way to production.
12. Press Ctrl + S to save the build definition. Everything should now be in place for a continuous integration scenario where a source check in will trigger both a build and a release.
Figure 64
Loading FabrikamFiber solution
14. Launch Internet Explorer from the taskbar and select the FF DEV button from the favorites bar to load the Fabrikam Fiber site currently deployed to the development environment. You’ll have to play along with the scenario here, as the QA and Production versions of the site are also on the same machine, albeit on different ports.
Figure 65
Location of Fabrikam Fiber link (development)
Figure 66
Location of _Layout.cshtml
16. In _Layout.cshtml, locate the h2 tag that contains the “Support” text and change it to be “Support v2.0”.
Figure 67
17. In Team Explorer – Pending Changes, select the Check In button. Select Yes if prompted to save changes and check in.
Figure 68
Checking in change
18. If you quickly open Team Explorer – Builds, you should see that the check in triggered a build.
Double-click on the build.
Figure 69
Opening build in progress
Figure 70
Location of View Log link
20. If you scroll down through the activity log, you should see the steps that have to do with Release Management.
Figure 71
Activity log showing steps involving Release Management
Figure 72
Loading Release Management traffic overview
22. Note that the Fabrikam Call Center release path now shows that another deployment is in process in the development stage.
Figure 73
Traffic overview showing release in process
Figure 74
Loading detailed traffic view for “Dev” stage
24. Assuming that you have waited long enough for the deployment to complete, you should see that the most recent release (top) is currently in the “Dev” stage waiting for approval.
Figure 75
Release history showing release in progress
Figure 76
“Dev” stage workflow configuration
26. As you can see, the acceptance, deployment, and validation steps are all automated while we are awaiting explicit approval before moving on to the “QA” stage. Specifically, the release is waiting for Brian Keller to provide approval. Although not configured in this virtual machine, Brian would receive an email alerting him about the pending approval.
27. Select the My Approval Requests link to view pending approvals.
Figure 77
Location of My Approval Requests
Figure 78
Loading pending approval details
29. Select the View Sequence link.
Figure 79
Location of View Sequence link
30. We can look at the deployment sequence and see all of the specific parameters that were (or will be) used for each stage. Note that these ultimately become historical (and read-only) for each specific release.
Deployment sequence showing parameters
31. Select the View Log link.
Figure 81
Location of View Log link
32. The log shows the details and status for each step of the release process. This shows that the deployment was automatically accepted, deployed, and validated for the “Dev” stage and is now waiting for approval. The Deploy step also has additional details – select the ellipses button in the details column.
Figure 82
Location of ellipses button
Note: You can view future steps as well by selecting the “Include Future Steps” option at the
bottom of the log.
33. The deployment log shows details for each action performed. Select the View Log link for the
Remove Web Site action. Use Notepad if prompted.
Figure 83
Note: If a specific action fails, the output from the underlying tool used should provide
debugging information to help determine if there is a problem with the target environment or the deployment sequence.
34. This log shows that the FabrikamDev web site was deleted successfully.
Figure 84
Viewing successful action log
35. Close the Notepad window and the Deployment Log window.
36. Before we approve the release, let’s look at the deployed site in Internet Explorer. Select the “FF DEV” link from the favorites bar.
Figure 85
Location of “FF DEV” link
Figure 86
Deployment to the development environment is a success
38. In Release Management, return to My Approval Requests, select the release, and then select the Approve button.
Figure 87
Approving the release to the development stage
39. In the Approval Confirmation window, enter a comment such as “Dev deployment looks good” and then select the OK button.
Figure 88
Adding an approval confirmation comment
Figure 89
Waiting for QA approval
41. Once the deployment is complete and automatically validated, someone from the QA Team will need to approve the release. Brian is a member of the team, so go ahead and load the QA site using the “FF QA” favorite link in Internet Explorer.
Figure 90
Deployment to the QA environment is a success
42. Back in Release Management, navigate to the My Approval Requests view, select the deployment, and then select the Approve button.
Figure 91
Approving the release to the QA stage
Figure 92
Adding an approval confirmation comment
44. As a refresher, let’s take a look at the “Prod” stage workflow once again. No need to navigate there in the application, just refer to the screenshot below. Note that the acceptance step is not automated as it was for the previous stages. This means that the assigned approver must explicitly sign off on before the actual deployment to production will begin.
Figure 93
“Prod” stage workflow configuration
approve the deployment to production since the QA Team signed off on the previous stage. Select the Approve button.
Figure 94
Approve the request to deploy to production
46. In the Approval Confirmation window, enter a comment such as “Ready for production”. Note that you can deploy immediately or schedule the deployment for later. Use the default option to deploy now and then select the OK button.
Figure 95
Approve the request now
47. Go ahead and load the production site using the “FF PROD” favorite in Internet Explorer. You may need to refresh the page a few times before the build is deployed (displaying “Support v2.0”).
Figure 96
48. The Ops Team now needs to validate the deployment. Select the request and then select the
Approve button.
Figure 97
Approving the deployment to production
49. In the Successful Deployment Confirmation window, enter a comment such as “Ops approved” and then select the OK button. The operations team may have their own suite of tests that they run to ensure that everything is running as expected and ready for end-users.
Figure 98
Approve the deployment to production
50. Navigate to Releases | Releases and note that the release has made it to the target stage of “Prod” and has a Status of “Released”.
Figure 99
Build has been released to production
51. There are circumstances where being able to manually trigger the release of a specific build using the same process would be useful. For example, let’s say that it has been discovered that the production site has some scaling issues (that unfortunately weren’t discovered before the previous release made it to production), and we would like the QA team to do some
52. Select the New button in the Releases tab.
Figure 100
Location of New button
53. Name the release “QA Testing”, select the Fabrikam Call Center release template, select a target stage of QA (we don’t want to go all the way to production), and then select the “Select…” link next to the Build field.
Figure 101
Manually triggering a release
54. In the Search Builds window, select the oldest build.
Figure 102
Selecting a build
Figure 103
Location of Start button
56. If you followed this release to the target stage using the same process as before, you would end up with the desired build deployed to the QA environment.
57. To learn more about Release Management, please visit
http://www.visualstudio.com/explore/release-management-vs.
To give feedback please write to [email protected]
