In part I of this series, we created a vFabric Application Director blueprint and deployment profile for a corporate network and deployed it with an execution plan. Then, we showed how simple it is to redeploy on Amazon EC2.
In part II of this series, we took the same application, and swapped out one of the application components, specifically we changed the data store to vPostgres on an external node, and updated the application running in production.
In this post, we will pick up with the now vPostgres-driven app and see how we can perform a scale out and an update to the application itself that will include adding 3 nodes to the app cluster and updating the WAR file.
In total, there are 8 steps explained in this post and a video, at the end, with the whole process:
1. Scale Out–Increasing Cluster Size
2. Scale Out–New IPs Resolve
3. Scale Out–Review Execution Plan
4. Scale Out–Test App
5. Update–Change Configuration
6. Update–Modify the Script
7. Update–Review Execution Plan
8. Update–Confirm Node Updates
1. Scale Out–Increasing Cluster Size
We start with our last app, version 1.5.0. Then, we create a scale-out update by increasing the cluster size from 1 to 4. This is as simple as finding the blueprint, a logical model of the deployment, and changing a property. When we click update, we can see the scale out happen in real time.
2. Scale Out–New IPs Resolve
In this part of the video and in the screen shot below, we can see a status dashboard of the deployment. The original two IP addresses stay the same, while three new ones are resolved. vFabric Application Director’s dashboard view of the deployment is shown below:
- The top, left panel shows the task details–this includes details about the task like task ID, start time, end time, deployment location, and more.
- The top, right panel shows the blueprint–this is the logical model, built in vFabric Application Director, that visually depicts the entire application topology.
- The bottom, left panel shows the VM details–this information includes details about each VM like the cloud template, host name, IP address, and network.
- The bottom, right panel shows the execution plan–this is a visual depiction of the sequence of activities performed by the system.
3. Scale Out–Review Execution Plan
We can expand the bottom, right panel (from above) to look at the Execution Plan within vFabric Application Director. We can see the work being done by the system to set the additional nodes.
4. Scale Out–Test App
In the last step of the scale out, we open up each node to test that the app is working.
5. Update–Change Configuration
Next we are going to complete an update to the application itself. Here we will change the WAR file value to 2.0, effectively renaming this to version 2.0. This is as simple as updating a property.
6. Update–Modify the Script
As well, we will go into actions and modify the update script. This script is used when the application is redployed.
7. Update–Review Execution Plan
By expanding the Execution Plan from the status dashboard again, we can see that the only update is on the node array for the contact app and, specifically, the update script. The Execution Plan shows that only certain changes are redployed–not the whole app.
8. Update–Confirm Node Updates
From the status dashboard drill-down into the Execution Plan, we can again see that each node has been updated, and can go into the application to test as well.
Then, we go log into the app to see it running.
The Video: Scale-out & Update for a Clustered Application
Learning more about vFabric Application Director:
- Learn How Apps will be Deployed in the Future on services like vFabric Application Director
- Learn what “Cross Platform Cloud Application” means
- Learn about Model-Driven and Workflow-Driven Apps
- Read how VMWare IT is managing Policies with Provisioned Apps to Save Money
- Download a free trial and learn more about vFabric Application Director
