How to containerize Workload Automation

Workload Automation (WA) 9.5 Server, Console, and Agent components are available as Docker images that can be deployed easier and faster.
 
You may ask yourself whether there is a way to take advantage of the benefits even though you are not yet planning to deploy to the latest version. Is there a way to do this? yes, there is. You can still enjoy the benefits of Docker with the previous WA versions, in this blog post we will show you how.

 
WA is not designed to be containerized, therefore there are no supported ways to containerize IWS and DWC master. Although this applies to production environment, it does not necessarily apply to the test/QA environment.  We run IWS instances in containers along with virtual and physical machines in our test/QA environment. We also maintain a registry of docker images for all releases and components, this allows us to compose and deploy virtually any WA environment in a matter of seconds.   
 
In the same way, containerizing any WA version 9.x can be beneficial for any customer that wants to deploy WA in their test/staging environments. The main benefits of this approach are:

• Density: On the same machine, both physical or virtual, you can run greater number of WA instances compared to other approaches (i.e. virtual machine) as all instances share the same operating system.
• Isolation: Each instance runs in a separate container and therefore no more configuration issues (you can use the same deployment template using the same listening ports, installation path and so on).
• Performance: You can deploy any WA component in a matter of seconds.
• Easy: You can compose and deploy any IWS environment by just editing a ‘declarative’ file.

Figure 1: Test environment.

​Let’s start with building an image for the Master Domain Manager (WA Master Image) and an image for the Dynamic Work Console (DWC Image)

Figure 2: IWS, DWC images (layer details).

The first step is to build a WebSphere Image that can be used as a base image for both the WA and DWC Master Image. This image contains both Installation Manager and WebSphere products.

​The following is an example of the DockerFile:  

In the example, we used centos:7 as a base image, however you can use any Linux supported distribution, and we assume to have all the packages served by an HTTP server.
 
From this image, we can proceed to create the WA Master Image. 
Now the installation of the Master requires a supported RDBMS up and running for storing both the configuration and data. At the end of the installation we need to save the WA database and make it available every time we spin up a new WA instance.
 
An easy way to proceed is to package a RDBMS in the same image along with WA.

At this point we have a base image with all the prerequisite in place and we can proceed with the WA installation. Notice that both the DB2 and WA instances are bound to the hostname of the machine. We have two options:
 

1. Build the image as usual via a Dockerfile, every time you start the container, make sure to reconfigure the hostname in DB2 and WA
2. Use the same hostname when building the image and when running the container.

 
In the following example, we use the second option:
 
Let’s go ahead and build the WA image. Once you customized the response file, run the image specifying MDM as the hostname and install WA in silent mode.

At the end of the installation we have a container where WA is up and running. Let’s commit the container.

Now we can deploy the WA master image multiple times on the same, or on different machines. The main advantage to this method is to install one and deploy N.

Figure 3: Build once deploy everywhere and any times.

With the same approach, you can build an image for the DWC.
 
At this point you can compose an environment with DWC MDM running on a private network.

Figure 4: Multiple isolated environments.

The following is an example of composing an environment with DWC 945, IWA 945.

​At this point you are ready to run the:

​Conclusion
You can enjoy the benefits of Docker with the Workload Scheduler in just a few simple steps.

Maurizio Simeoni, Senior Software Engineer
Maurizio is Senior Software Engineer at HCL Technologies with 15+ years experience in developing enterprise-wide applications and continuous delivery infrastructure. Strong skills covering all the software life cycle development roles and customer facing activities like advocacy, product delivery and support. Agile development, Continuous Integration and Continuous Delivery are the cornerstone of my work. In the last year I embraced the DevOps career path, leveraging the strong skills in continuous delivery process, test automation, containerization, micro services, cloud infrastructure and DevOps tools (Jenkins, Udeploy, Chef, Ansible, Docker). I think continuous learning is the key for any personal growth. I like learning new technologies, languages, tools that can help me and my team to be more productive.