As part of the VCAP7-CMA Design study guide, I have put together a use case which will be used to provide context when addressing each of the exam objectives. The study guide can be found Here Use Case Mega-Awesome Transit have identified a need to increase the reliability and reduce the complexity when updating their dispatching system during update periods. Currently updates to the dispatch system cause interruptions for up to 12 hours.
Gather and analyze business requirements
I recently updated my Ubuntu install from 17.04 to 17.10, which meant the kernel was updated to 4.13. As a result, I was no longer able to run VMs with VMware Workstation. I found some posts regarding the error and that it was due to a change in the way the Linux kernel handles paging. To resolve the issue I needed to replace the vmmon modules. I didn';t take down the entire error message, but this post references the same error.
Today I came across the need to use PowerCLI on a computer without internet access. Thankfully I found this blog post detailing the steps that I needed to perform, however not all was rainbows and unicorns. I quickly ran aground when attempting to import the modules. After a bit of research, I learnt that Microsoft had changed the module directory structure in PowerShell v5. This new structure is not compatible with previous versions.
vRA uses Custom properties to help define parameters and the behaviour of requests. In the previous post CHEF: vRA Integration, Property Groups and Blueprints we used custom properties to define the properties of our Chef deployment. In this post, we look at using custom properties to assign the node environment. The Chef default client deployment workflows form the basis of the examples. A brief on custom properties We use custom properties to manipulate many aspects of workflow execution.
The Chef plugin provides some workflows to help with getting the vRA integration up and running with minimal effort. I would suggest treating them as samples to build functionality on top of. There are some limitations from the default state which might not provide sufficient flexibility for production usage. Setting up the Property Groups: Launch the vRO Client and navigate to the Chef workflow ‘Create Property Group for Chef EBS Workflows’.
Chef Integration with vRA: Part1 In the previous Chef integration with vRA post, we covered how to install the Chef plugin for vRO. In this post, we will look at provided workflows used to install the Chef Client on Windows and Linux VMs. The workflows are called by the Event Broker Servers (EBS) after a VM is provisioned. The Client Provision Workflow To begin launch the vRO client and navigate to the Chef workflow folder.
The Chef agent is installed on a VM after the VM has been deployed and completed the ‘Machine Building’ stage of deployment. This is achieved by creating an event subscription through vRAs Event Broker Service (EBS). When an event triggers an EBS subscription, vRA communicates with vRO causing a workflow to run. This means the first step of configuring Chef integration with vRA is to configure the Chef vRO plugin.
Building use cases to learn new products allows for new knowledge to be applied and persist much longer. Learning new topics and maintaining that knowledge is a skill. By dedicating time to understanding a range to topics, you start to understand how to learn and become more efficient. Watching videos, reading, rewriting and review are great ways to get knowledge in. But, unless you apply context and apply learning the knowledge will slip away quickly.
One of the challenges many of us come across when taking a more automated approach to infrastructure is the sheer number of tools available. This is especially true when you come from a one tool for multiple roles environment. Every week, there appears to be another tool for you to choose from. Finding where to start is quite daunting, and the opinions of others leads to much second guessing. One of the lessons that would have helped me earlier on, is understanding what the role of a specific tool is and how it fits into the overall process.
Terraform is an awesome tool used to manage infrastructure using the Infrastructure as Code philosophy. Modules called Providers enable Terraform to communicate with a number of different cloud providers. Post deployment tasks are performed through a separate set of modules called ‘Provisioners';. A provisioner is used to execute commands locally on an instance after it’s been created. One such provisioner enables the Chef client to be installed on the newly provisioned instance and the instance to be added as a node to the Chef Server.
Configuration management systems are used to deploy and ensure a standard environmental state. You might use a configuration manager to ensure that a file exists, or a certain setting is always applied to a system. As your IT infrastructure grows or becomes more distributed ensuring a consistent state becomes more difficult. Through deploying and ensuring consistent state across multiple systems based on central policies configuration managers help to lower the administrative overhead, improve security and reliability.
While at Interop ITX 2017, I met with Eric Pulvino from Cumulus Networks and learnt a bit about where Cumulus was heading and what’s new. One of the topics that got my attention was using a CICD pipeline for end to end testing of network configuration changes. Cumulus Linux in a networking operating system (NOS) from Cumulus Networks. At its heart, this is a Debian Jesse distribution with additional networking smarts.