Kubernetes clusters may have a well-deserved reputation for being a challenge to set up. Which is why some vendors are developing tools that make it simpler to deploy and manage Kubernetes. The latest example of those tools comes in the form of an update to the Tectonic distribution of Kubernetes curated by CoreOS.
Version 1,6 of Tectonic provides support not only for the latest version of Kubernetes, but also for Terraform, a tool developed by CoreOS that helps make it possible to use scripts to install Kubernetes on Amazon Web Services (AWS) or on a bare-metal server. Via Terraform, an IT organization can publish configurations that can be referenced by Kubernetes. To make that possible on bare-metal servers, CoreOS integrated Terraform with Matchbox, a set of tools for booting and provisioning Linux container clusters.
The latest release of Tectonic also promises to improve workload separation on Kubernetes clusters via beta support for Kubernetes taints, tolerations and pod affinity, which provide a layer of isolation between the Kubernetes control plane and user workloads, in addition to spreading Kubernetes services across multiple nodes. By employing this approach, workload separation is enabled by default when users deploy multiple controllers and worker nodes.
Other new additions to Tectonic include support for role-based access controls, improved auditing capabilities and support for the Security Access Markup Language (SAML).
Use of Kubernetes in the cloud and on bare-metal servers is starting to increase, says Mackenzie Burnett, product manager for CoreOS Tectonic, and, in the case of on-premises deployments, utilization rates of servers significantly increases. However, some organizations still prefer to deploy Kubernetes on top of, for example, OpenStack or VMware because they have existing investments. Depending on the level of maturity of the organization, the rate of transition away from management frameworks based on virtual machines will vary considerably.
Burnett notes the three biggest inhibitors to Kubernetes adoption are organizational inertia, a lack of IT staff with Kubernetes skills and the frequency at which Kubernetes as an emerging technology is updated. To the solve the latter problem, Burnett says CoreOS automatically updates Kubernetes installations on behalf of customers. CoresOS finds that most customers have a Kubernetes project underway, but when they move that project into a production environment, they quickly develop a marked preference for CoreOS to manage the update process.
The education issue, meanwhile, is being addressed by certification initiatives led by Cloud Native Computing Foundation (CNCF).
Of course, Kubernetes is only one of several platforms for hosting containers in a production environment. The strength of Kubernetes comes from the fact that in addition to being employed by Google in production environments, the number of vendors lending support to the open-source project is extensive. That doesn’t necessarily guarantee adoption within traditional enterprise IT environments, but it does go a long way to make Kubernetes a cluster platform that more IT professionals have had some exposure in the last 12 months.