Kabanero

Kabanero is an open source upstream project which brings together open source projects for usage on Red Hat's OpenShift OCP product.

Kabanero Operator

Kabanero delivers a Kubernetes operator which provides day 1 dependency install scripts, operator registry and OLM dependency management, and day 2 operations reacting to changes in configuration (through custom resources) and managing upgrades.

Prerequisites

Temporary move to OpenShift Container Platform (OCP)

The Kabanero Open Project intends to build on open source distributions of Kubernetes; however, the current distribution of (OKD) "The Origin Community Distribution of Kubernetes" is lagging the commercial distributions. We are trying to build leading-edge capabilities and look to leverage new features across the integrated open frameworks. Therefore, at least for the short term, we have chosen to focus on building version 4 capabilities of the commercial distributions. When the OKD Community more generally releases version 4 and are available, our community will consider returning to hosting on OKD.

Software Installation

Release 0.6

• [OCP] (https://www.openshift.com/products/container-platform) V4.3.0

Release 0.3-0.5

• [OCP] (https://www.openshift.com/products/container-platform) V4.2.0

Release 0.1 and 0.2 pre-requisities

• OKD v3.11.0+

Install

While logged in as a cluster-admin execute the install script from the kabanero-operator release

curl -s -L https://github.com/kabanero-io/kabanero-operator/releases/download/0.6.0/install.sh | bash

Upgrade

Release 0.5 to 0.6

Upgrading from 0.5 to 0.6 includes the transformation of Kabanero artifacts from Collections and Collection Hubs to Stacks and Stack Hub. When moving from an existing 0.5 install, the operator will transform the Kabanero Custom Resource spec.collection to spec.stack. The operator will also transform Kind:Collection to Kind:Stack. (Due to incompatible changes between Triggers 1.0 and Triggers 2.0, Existing webhooks must be reestablished.)

Release 0.4 to 0.5

Tekton technology transitions from Knative events to Tekton Triggers. To leverage Tekton Triggers Collection Hub requires a Tekton TriggerBinding and TriggerTemplates for each pipeline. (The Kabanero 0.5 version of the Collection Hub is enhanced with bindings and templates.) Existing webhooks must be reestablished.

Release 0.3 to 0.4

While logged in as a cluster-admin execute the install script from the kabanero-operator release.
The installation script will update the CatalogSource and previously installed operators.

curl -s -L https://github.com/kabanero-io/kabanero-operator/releases/download/0.4.0/install.sh | bash

Uninstall

While logged in as a cluster-admin execute the uninstall script from the kabanero-operator release

curl -s -L https://github.com/kabanero-io/kabanero-operator/releases/download/0.4.0/uninstall.sh | bash

Cluster Hardware Capacity

The full suite of Kabanero foundation components include:

• Istio (ServiceMesh)
• Knative
  • Eventing
  • Serving
• Tekton
  • Dashboard
  • Pipelines
• kAppNav
  • Operator
• Appsody
  • Operator
• Che
  • Operator
• Kabanero
  • Operator
  • Dashboard
  • System-Management CLI

The default configuration for the Kabanero foundation components require a scheduling capacity of approximately:

• 6 CPU cores
• 16GB of memory

The following table illustrates the default container resource requests and limits in a 3 node cluster.

appsody Application Requirements

appsody Hardware Capacity

The default Kabanero collections are built from UBI sourced containers.

appsody currently includes 5 different stacks (compressed container size):

• java-microprofile (308 MB)
• java-spring-boot2 (269 MB)
• nodejs (279 MB)
• nodejs-express (275 MB)
• nodejs-loopback (288 MB)

Estimated required capactiy for running 10 appsody applications (Assuming roughly 16MB per application):

• java-microprofile: 3240 MB
• java-spring-boot2: 2850 MB
• nodejs: 2950 MB
• nodejs-express: 2910 MB
• nodejs-loopback: 2940 MB

Kabanero Requirements

• The Kabanero Operator requires cluster-admin privileges.