Chris Phillips’ Blog - API, Integration and Governance

Today I sat down and installed Bind9 to act as a local DNS server. As this took me far to long I have attached a config similar to my home one explaining the changes I made from the default.

Today I updated the API Connect WhitePaper to 1.0.9 with some very minor changes. This can be downloaded from the same location as before, https://www.ibm.com/downloads/cas/30YERA2R

When designing your enterprise API Connect environment there are a number of different patterns that can be followed. In the API Connect WhitePaper I present two common patterns. This article is a third pattern that I will propose to put into a future release of the WhitePaper.

Helm is the de facto standard for deploying to Kubernetes. OpenShift does not ship with helm but instead provides their own system called Templates. Thus far I have never used OpenShift Templates and so I focus on helm. This article does not go through the differences but just shows how to get helm v2 up and running.

VMWARE ESXI is a free VMWare hosting system. The primary differences between this and the full vSphere are support, limit to one host and no of supported cpus.

TLDR : Use (Rufus)[https://rufus.ie] on Windows not DD or anything else.

WARNING: What is covered in this article is not supported. It is recommended that this is only used for a POC where Helm 2.x cannot be used.

This is a work around to a problem I have seen a couple of times. I expect this problem to be fixed in a future release of APIConnect.

When APICUP (the API Connect Installer) configures a DataPower Gateway to be deployed it internally generates a Helm chart. This Helm chart includes a stateful set that includes both an init container as well as the main container.

Out of the box the DataPower init container is set to be a busy box image deployed on docker hub. Therefore if the environment that is being used to host the Kubernetes deployment does not have internet access then it fails.

Thanks to Sara Hagger

DataPower is the gateway/runtime for API Connect. In order for API Connect to synchronise APIs with the Gateway it uses a new DataPower piece of function called the API Connect Gateway Service. When you deploy DataPower in Kubernetes with APIC UP a special log target is configured to store the API Connect Gateway Service logs. However on all other forms of DataPower this is needs to be enabled.