One of our customers is in the process of decommisioning their OpenShift v3.11 cluster. This cluster is currently still used for building customer specific base images. Over time quite a few elaborated pipeline builds (based on Jenkins) have been developed for that purpose.
The customer wanted me to migrate the existing pipeline builds on their v3.11 cluster to Tekton (aka OpenShift Pipeline) builds running on their new v4.9 cluster. This task turned out to be quite pesky. Tekton is a beast in many aspects.
Today software often needs to be run in cloud environments. Newly developed software, especially microservices are developed with cloud readiness in mind.
But we not only have microservices in business environments, we also have integration software. This type of software is developed and designed to connect external services to internal ones.
Some time ago, I started a project to create a Helm based operator for an OpenShift application. I used the Operator SDK to create the Helm operator. The Operator SDK documentation describes the parameters pretty good, and it contains a simple tutorial. it does not, however, describe the complete development cycle. This article aims to describe everything from creating the operator to the point where you can upload your operator to OperatorHub.io. We start with a basic Helm Chart. With this, you can install Nginx as a StatefulSet. You can find the source code in my github repo. Before we can start with creating an operator, we need to fulfill some prerequisites.
Last summer I watched the Red Hat master course about Kafka from Sébastien Blanc. The Kafka setup in Kubernetes presented in the course looked pretty easy. The Kafka client implementation for Java seemed to be easy as well. Furthermore, I wanted to use Kafka for a long time, so I got the idea to extend my Istio example. Each time a service is called, a message is sent to a topic. The service (implemented in Quarkus), as well as the Kafka cluster should be in an Istio Service Mesh and secured with mTLS. I found descriptions of Joel Takvorian that Kafka works with Istio, so I knew (or at least hoped) that my plan should work.
This article will describe the overall architecture of the example and which obstacles I encountered during deployment.
So you have this nifty web application deployed on your OpenShift cluster and you want to make it accessible by the whole world with HTTPS under the name
coolapp.<mydomain>. Unfortunately you face several issues:
Exposing the service to your web application leaves you with a route using the self-signed certificate that was generated during setup of the cluster. None of the browsers in the wild will trust this certificate.
The self-signed certificate dictates URLS of the form
https://<appname>.apps.<clustername>.<mydomain> (or whatever domain suffix you configured). Not very nice.
You might mitigate the previous issues by getting an official certificate signed by a generally trusted institution. But you will have to pay for it.
And you will have to pay for it not only once but every year (latest every 389 days) thanks to recently tightened certificate policies installed by all major browser vendors.
Worst of it all: You must not (by any means) forget to apply for a new certificate in a timely manner and replace the certificate in your route before the old expires. Otherwise some people might get pretty angry about you.
Let’s Encrypt to the rescue!
During a discussion with a customer, we talked about which steps are necessary to add an application to a services mesh. Which should be no big deal. Unfortunately, there is not a simple guideline how to do that for the Red Hat OpenShift Service Mesh. Furthermore, I was not sure how the requests for the application would look like in Jaeger. To clarify these points, I created a small application. Which I then deployed on OpenShift and added it to a service mesh control plane. This is the documentation of the steps that I have done.
During this year’s Red Hat Summit I had the chance to get a glimpse of the latest version of Kiali. This version had some nice features, like the traffic flow of the application graph during a time period (Graph replay). It also contains wizards to create destination rules and virtual services. This demo has struck my curiosity to get the hands on this Kiali version. One obstacle for me was that my Kiali is running in Red Hat OpenShift Service Mesh and is controlled by the Kiali operator. Currently, it is using version 1.12. The version that I wanted to try was the latest release version (1.17). The Red Hat OpenShift Service Mesh does not support this version. This article describes what we need to do in order to replace the Kiali version of an Red Hat OpenShift Service Mesh with the latest version of Kiali.
Some time ago, I did a webinar about the RedHat Service Mesh, which is based on Istio. For this webinar, I prepared a demo application. Among other things, I wanted to show how to do the authentication with JWT token in general and, more specific, with Keycloak. This article will describe how to configure Keycloak. In the second article, I will show you what problems I encountered running the application in Istio and how I figured out what was wrong in my configuration. You can find the article here
In the article, I’m going to describe what we can do, if we configured our application to use Istio, but it is not working like intended. Originally, I wanted to give a detailed description what problems I encountered during the creation of my webinar and how I fixed them. However, I came to a point where this would be a very long one. I hope that you don’t mind that I shortened it and just describe which tools are available to
debug the Istio configuration. In my previous article I described how to configure Keycloak for my webinar. So without further ado, let’s start.
In this article, I will show you how to install Red Hat OpenShift Container Platform 4.3 (OCP) on VMware vSphere with static IPs addresses using the openshift installer in UPI mode and terraform. In contrast to the official OpenShift 4.3 install documentation, we will not use DHCP for the nodes and will not setup the nodes manually - instead we will use static IP addresses and terraform to setup the virtual machines in our vCenter.
So here is another one of our series Installing Blahblahblah on OpenShift. This time it is about getting MongoDB to run on OpenShift - the way recommended and promoted by the MongoDB guys. The whole setup is still in beta stage as indicated on these two entries in Red Hat’s container image catalog. You can get your MongoDB instance up and running on OpenShift. But most of the required steps have to be performed on the command line, contrary to the impression given by MongoDB, Inc that once you get the MongoDB Operations Manager up and running everything can be achieved via this tool’s GUI. Some operations in the Operations Manager simply do not work (yet) on OpenShift.
With the release of OpenShift 4.x Red Hat left no stone unturned (compared to previous 3.x versions). Among many things Minishift became Red Hat CodeReady Containers. Having been a big fan of Minishift I recently wanted to give CodeReady Containers (aka CRC) a try.
Turned out this is not that easy - at least if you want to run CRC on a Linux that does not come from Red Hat (or its community). This article gives instructions for all those people out there who want to run CodeReady Containers on Ubuntu.
We recently had to install a bunch of applications on a customer’s shiny new OpenShift 3.11 cluster. Among others also GitLab. Turned out getting GitLab up and running on OpenShift is not so easy. What I found on the Internet about installing GitLab on OpenShift was partly outdated and not 100% accurate. Most information was about getting GitLab into a Kubernetes cluster. So I had to adapt these information to the situation in an OpenShift cluster.
This article is the conclusion of all these findings and efforts and gives a step-by-step recipe on how to install GitLab on OpenShift.
Recently, I stumbled on a situation where I wanted to add a couple of values to an OpenShift deployment configuration. Previously I had modified or added a single attribute in a yaml file with
oc patch. So I started to wonder whether it is possible to update multiple attributes with
oc patch as well. To get right to the result: Yes, it is possible. This article will show you which features
oc patch and likewise
kubectl patch really have, beside a simple modification of one attribute.
After some time, let’s move on to another topic around making OpenShift environments more developer friendly. This time we are going to look at what happens, when a system test actually failed, and how to enable developers to properly react.
Kubernetes and OpenShift have a lot in common. Actually OpenShift is more or less Kubernetes with some additions. But what exactly is the difference?
It’s not so easy to tell as both products are moving targets. The delta changes with every release - be it of Kubernetes or OpenShift. I tried to find out and stumbled across a few blog posts here and there. But they all where based on not so recent versions - thus not really up-to-date.
So I took the effort to compare the most recent versions of Kubernetes and OpenShift. At the time of writing v1.13 of Kubernetes and v3.11 of OpenShift. I plan to update this article as new versions become available.
The OpenShift command line interface is a very powerful tool which is quite useful for beginners and advanced user of OpenShift alike. Some of its features are not well documented or not documented at all. In this article I would like to shed some light on commands that I personally find useful and that are, from my observation, not widely in use. So without further ado, let’s start with the commands:
Our world is full of various processes: tracking of goods delivery, currencies trading, monitoring of server resources, hotel bookings, selling goods or services etc. Since these processes occur over time, they can be described by time series data.
Successful businesses always take advantage of their data by analyzing it and then making predictions (e.g. predicting volume of sales for the next month) and business decisions (e.g. if the volume of sales grows then additional goods need to delivered to a warehouse).
There are a number of technologies for analysing the time series data. This article gives an introduction to one of them which is called TimescaleDB which is an open source solution for time series data analysis based on battle-tested PostgreSQL DBMS.
Let’s move on with this little series about how OpenShift environments may fall short in terms of developer experience.
Today we focus on the role that system tests have in an OpenShift infrastructure and what might possibly go wrong here testdata-wise.
In some OpenShift environments for building and delivering software we notice that the needs of developers, arguably a group of people who will have a great deal of contact with the platform, are not met as thoroughly as would have been possible.
Especially when it comes to software testing there is often much room for improvement. The usage of container platforms can improve testing techniques a lot but might also be a major blocker when it comes to the provided infrastructure. Good testing is already hard. Everything that makes it even harder, by forcing your developers into workarounds or compromises on testing quality will result in larger round trips, more testing effort, less valid testing, in short: wasted time.
So in this mini series of blog posts we will have a look into some possible fields of improvement and give recommendations on how to fix the respective situation.
Today we evaluate the fact, that some CI/CD setups for OpenShift may spoil the most simple type of testing a developer uses: Just running the software locally - in OpenShift.
Docker Headless VNC Container 1.3.0 has been released today. The different Docker images contains a complete VNC based, headless UI environment for testautomation like Sakuli does or simply for web browsing and temporary work in a throw-away UI container. The functionality is pretty near to a VM based image, but can be started in seconds instead of minutes. Each Docker image has therefore installed the following components:
There are a lot of articles that show how to monitor an OpenShift cluster (including the monitoring of Nodes and the underlying hardware) with Prometheus running in the same OpenShift cluster. This article however is based on a different scenario: You are responsible for an application on an OpenShift cluster and want to monitor just this application, but you don’t have any administrative permission on it. The reason for this can be that you are working in a big company where the operation of the OpenShift environment is outsourced or the process to introduce a new monitoring solution takes way too long or the current monitoring solution doesn’t match your requirements and so on.
In this article I’m going to show you how to setup the monitoring of a demo application in 6 easy steps. The example is built in that manner that it will be easy for you to do the same for your application. A side note: If the OpenShift cluster that you are using will be monitored in the future with a different Prometheus setup, you don’t need to start from scratch. You might need to tweak the configuration of your scraping a bit and you need to move your dashboard to a different Grafana but that should be it.
Docker Headless VNC Container 1.2.0 has been released today. The different Docker images contains a complete VNC based, headless UI environment for testautomation like Sakuli does or simply for web browsing and temporary work in a throw-away UI container. The functionality is pretty near to a VM based image, but can be started in seconds instead of minutes. Each Docker image has therefore installed the following components:
The GitHub repository toschneck/openshift-example-bakery-ci-pipeline contains, the sourcecode for the examples of the talk Continuous Testing: Integration- und UI-Testing mit OpenShift-Build-Pipelines at the Redhat/ConSol OpenShift-Day:
Sowohl End-2-End-Testing als auch End-2-End-Monitoring folgen dem gleichen Paradigma – sie betrachten eine Applikation aus der Sicht des End-Users. Hier darf es keine Rolle spielen, in welcher Oberflächentechnologie die Applikation geschrieben ist oder in welcher Art sie mit dem End-User in Verbindung tritt. Genau an diesem Punkt setzt das Open-Source-Tool Sakuli an.