It’s been a while since I posted here, but as part of my work on OpenShift AI (that transition could be a topic for another post), I expect to have more interesting content to share. This post covers an exploration I participated in, focused on sending traces to MLflow for monitoring agentic workflows in production.

This post describes the next step in our journey to deploy MTV (Migration Toolkit for Virtualization) on Openshift Local: deploying a custom image to Openshift Local without going through an external registry like quay.io

Recently I wanted to experience with Forklift, a project that enables to migrate virtual machines from Red Hat Virtualization (RHV) to Openshift Virtualization. Here, I’ll describe how I did this using Openshift Local which can be useful for development or experiments (Openshift Local is not meant for production use).

Few months ago I’ve examined the ability to propagate configurations to Kubernetes clusters in a multi-cluster environment. Here I describe the process and tools in the hope that it will be useful for others that try to do the same thing.

In this post I share a solution I have implemented for publishing nightly builds on GitHub.

In this post I describe how to run a plain (vanilla) Kubernetes cluster that spreads across physical machines installed with Fedora 31.

Ten years ago (2008) I submitted my first contribution to an open source project named muCommander that I maintain to this day. This post provides a short description of the project, recent changes we have made, challenges we are facing, and some future plans.

When I was a kid we used to play this game in which someone thought about a message and told it to his friend. The latter told the message he heard to another kid, and so on. Eventually, we were amused to see the difference between the original message and the one that the last kid has heard. This post describes our approach for addressing a similar problem that was caused by having too many abstraction layers in oVirt. Each layer converted its input in order to report it to the next layer “in its own words”, resulting in cumbersome and error-prone business flows in our platform.

Few weeks ago I attended a session on the new API specification in oVirt. While the motivation was well explained and the overall design of the solution made a lot of sense, the presented language made me wonder whether it is the best language for the problem at hand. In this post I argue we can achieve a better language for the API specification in oVirt by using an external DSL rather than an internal DSL.

Recently I’ve been working on improving the scalability of monitoring in oVirt. That is, how to make oVirt-engine, the central management unit in the oVirt management system, able to process and report changes in a growing number of virtual machines that are running in a data-center. In this post I elaborate on what we did and share some measurements.

This post demonstrates the development and use of Application-Specific Aspect Languages through an example of such language for synchronization in oVirt.