The latest iteration of the oVirt Workshop rolled into Düsseldorf on October 16th, with loads of new content to share with participants and the start of new dialogs about the future of the open source virtual datacenter management platform.
The day started with a well-received presentation from Michal Skrivanek on what’s new in oVirt 3.5, which unbeknownst to the attendees, would be released the very next day.
Next up, Antoni Segura Puimedon walked everyone through the details of how oVirt and OpenStack are integrating, which generated some good discussion, primarily because of all of the technologies out in enterprise IT, people are very interested in OpenStack and how they can use it. This is a topic worth exploring in detail in the near future.
Barak Azulay then talked about ways oVirt works with Foreman. Foreman integration has been a big part of the oVirt story over the past year, as admins and users make the jump to better life-cycle management. Combining that kind of management with a superior virtual machine manager makes a lot of sense.
After the break, Federico Simoncelli rolled out a look at how oVirt is progressing in the world of a hot topic in IT right now: Docker. There are two ways oVirt will work with Docker: using oVirt to manage Docker containers directly (with a little help from Kubernetes for orchestration), and by containerizing oVirt Engine to run inside Docker and create the fastest and simplest oVirt install ever.
Things got ahead of schedule by this point in the day, so Fabian Deutsch graciously stepped in to pinch-present an overview of oVirt Node: what it is and what’s ahead for this just-enough OS tool. There’s a lot of work being done on Node, and the audience was keenly interested in hearing more about it.
Following lunch, the afternoon sessions became much more free-form and interactive.
Leading this afternoon off were presentations from three oVirt users who shared their own stories on how and why they use oVirt within their organizations. Jorick Astrego, René Koch, and Markus Stockhausen talked about the whats and the whys, and garnered a lot of questions from other users and oVirt developers alike to see what use cases they had for oVirt.
Using the brand-new oVirt USB sticks created just for the oVirt Workshop and LinuxCon EU, Sandro Bonazzola demoed the key new features in oVirt 3.5 and also showed audience members some tricks and techniques they might not have known about otherwise. Martin Sivak jumped in with an impromptu follow-up on how to set up pre-defined and custom scheduling policies within oVirt.
The final presentation of the day had Barak returning to the podium, joined by me, to lay out what is going to be coming with future versions of oVirt, which as hot unplugging of CPUs, Ceph integration, load balancing based on memory, more integration of the PatternFly interface, and CentOS/RHEL 7 support for oVirt Engine, to name a few. Attendees also talked about what they’d like to see in upcoming versions of oVirt, which included "stability;" "serial consoles;" saving snapshots to disk as a poor-man’s backup but through the API; and enhanced integration with KSM. There were more suggestions, which will be detailed in another report.
Overall, the oVirt Workshop was a very positive experience for all who attended, and it wasn’t the only oVirt presence in Germany last week. Dozens of people stopped by the Red Hat booth to talk about oVirt, and several engineers and community members spoke about oVirt in the co-located LinuxCon, CloudOpen, and KVM Forum events. oVirt is gaining quite the reputation within the IT community, and much of it is very, very good.