Hi folks, it seems we definitely hit something interesting based on the number of responses to part 1 of this article.. Talking about the Next Generation Hypervisor, or “Red Hat Virtualization Hypervisor” as it will officially will be known as, has definitely gotten positive attention. We’re going to take another look today, with a deeper focus on “Cockpit”, the hypervisor administrative plugin and web console.
Let’s get started.When I wrote part 1 of this article, I had been working on the “beta 1” release of the product. By the time I got to work on part 2 of the article, “beta 2” had been released, and several things had already been ironed out. Most notably, the ability to update the hypervisor itself.
If you remember my original description of the new design, it’s somewhat of a hybrid of a “thick” and “thin” hypervisor in that it allows the flexibility of a thick, but leaner. You can add additional packages using RPM and YUM, for monitoring agents, new drivers, etc. You could even update the entire hypervisor image via YUM. Just note that the RPMs that you added prior to the image update will need to be re-added.
In the embedded video, I show the process of updating the image; it’s actually very cool. Typically you would do the updates from within the management console, but as I note in the video, sometimes you need to see what’s happening under the covers.
Once the image is updated, I move on to “Cockpit”. Cockpit is such a very cool addition to the hypervisor that I had to do a brief walk-through. As I point out, it is NOT a replacement or competition for RHEV-M, it is a complement. Watch the video, I think you’ll agree.
What I don’t show, is accessing Cockpit. It’s as easy as “https://ip_or_fqdn:9090” in your web browser, then logging in with your credentials. Adding additional hypervisors is also very easy – I show you how to do that. Graphs and stats are color coded to the different hypervisors, making it easy to match metrics to nodes.
Next in the walk-through, I highlight services. Specifically the ability to not only view what services are available, but their status, their respective logs, and the ability to start, stop, restart, and reload specific services. This was not previously available in the old admin tool or RHEV-M.
Next up, we can view the system logs. We can even filter them on their specific error level (error, warning, etc). Storage and networking resources can also be interacted with in ways that couldn’t be done without logging into the hypervisor.
The ability to run a diagnostics report and view a console right from Cockpit also adds to the functionality. I seriously can’t wait until people can start using this in production.
Granted, sometimes you need to “ssh” into a node and that ability is absolutely still there, but clearly the need for “operations” level tooling has existed. That’s what Cockpit represents to me: operations level tooling, and this is a good thing. The troubleshooting gets easier, the sense of what is going on within that hypervisor (and other hypervisors), is greater.
Check out the video. Best viewed in fullscreen:
Let me know what you think in the comments section.
Hope this helps,