Using RHEV 3.6 (Live Snapshot)

Don’t you wish there was a way to easily save your hard work beyond building a VM template in RHEV but before it goes into production? That pristine state that matches the template, complies with security, complies with business rules, has all patches, and it has all binaries. Or 2 months later when you get all of your new updates and you add them to your VM’s and you wish you could just build new VMs based on that newly patched VM? You can.

Live Snapshots.

This week I’m continuing my series on getting actual work done in Red Hat Enterprise Virtualization and pushing on with something that is often overlooked. Live snapshots have been been around for several minor releases now and can be a real time saver. I hinted at the 2 big use cases above, now I’ll explain them a little better.

Use case #1: Providing a tactical means of rollback (change management) – this is really the primary use case for live snapshots. You’ve created a foundation for your VM(s) by way of one or more templates, and that saves you time in how many steps it takes to create VMs and removes “one offs”. But the live snapshot extends that functionality by allowing you to take point in time views of the VM that may be critical in the lifecycle. Such as just before it goes live, just before it gets updated, just before it gets a configuration change, or just before you go on vacation.

Think about it, if you’re bored at 9:30 in the morning on a Tuesday at work, sure you can troubleshoot who did what to your VM. But if you’re at Disney World with your family answering your cell phone and your significant other gives you the death stare, you should be able to just say, “Roll it back to my last snapshot… yeah, the one I took 10 minutes before I left for vacation. Goodbye.” Crisis averted. You’re welcome.

Use case #2: Creating new VMs – This is something that isn’t often thought of with live snapshots, but is absolutely available. Think about it: You’ve created the VM from template, and now it’s up to date with the latest patches and such. The live snapshot function in RHEV will allow you to create new VMs from snapshot. And why not? Why do the work over again each time you create a new VM? Or simply use it as a temporary stop gap until you get your templates update with those same patches and updates.

In the demo I cut below, I highlight how to take the snapshot as well as how to rollback, and then “rollback the rollback”. As the feature name implies, live snapshots are taken while the VM is running, so that there is no disruption to the VM, the application, or the users. However, if there is a need to rollback to a snapshot, you will need to shut the VM down. But think about it, if the problem is bad enough to cause you to rollback to a snapshot, there is already a disruption, so shutting the VM down is likely not the worst thing in the world here. But what is this “rollback the rollback?” What I mean by that is that simply switching to a snapshot does not commit you to that snapshot. You literally have to click on the “commit” button in order to tell RHEV that the official version of that VM is now the snapshot. If you decide that the snapshot isn’t what you really wanted, click on “revert” instead, and you go back to the original VM.

Let’s check out the demo (fullscreen is recommended):

Now, what I didn’t show in the demo is the “guest agent”. It is not required, but it will help if you also need the memory state saved as part of the snapshot. If you aren’t concerned with memory state, don’t worry about the guest agent; you’ll just see a warning message about memory state when you go to run the snapshot.

As always, comments and questions are welcome.

Hope this helps,

Captain KVM

Agree? Disagree? Something to add to the conversation?