For those that don’t know, recently Rubrik GA’d their 5.1 release of the core product Cloud Data Management (CDM). Now with this release comes a slew of new features and enhancements to existing feature-sets. I wanted to take a minute and highlight a couple of my favourites – a couple I think are pretty awesome which centre around the vSphere ecosystem.
First up, One-Click Continuous Data Protection
That’s right – we now have CDP for CDM – all accomplished with one single click! But before we get right into the new, let’s talk a little about the old.
Up until Andes 5.1 Rubrik only protected VMware workloads by taking vSphere snapshots and performing incremental backups utilizing VMware’s Change Block Tracking (CBT). While this approach works well for taking hourly backups, it’s hard to implement this strategy if we need a lower RPO. The challenge comes with the creation/removal of snapshots – this takes time, and when configured to occur, say, every minute could increase the load on both the ESXi hosts and the underlying storage system.
Continuous Data Protection takes a different approach – rather than utilizing CBT and the snapshot technology, the vSphere APIs for IO Filtering are leveraged in order to intercept, or filter, all of the IO being sent from a virtual machine to the storage system. Here, Rubrik can sit and transfer those blocks in nearly real-time – allowing a much, MUCH lower RPO to be met. Let’s take a look at just how easy this is…
As we can see above, the enablement of CDP is configured within the Rubrik SLA Domain by simply toggling the feature on and setting up the retention time for the restore points created. Once enabled, we can simply assign those important VMs, the ones which we wish to have CDP restore points, to the SLA domain. Rubrik then takes care of everything else for us – it’s as simple as that.
As shown below, we can now perform data recovery functions on our VM from nearly any point in time in the designated retention time set previously. This includes Live Mounts, Instant Recovery, and Exports (traditional restores)
CDP is a pretty awesome feature IMO – allowing our customers to simply toggle a switch on an SLA Domain and provide incredible RPOs to their organization.
Leverage those vSphere Tags
Being an automation junkie when VMware first introduced vSphere tags I was ecstatic. Finally, I had a way to organize my VMs by all of the craziest, wonkiest ways I could think of. I could tag VMs by department or application owners, I could tag VMs by support status, I could even tag VMs in order to create logical groupings of applications. The best part though, I could leverage all of this crazy tagging to perform automation against the logical groups I created.
Rubrik wise though, this means I can use vSphere tags to drive the assignment of VMs to SLA Domains. Now, we have always had a PowerShell cmdlet that allowed us to sync vSphere categories and tags with SLA Domains, however with Andes 5.1 this functionality is now provided natively within the CDM product – and in true Rubrik fashion, it’s super simple to configure.
As shown we can now simply select our VMs by a tag grouping from within the vSphere VMs section of the Rubrik UI – then, use the ‘Manage Protection’ button to assign any VMs containing this tag(both now and in the future) to an SLA Domain. This drastically simplifies things for a vSphere Administrator. Imagine creating a vSphere tag that maps to an SLA Domain within Rubrik which provides CDP – without leaving the vSphere client the administrator could grant near-zero RPOs to a virtual machine! Pretty nifty!
And of course, performance!
As with any product release, it’s not good enough to simply deliver new features – you have to revisit architectures and decisions made in order to enhance or build upon existing features as well. Andes 5.1 has been rearchitected in such a way to provide incredible performance enhancements to our vSphere protection as well. The way in which we identify unallocated and allocated blocks has been improved, allowing us to ensure we aren’t ingesting zero-based blocks, nor are we exporting them back when performing a restore – this drastically decreases our time to backup and time to restore benchmarks. But there is no stopping there – Andes 5.1 provides asynchronous IO when we are ingesting backups – allowing us to more efficiently utilize the nodes within the Rubrik cluster in order to provide smaller backup windows for our customers!
Above we can see a first full backup of a simple 20GB Ubuntu VM being performed with Rubrik Andes 5.0.
The same VM, with a full backup in Andes 5.1 is shown above. As we can see we have increased throughput and decreased our overall backup duration. Now – I know this is only a matter of about half a minute in the case – but keep in mind we are only dealing with 20GB – Imaging VMs which are terabytes in size – not a bad improvement 🙂
Don’t forget the APIs!
From day zero Rubrik has been 100% utilizing APIs to drive it’s UI – which means all of the new features we’ve discussed are completely available for your automation pleasure via APIs as well!
Although today we only talked about some of the vSphere enhancements and features within Andes 5.1, there is a whack of other features and enhancements in the release! From NoSQL support to SAML support to cloud tiering, you can read the complete list here! Thanks for reading!