It has a quick setup and then it is hands-off. It’s a backup product, and you want to set it up and forget about it. For the VMware side, you can set up policies by resource pools, clusters, and folders. This helps with the "set up and forget" process. They have added physical servers (Windows and Linux), plus SQL. This is a huge plus as everything can be managed under one policy.
VM backups are common and there are a lot of vendors. The thing that sets Rubrik apart is the instant recovery feature. While other vendors have this, the storage is on you. In my case, this storage was always on a cheap disk. With Rubrik, you get an array that can support high workloads until you have a main window from which you can migrate the VM back to production storage.
Replication is also great. With one click of a button, you can replicate between Rubrik devices and public cloud storage.
The reporting can be vague, especially when something is failing. AD integration is cumbersome. For highly compliant environments, Rubrik does not patch the IPMI firmware on a regular basis. Be aware that you should design the roll-out with this in mind. Otherwise, your security teams will be calling about vulnerabilities that they can reach.
We have used this solution for over two years.
There have not been any stability issues.
There have not been any scalability issues. This is one of the pluses and a driver for buying Rubrik. If you need backups to run faster, or if you need more storage, you just add another node or Brik.
The technical support is good. They will escalate quickly if it is needed.
We had other solutions before this one. We used Veeam and VMware backups. We left Veeam due to storage management. VMware backups were not as feature rich as Veeam or Rubrik.
The installation is straightforward. You fill out the sheet with names and IPs, and Rubrik sets it all up. However, it does use a lot of network ports, at least one 10-gigabit and one 1-gigabit port per node. That adds up quickly with multiple Briks that can have four nodes each.
Make sure you create different networks and segmentation. The IPMI 1-gigabit networks need to be locked down and not allow other networks to be routed to them.
Understand what you really need to backup. Do you need a backup, or can you use storage snaps? Run their PoC trial. It will give you a good idea of what your compression will look like and it will allow you to estimate the size you need.
We used Veeam and VMware before. We also looked at Commvault.
Understand what you want to backup. Rubrik makes it very easy to go in and back up everything, but that might not be in your budget.
Originally posted at https://vinfrastructure.it/2016/01/rubrik-reinvent-the-backup/
Some months Rubrik announced the release of version 2.0 of Rubrik Converged Data Management to deliver complete data protection and rich data services at a global scale (see also my previous post).
The idea is quite simple but disruptive: reinvent the backup models in order to eliminate backup software by integrating data protection, instant recovery, and DevOps infrastructure into a single scale-out product!
Data protection is usually something complex, or at least with several components:
Rubrik can remove all of them:
An simple move to converged, scale-out and appliance based architecture:
The idea is similar to Cohesity approach (for example) but Rubrik stats with the data protection in mind as a first scope of their solution, so several features (in this area) are better developed (yet). It’s interesting see also people from Symantec, Data Domain, Veritas in the investors list.
During the last IT Press Tour #17 I’ve got the opportunity to learn more about this solution, directly the board of Rubrik: Bipul Sinha (CEO), Arvind Jain, Arvind Nithrakashyap. For the technical presentation there was also Chris Wahl.
As written the solution is (physical) appliance based, using the model of Data Domain (that has change the tradition backup infrastructure), but using a “Web-Scale” model typical of hyper-converged product, with a grow as you go approach in 2U increments. And with a really easy setup and configuration: it can be up and running in 15 minutes!
One of the best quote is: “it looks simple because we make it simple”, Bipul Sinha. It’s really simple to use and with a fast interface.
The r300 Series is a Converged Data Management appliance that provides automated backup, instant recovery, unlimited replication, and data archival at infinite scale.
The Rubrik cluster should start from three appliances, each of them can have 12x4TB HDD or 12x8TB HDD, plus 4x400GB SSD, and can grow as you need. Deduplication is cluster-wide to improve space usage.
Actually it support VMware vSphere 5.1, 5.5 or 6.0 and you have only to point to a vCenter Server and use policy based rules (called SLA Domains) to protect your VMs. At this time only VMware platform is supported, but the entire architecture is platform agnostic, so could be possible see new virtualization platform supported in the future.
In order to provide application consistency, Rubrik push a VSS provider inside each VMs.
As show in the previous picture it’s possible replicate or simple increase the retention with an external object storage: S3 is supported, but for on-prem infrastructure (if needed) could be used S3 (for example with Scality) or Swift. Also NFS could be used for on-prem long term retention.
For site replication more Rubrik appliance could be used, both for DR purpose or Dev/Test cases.
And what about the data services, like the ability to have application-level recovery? Actually this is not included or integrated and it’s far from other products (like Veeam). Rubrik has a partnership with Kroll on Track to provide this services, but I think that simplicity and convergence mean also have those kind of services inside the solution itself and I hope to see something in the next releases.
Great review! I am happy to update that Reporting has gotten better in their newer code releases (3.0 & 4.0). In regards to the security scans, we only found issue with standard ports and configuration on the IPMIs, which was easily fixable. Have you found other security concerns that have been a bigger headache to deal with?