Select Page

Rubrik – Cloud Cluster

Rubrik – Cloud Cluster

As announced in April, Rubrik is no longer available in only its initial, physical form. You can now run Rubrik on both Amazon Web Services and Azure, providing the same functionality and simplicity for your workloads running in a public cloud. As Eric Shanks wrote in his article here, a huge benefit of this is reducing the requirements for that backup data to traverse your VPN or DirectConnect. Additionally, with the addition of Rubrik Edge, customers who wish to maintain a minimal footprint for their on-prem data center can still reap the benefits of the simplicity of Rubrik.

Similar to a physical appliance, the Cloud Cluster consists of multiple (minimum of 4) nodes which run as Amazon EC2 instances within your VPC.  The node sizes are suggested to be m4.xlarge, which include 4 vCPU, 16GB of memory with 750 Mbps of dedicated EBS Bandwidth. And again, just like the physical appliance, each node is preconfigured with a 400Gb SSD (gp2) for cache along with 3 additional 1TB volumes (st1) for storage.

The deployment of the Cloud Cluster is fairly straight forward and should be second nature to anybody familiar with launching EC2 instances. In short, you select the custom Rubrik AMI within your account (Rubrik must give you access to it) and launch the instances from it. There is, of course, a bit of prework required before launching, including VPC design and configuration along with proper networking and security groups in place. Once launched, log into the first node via SSH and run the user configuration. This sets critical information for the cluster such as the name, IP address of each node, NTP, DNS, and more. In my experience, the deployment takes 30-45 minutes to complete. Once deployed, you can log into the console and manage the appliance.

The only way to backup data, in version 3.2, is the use of agents on your EC2 instances. There is both a Windows and Linux agent which can be downloaded directly from the console and installed on your instances. Due to the nature of using agents, don’t forget to modify the security groups associated with your instances to allow the appropriate ports (12800,12801) inbound from the Rubrik cluster.

A couple of AWS features that I believe might be beneficial to a Rubrik Cloud Cluster include an EC2 Placement Group for low latency, 20GBps connectivity between the nodes or possibly launching larger EBS-optimized nodes to take advantage of a larger amount of dedicated EBS bandwidth (of course there’s a cost associated with such). In addition, choosing an S3 bucket within the same region should provide higher throughput for archival data versus choosing a different region. Use the built-in Cross-Region Replication if your requirements dictate you keep the data in an additional AWS region.

Last but not least, keep in mind the pricing of running such a beefy cluster. As mentioned above, the nodes are of instance size m4.xlarge, each of which can run as much as $0.20/hr using On-Demand pricing or as little as $0.075/hr using 3yr All-Upfront Reserved Pricing (us-east-1 pricing). That does not, of course, include the cost of the 400Gb gp2 volume along with additional 3 data volumes per node (st1). If you need to calculate or estimate costs, head on over to the free calculator that AWS provides to check out what you might spend on such a cluster.

All in all, this is pretty exciting for customers that want the simplicity of a Rubrik solution without having to manage a physical footprint to protect their data in the cloud. I can’t wait try out all the new ‘snazzy’ features will be added in the 4.0 release. Rubrik is a sponsor of Cloud Field Day 2 and I can’t wait to hear from them in person.

 

 

About The Author

Bryan Krausen

Bryan Krausen is currently working as a Technical Architect with experience in a vast number of platforms. Bryan has been active within the VMware vExpert community for several years and is the leader of the Louisville VMware User Group (VMUG) and Louisville AWS User Group.

Leave a reply

Your email address will not be published.