Introducing Rubrik

New to the backup protection and recovery space is a company called Rubrik.  Their platform promises a simplified way of protecting, recovering, and archiving VMWare virtual machines with an easy to use, highly scalable appliance.  Let’s look at the Rubrik r300 series now.

Simply put, this is a backup and restore appliance for VMWare virtual machines.  With this platform, you can quickly backup your VMs, store them in a deduped format, and be able to quickly create multiple restore points for your VMs based on your Service Level Agreement Policy, or SLA. To recover a VM, the Rubrik will act as your datastore and using a combination of the disk and SSD, you will be able to quickly recover a VM from a failure or just for testing in a matter of seconds.  VMs can reside on the local Rubrik platform for backup, a remote Rubrik cluster, or an archive destination such as AWS.


Hardware-wise, the Rubrik platform is a converged storage and compute platform.  Each cluster, or “Brik” consists of 4 Nodes in a 2U enclosure.  Each node is identical in hardware, side from the hard disk size.  Currently, there are 2 models.  The r344 and the r348. 




 
r344 per node (4)
r348 per node (4)
3x 4TB HDD
3x 8TB HDD
400GB SSD
400GB SSD
Intel 8 Core 2.4Ghz Haswell
Intel 8 Core 2.4Ghz Haswell
2x 10Gbps SPF+ Ports For Data
2x 10Gbps SPF+ Ports For Data
2x 1Gbps Copper for iSCSI Transport
2x 1Gbps Copper for iSCSI Transport
1x 1Gbps for IPMI traffic
1x 1Gbps for IPMI traffic

As you can see, the only difference between the two nodes are that the r348 has twice the available storage.  Briks can be added to the Rubrik cluster as needed to increase available storage and available nodes. 

After the cluster is installed, cabled, powered, and initialized, you can bring up the interface on any HTML5 browser.  Chrome works well for this purpose.  Logging in, you will be presented with the Rubrik dashboard.



Here, we can add our vCenters, picks are VMs, and assign them to a SLA.  Let’s look into how we can protect a VM.

I’m going to assume you already added a vCenter to your environment and we are going to use one of the built in SLAs. 

We are going to first click on VM Protection and then find the VM that we want to protect.  Type in the name of the VM in the search Window and it will predictively find the VM for us.  We’ll check the box by the VM and then hit the “Assign SLA” to give the VM a SLA assignment.  We can select single VMs, multiple VMs, or pick by Folder or Cluster/Hosts.

And that’s it.  Soon, your VM will be backed up to the Rubrik Cluster.  When the VM is backed up to the Rubrik and how often is based on the SLA.  Let’s look at the SLA a bit more to get a better understanding.


Selecting SLA Domains and then Local Domains, you will see the domains listed on your local Rubrik cluster, which you will be able to manage. By default, you will have 3 SLAs created from Rubrik, Gold, Silver and Bronze.  You can modify these as well as create your own SLAs.  The properties you can give to a SLA are how many snapshots to take every hour, day, month, and year, how long to keep these snapshots, should it replicate it to a remote Rubrik cluster, or should it archive to AWS.  Let’s look at the built in plans and discuss how they differ.

Our current cluster is just a single cluster with no replication or AWS setup, so we will not go into those settings.
Bronze
Silver
Gold

Take
Keep
Take
Keep
Take
Keep
Every Hours (Keep for Days)


12
3
4
3
Every Day (Keep for Months)
1
1
1
1
1
1
Every Month (Keep for Years)
1
1
1
1
1
1
Every Year (Keep for Years)
1
2
1
2
1
2

So, here we can see the Bronze SLA will only take 1 snapshot a day.  These daily snapshots will last for one month.  The first one of the month, will be kept for one year, and the first one of the year will last for two years.  The Silver SLA will take a snapshot every 12 hours, keep for it 3 days, and so forth.  From here, we can see that the Gold plan will have significantly more restore points, but at cost of more storage.  Remember, that everything is deduped and incremental backups are done after the first full backup.

When the backup is copied to the Rubrik cluster, it will be triple mirrored to 3 separate nodes.  This means that you in a single Brik cluster, you can lose two nodes and still be able to completely recover all of the virtual machines on the Rubrik cluster. 

At this point, we have protected our virtual machines. Recovering our VMs are a simple process as well.  We will go over that process in a future blog.