Cohesity Secondary Storage – a look into VM Protection

Ethos Technology is allowing MirrorSphere to use one of their Proof of Concept Cohesity secondary storage units.

For those who have not heard of Cohesity, they are providing a “limitless data platform for secondary storage,” to quote them directly.

The Founder and current CEO, Mohit Aron of Cohesity was also a founding member of Nutanix and regarded as a pioneer of hyper-convergence. He also worked at Google from 2003 to 2007, where he helped design the Google File System.

Cohesity, started in June 2013 has had two funding rounds and in total have raised $70 million. All though still a small company, at least 25% of the employees have a Google pedigree.

The Cohesity Infographic explains nicely what Cohesity Secondary storage is aiming at becoming.

For many IT systems, companies run overly expensive platforms to support some relatively low brow requirements. Most IT teams will have systems for end users that do not require fast backend storage, and where IO is not the most important factor.

Where the Cohesity system helps, is due to a distributed system and, although you have to have at least three nodes and twice as much raw storage under a redundancy factor of 2, it is based on commodity hardware.

The stepsof enabling the protection of any or all Virtual Machines (VMs) that you might have in a VMware environment are as follows…

The Cohesity comes preinstalled and only requires some very rudimentary information to get the cluster setup and running.

The only requirement is to have Bonjour installed, and IPv6 enabled on your computer. A little tip, to help the discovery of the nodes, is to ensure that all other NICs were disabled.

To configure the 2U Cohesity units, all you need, apart from network connectivity, is a serial number from one of the four nodes in this Cohesity unit. The serial number is used in a URL to access the web-based Cluster Configuration.

With a few key networking details, like a range of IP addresses, the creation of a cluster took about 5 minutes to perform.

The Cohesity system performs some magic, and then the webpage automatically updates with an IP address and just a reminder of the administrator account name and password you entered.

Once logged into the Cohesity cluster, you see their main dashboard.

The dashboard provides basic statistics, like the health of the system, storage and protection jobs. The Performance pane allows you to click through and see more information.

Before we can go any further, a Partition is required to be provisioned. Creating a Partition is a wizard driven process.

Simply provide a Partition name.

Assign some Virtual IP Addresses (VIPs).

You can also select which Nodes that you wish to add to this Partition bearing in mind that Cohesity Secondary Storage always need three as a minimum.

As I am using a few VIPS, in my example, I can load balance across the nodes that are hosting this partition using DNS round robin. A VIP moves from a failed node for the duration of that failure.

Before we can configure a VM protection job and after the partition has been created, we require a View Box.

A View box defines operational aspects, like if deduplication should be enabled and whether that should occur in-line or post-write operation.

Another feature of a View Box is the ability to extend it to external targets, like Google Near-Line or AWS Glacier.

Before we can protect a Virtual Machines in VMware, we need to add a source, which would be the vCenter IP address of FQDN.

From the Main Dashboard, under the Protection Menu, select Sources.

By clicking on Register Source, it provides an easy step by step wizard.

Entering my vCenter FQDN or IP Address and credentials and click on Register Source is all that is needed. An option to use the Throttling option, which places small pauses with the API calls into vCenter, this is put in as an option to stop the Cohesity request potentially from overwhelming the vCenter server appliance.

The Cohesity system, quickly provides an overview of your environment, showing how many VMs are protected or unprotected.

Clicking on the Protect a Source button provides another wizard with four steps.

From here we can protect everything under a Data Centre of, all VMs on a host.

The Automatic Protection will automatically add new VMs.

For instance, if you are protecting everything on a particular host, if a new VM appears or created, that will be enabled within next Protection job.

You can create your own, policies, but Cohesity do provide three basic policies out of the box.

Each policy determines how often you run the protection job and how long that the protected data will remain.

Easy to set up, after setting the data and time, the only other two options that you need to consider, whether you want App-Consistent backups, for applications like SQL. The other that is on by default if you want the VM files indexed. Very handy if you want to recover a particular file.

A final review of the job.

Now it is possible to see how many protected VMs there are against the entire VM estate on that Source.

The first jobs can take a bit of time as they will all be full. The subsequent backups will use Change Block Tracking (CBR) and are much faster.

Cohesity provides a new clean UI for the Protection Jobs status, so providing a quick method of determining your protection status and any failed jobs.

Expand each job and you can view any current running or previous tasks. Here you can also drill down and see additional information.

While a protection job runs, the percentage completion of the jobs is displayed and further clicking on the job provides additional information

The Cohesity secondary storage that I am currently testing is an absolute breeze to use.

The web interface is well designed, responsive in performance and the Wizards make the entire system a delight to use.

The Cohesity secondary storage also provides some big data technology, as it can index pretty much anything placed on it.

Cohesity also offers a Test and Dev feature. This feature clones protected VMs on the Cohesity Secondary storage, so an another location on their storage. The operation then integrates with vCenter and mounts Data Stores to hosts and add the VMs.

As a new platform to the market, it is already very mature in the look. As with any new to the market platform, Cohesity secondary storage has been very well backed in the VC arena, so future features or enhanced will come quickly.

One features that might work well would be native CIFS shares with Active Directory integration. With this and then add cloud spill and their big data indexing, archiving with secure search could become a viable option.