Marc Huppert

By: mhuppert | April 09, 2017

I got the opportunity to test the new Cohesity Platform Virtual Edition. I was the first guy in the world to install this product in my Home Lab and provide an online review. I am really honored for that. Thanks a lot Cohesity.

What is Cohesity and what is the Platform Virtual Edition?

Cohesity, a relative new startup from Silicon Valley has a mission: Hyperconverged Secondary Storage. Here is the official statement:

“Cohesity makes large organizations productive by consolidating, protecting and sharing your non-mission-critical data assets. Your essential data is instantly available when you need it, where you need it. Our ground-breaking distributed systems technology hyperconverges all secondary storage workloads into an efficient, agile and infinitely scalable resource pool. This greatly simplifies both your infrastructure and the resources to administer it.”

I first recognized Cohesity two years ago, when I thought it was a nice idea. Today it is a solution! An awesome one!

“Cohesity DataPlatform is the industry’s only hyperconverged platform designed to consolidate all secondary storage and data services at web-scale. Consolidate backup data, files and objects on distributed storage with best-in-class space efficiency and zero-cost copies. Converge your data services on Cohesity, including data protection, in-place search and analytics, and provisioning of test/dev copies. Build a data fabric that extends from the edge to the cloud and enable multicloud data and application mobility.”




They offer Secondary Storage, called DataPlatform with Global Dedupe, Compression and Encryption. Basis for that is an Intel Whitebox Server System with 4 Nodes within a 2U HE chassis. 3 Nodes are minimum required. They are not using a hypervisor, they installed their own Operating System and Cohesity Distributed File System. The system can scale to any capacity and infinite number of nodes.





What if I wanted to start smaller? Maybe in a ROBO environment?


The solution for that is the Virtual Edition:

“DataPlatform Virtual Edition (VE) is DataPlatform delivered as a virtual appliance. Simplify data protection and consolidate secondary storage in a simple and cost-effective virtualized form factor for remote and branch offices (ROBO). Manage remote offices through the same UI, APIs, and policy-driven automation used in central datacenters.”



The Virtual Edition has the same features like their big brothers, but comes as an OVA file.

It will be possible to use it for example inside remote offices and replicate it to the main Datacenter, where a Hardware Cohestiy DataPlatform is waiting for the data. It also can replicate into the Cloud.


Besides it is a perfect solution for our HomeLabs…..

It comes at two different layouts: Demonstration and Production

Here are the requirements for the Demonstration deployment:

4 vCPUs, 16GB RAM, 50GB SSD and 1TB HDD….


If you want to unleash the full power, you need to deploy the Production version:

16 vCPUs, 64GB RAM, 400GB SSD and 8TB HDD…



I deployed first the Demo version at my HomeLab to play with it. I made several tweaks to the virtual appliance to increase the performance. I had the opportunity to present my changes and feature requests to the Product Management Team at Cohesity, their feedback was amazing.


I started with the environment and really loved it from the beginning. The UI is so nice, the performance awesome and the features insane.


I decided to buy a new dedicated server just for the Cohesity DataPlatform Virtual Edition.



I ordered a HPe DL160Gen8 with the following specification:


2 x Intel Xeon E5-2650L (16 Cores and 32 Threads)

128 GB RAM DDR3-1600MHz

SmartArray P420 with 1GB Cache

2 x SanDisk X400M 1TB SSDs

2 x Seagate 8TB HDDs

1 x SanDisk FusionIO ioDrive2 1,2TB PCIe Flash

1 x 10GbE LOM Card

After installing ESXi 6.5 and the final configuration I deployed the Cohesity Virtual Edition in Production size on it.


My DataPlatform Virtual Edition has the following configuration:


4 vCPUs

64GB RAM with 100% Reservation

1st pvscsi controller with 62GB OS disk

2nd pvscsi controller with 400GB MetaData vmdk on SanDisk ioDrive2 PCIe Flash

3rd pvscsi controller with 7TB Data vmdk Eager-Zeroed-Thick (the creation took a while….)

vmxnet3 virtual NIC



It is not possible to add more than 2 vmdks to the appliance. One is used for MetaData information, this vmdk must reside on flash, the other one for data residing on magnetic spindles.


I configured the DataPlatform Virtual Edition to fit into my HomeLab environment.

·  DNS

·  Domain Name

·  NTP

·  SMTP Server

·  Cluster Name

·  License Key


First you need to create a partition within the system. A partition needs an internal name and a Hostname:

After that a View Box must be created with several options available:

I enabled Deduplication (Inline) together with Compression (Inline)



The third step (optional) is to create a View, that is a share via CIFS and/ or NFS. I created a View for my Veeam Backup Server. This share from the Cohesity Virtual Edition is configured as a backup repository within Veeam Backup and Replication. Cohesity has a very nice whitepaper, where the required steps are described in detail.

Cohesity can use its own DataProtect to protect almost anything. But at the same time Cohesity is able to expose SMB/CIFS shares used for many use cases like:  MS SQL dumps, Oracle RMAN, Log dumps or even be used as target storage for another backup application like Veeam. Cohesity is built with Global De-Dupe and Compression so it will save a lot of space for the Veeam backups that are only de-duping on the individual job level.

You can use any of your existing backup solutions together with Cohesity, all benefits from the global deduplication, compression, encryption and replication functions.



I wanted to use the build in DataProtection feature of my new storage/backup solution.



I registered my vCenter Server Appliance running in HA mode and was able to create the first virtual machine backup job.

From the vCenter Server repository I registered my Microsoft SQL 2014 VM as a SQL Server.

I entered the required Windows Credentials. No agent of something else needs to be installed onto the SQL VM.

Now it was time for the first Protection Job, there are several options, depending on what you would like to protect:

Let`s start with a MS SQL Server Protection Job:

The system uses the VMware VADP API to do VM and SQL backups. The options are defined within the Protection Policies. There are pre-configured policies in place, but you can also configure your own.

You need to enter a Job name and you have the ability to configure advanced job options.

Choose your SQL Server VM (registered as a SQL Server before) and add it to your Cart.

That´s it. You have successfully configured a MS SQL Server Protection Job. It automatically starts the backup process for this protected VM. From now on, your SQL Server gets backed-up every hour and the Transaction logs every 15 minutes... (depends on your policy).



You can check the status of your protection job at the Manage Protection Jobs Page:

I also configured several Virtual Machine Protection Jobs with different options.

I tweaked the system and I was really impressed by the performance of this little beast…

Really awesome is the Data Reduction / Storage information:

64,97 TB Logical are using within my Virtual Edition only 844,32 GB of real space!!!



The best thing is the Recover Option! Simply choose Recover from the main page:

What would you like to recover?

Let´s recover an individual file from a Virtual Machine Protection Job:

Simply search for the file….

Full text indexing is configured within each Protection Policy (You can disable it for sensitive information)


How freaking amazing is that???


You can Recover to Server or simply Download the File…

It is absolutely simple to restore a MS SQL Database:

Or a virtual machine:

The virtual machine is powered on running on a NFS share that is presented to the ESXi Server. Instant VM Recovery…. and then migrated using Storage vMotion onto the choosed datastore.


In Addition, you can backup physical machines, Windows or Linux. I tested it with my wife´s notebook running Windows 10. Absolutely no issues, works like a charm.


But you can even do more….

You can configure External Targets, like Cloud Storage, NAS Systems or Tape Libraries….

I setup my All Flash Custom Build NAS System as an external target and configured the archive function:

Moreover, you can Clone a VM or a View within the Cohesity Platform:

You can also create reports about your Backups:

The official boot logo from the Virtual Edition is “One Platform, Infinite Possibilities”


That is absolutely true!!!! It is an awesome solution. I really love it.



Test it out by yourself, when available. It will be worth it...

Category: Cohesity 

Tags: VCDX, vExpert 



Posted on : December 19, 2017

I like to read all the news which you are sharing in this blog. Always
give us such a good quality post. Thank you so much for this post.Your
mode of telling all in this article is actually good, every one be
capable of simply understand it. I read your post and I found it as very
write my assignment for me informative like
[url=]cheap essay[/url]writing
service.I have visited and I got good writing style. Thanks a lot.<br />

Post a Comment