RCU 3.0 - VMware vCenter plug-in with Hardware Accelerated VM Clones and Redeploy

Eric Forgette, a member of our technical staff, has produced three new videos demonstrating features and functionalities of the RCU 3.0 vCenter plug-in. These new demo videos cover:

• Datastore Provisioning and Management
• Creating Hardware Accelerated VM Clones
• Hardware Accelerated VM Redeploy

You can click on a video in order to watch it in HD directly from Youtube

Video 1: Datastore Provisioning and Management

• Provision a new 100GB VMFS based datastore over FCP
• Provision a new 100GB NFS based datastore
• Grow a VMFS datastore
• Shrink a NFS datastore
• Review Deduplication savings
• Destroy a datastore

Video 2: Creating Hardware Accelerated VM Clones

• Create 6 new virtual machine clones
• Apply a sysprep.ini based Guest Customization
• Place them into an Active Directory Organizational Unit named Desktops
• Place the Virtual Machines into a VM folder named DesktopDemo
• Power on the VMs
• Complete in 1 minute and 20 seconds using almost no additional storage capacity

Video 3: Hardware Accelerated VM Redeploy

• Re-deploy 6 virtual machines
• Apply a sysprep.ini based Guest Customization
• Return each virtual machine to its current state
• Complete in 55 seconds using almost no additional storage capacity

Comments

[www.facebook.com/profile.php?id=658313066]

Watched the demo - how does the benefit of "HW accelerated" cloning compare on NFS datastores vs FC SAN datastores (shown in the demo)?

thanks!

[Vaughn Stewart]

@Fletcher (aka Facebook)

With NFS the ESX/ESXi host has access to storage virtualization at the VM level.

With VMFS the ESX/ESXi host has access to storage virtualization at the LUN or datastore level.

When the RCU makes clones on VMFS datastores we create VMware clones and on the array we deduplicate the datastore. If we need a mass number of VMs, we then clone the VMFS LUN.

The result in terms time VMs on NFS are immediate where on VMFS it takes time; however, bulk deployments which require multiple datastores the time to clone a VMFS and NFS datastore is equal.

In others words, cloning 1000 VMs on NFS may take 10 minutes where on VMFS it may take 60 minutes. In both cases it is very fast and space efficient.

Does this make sense?

[www.facebook.com/profile.php?id=658313066]

Vaughn, I was all excited to try this and test the speed since the last version - but while adding my storage controller credentials I ran into the same issue as before - vFilers (aka multistore) do not seem to fully support the same API as vfiler0 (eg ssh is not supported on vFilers in ONTAP v 7.3.1.1)

Any idea in what ONTAP version the vFiler customers will get to play with the great tools like RCU?

thanks!

[Eric Forgette]

RCU supports the vFiler context in Data OnTap v7.3.3.
Cheers

[www.facebook.com/profile.php?id=658313066]

Looks like I have a real good reason to skip 7.3.2 and go right to 7.3.3!

thanks Eric!

[Greg]

What is used to maintain the link from the "gold image" to the clones for redeploy? I've had a few instances where my vdi clones "lost" their link to the template and redeploy wasn't an option?

GregB

[Vaughn Stewart]

@Greg thanks for the question (and the discussion over lunch the other day)

With NetApp FlexClones (aka HW accelerated clones) the parent or gold image can be a VM or a VM template. The resulting VM clones will be space efficient as the share blocks with the source and one another, yet they are independent from their source with the exception of redeploy.

With HW assisted redeploy, the RCU restricts the redeploy to be from an update to the parent or gold image originally used in cloning the VMs.

[Jordan 5]

very cool space! I love your taste and it looks like it'd be so much fun to poke around in there... I used to live in DE but am now in VA, but if I ever go back that way to visit family I'd love to see it in person :)

[nike shox]

What is used to maintain the link from the "gold image" to the clones for redeploy?