Q

Is thick or thin a better format for vSphere Storage Motion?

An administrator can choose thick or thin provisioning on a destination volume for a data store, but which one offers better performance?

What formatting options are available when using Storage vMotion? How are they different?

When planning to move a VM from one data store to another, the administrator needs to select a format for the destination volume. A default option allows administrators to migrate a virtual machine file system (VMFS) or network file system (NFS) volume using the same format as the source volume, but administrators can also select a specific format -- thin provisioned or thick provisioned -- with Storage vMotion.

A thin-provisioned disk volume creates the full logical space, but only populates  it with a fraction of the actual storage capacity, adding extra storage to a VM when more is required. By reducing investment in upfront storage, thin provisioning may help administrators save money. However, the write performance of a thin-provisioned volume may be slower than other formats, because storage space is typically not prepared until it has been allocated and used.

By contrast, a destination VM volume can be thick provisioned, in which the entire space is allocated and prepared upfront. For example, if a 200 GB volume were thick provisioned, the entire 200 GB would be set aside immediately -- even if the entire space was not being used at that moment.

There is also a difference in the way that thick-provisioned space is initially zeroed or prepared. If the thick-provisioned volume is eager zeroed, all of the provisioned space is zeroed during the provisioning process., If the volume is large, this may take some time. However, because the space has already been prepared, the first subsequent write may be  faster, which would make it ideal for a storage migration. Of all the format types available, eager-zeroed volumes typically offer the best performance, and are often used with Microsoft Clustered Server (MSCS) and fault-tolerant VM volumes.

If the thick-provisioned volume is lazy zeroed, all of the provisioned space is allocated, but blocks are only zeroed before actual data is written. This makes the initial creation process somewhat faster, but the actual write performance may be lower than it is for eager-zeroed volumes. Lazy-zeroed volumes are often unused because they offer no particular benefit for the enterprise.

This was first published in May 2014

Dig deeper on vMotion and Storage vMotion

Have a question for an expert?

Please add a title for your question

Get answers from a TechTarget expert on whatever's puzzling you.

You will be able to add details on the next page.

0 comments

Oldest 

Forgot Password?

No problem! Submit your e-mail address below. We'll send you an email containing your password.

Your password has been sent to:

-ADS BY GOOGLE

SearchServerVirtualization

SearchVirtualDesktop

SearchDataCenter

SearchCloudComputing

Close