Home > Tintri VMstore™ > Knowledge Base > FAQ: Using Microsoft Windows Server 2012 R2 Deduplication on a VM Running on Tintri VMstore

FAQ: Using Microsoft Windows Server 2012 R2 Deduplication on a VM Running on Tintri VMstore

Applies To

  • Product(s):  VMstore
  • Product Version(s):  All
  • Bug(s):  N/A

Question

Should one use the Microsoft Windows® Server 2012 R2 Data Deduplication feature on a VM that is running on a Tintri VMstore?

Answer

 

These are guidelines to consider. Windows Server 2012 R2 Deduplication is post process and this makes it generate more IO load on a VMstore.  The process does the following (in this order):

  1. It writes all the data and the VMstore does all the work for inline deduplication and compression
  2. It again generates IO reads and writes to do the host-side deduplication and compression.

   Note:  Windows Server 2012 R2 deduplication does not deduplication across volumes/VMs. and enabling this feature has a minimal    effect on the compression rate for a VM on a Tintri VMstore.

Requirements for Windows Server 2012 R2 Deduplication feature

  • Data Deduplication is supported only on the following:
    • Windows Server operating systems beginning with Windows Server 2012
    • NTFS data volumes
    • Cluster shared volume file system (CSVFS) supporting virtual desktop infrastructure (VDI) workloads beginning with Windows Server 2012 R2
  • Deduplication is not supported on:
    • System or boot volumes
    • Remote mapped or remote mounted drives
    • Cluster shared volume file system (CSVFS) for non-VDI workloads or any workloads on Windows Server 2012
  • Files approaching or larger than 1 TB in size.
  • Volumes approaching or larger than 64 TB in size.
  • Deduplication skips over the following files:
    • System-state files
    • Encrypted files
    • Files with extended attributes
    • Files whose size is less than 32 KB
    • Reparse points (that are not Data Deduplication reparse points)

Recommendations

  • Tintri All Flash Arrays T5000s

It is recommended that you turn off Windows deduplication/compression because the VMstore already does inline deduplication and compression that is global across all data on the VMstore.

  • Hybrids

To get space savings from deduplication and/or compression, you may want to explore Windows deduplication/compression. But do consider the limitations details in the See Also section. 

  • T800s

Tintri does not have the ability for deduplication on HDD, therefore there is no space savings from block deduplication.

However, Tintri does have compression on HDD, and we support deduplication from cloning. If the data does not deduplication much, but compresses, Tintri recommends not using Windows deduplication/compression to eliminate the extra IO load of the post process host-side work because it will not obtain any additional space savings. 

  • T600s, T540

Tintri does not do deduplication or compression on HDD; therefore, Windows deduplication/compression may help reduce the capacity footprint on a Tintri VMstore, but at the cost of extra IOs.

See Also

Viewing 1 of 1 comments: view all
revised naming conventions for Data Deduplication and Windows versioning
Links were tested and work correctly.
Posted 23:53, 9 May 2017
Viewing 1 of 1 comments: view all
You must to post a comment.
Last modified

Tags

Classifications

This page has no classifications.