LAST UPDATED: August 8 2013
This was an easy and fun task to complete and it was only slightly more complex than enabling Shadow Copy… so you can do this!
Things you need to know first about Data Deduplication in Server 2012:
- If you have other dedup software, hardware or deduplication in your backup you will want to turn those off (or not use Windows Server dedup). Like antivirus, no good comes from running more than one of them at the same time.
- Deduplication should only be run on file servers. Specifically, it should not be run on Exchange, SQL, or other servers with a lot of complex file locking or databases.
- Deduplication is built into Windows Server 2012 NTFS and NOT the new ReFS. Resiliant File System will have dedup in the future, but does not have it in the year 2013.
- Deduplication occurs at the volume level
- Deduplication is not possible on your boot disk, so you need at least two volumes (one for Server 2012 and one for file shares, which can be dedup’d)
- Deduplication functions without problem on HyperV VM’s running Server 2012
- Deduplication is availabe in both Data Center and Standard versions of Server 2012
- Deduplicaiton is moderately CPU intensive so you want to schedule it to run off hours and not when your backups are running
Now on to the fun stuff. In my case, I built my first dedup’er in a Server 2012 HyperV VM on and Server 2012 Host.
- Get your self two volumes. In my case this meant I needed to use DISK MANAGEMENT to reduce the size of my single partition. This provided enough empty space for me to create a second partition and format with NTFS… which I will set to dedup!
- Copy a pile-o-data onto this new partition
- Open SERVER MANAGER and click the MANAGE button at the top right of the window
- Select ADD ROLES AND FEATURES
- Select ROLE-BASED OR FEATURE-BASES INSTALLATION
- Select your server (duh!)
- Expand FILE AND STORAGE SERVICES > FILE AND SCSI SERVICES
- Select DATA DEDUPLICATION
- Finish the wizard as is obvious
- Open SERVER MANAGER and select FILE AND STORAGE SERVICES
- Click VOLUMES
- Right click on the volume in question and select CONFIGURE DATA DEDUPLICATION
- For testing purposes set the DEDUPLICATE FILES OLDER THAN: to zero days. If you don’t do this you will have to wait the number of days specified for the the deduplication to run. In production this is a great idea but for testing… I wanna see it NOW!!!
- Click SET DEDUPLICATION SCHEDULE
- Click on ENABLE THROUGHPUT OPTIMIZATION and set a start time of about now
If you can’t wait for the time you set, just dig into the TASK SCHEDULER and start it. The task will be under:
- TASK SCHEDULER LIBRARY
In my case I saw 4GB of Office type documents (PowerPoint, Word…) change to 16MB which is a 256 times reduction… this is obviously not correct. What is happening is the duplicated bits are being relocated to the SYSTEM VOLUME INFORMATION folder so Windows Explorer does not see (or count) them.
The DEDUPLICATION RATE shows in VOLUMES, displayed only 48% which is the accurate number including the bits from the SYSTEM VOLUME INFORMATION folder.
If you are going to run data dedupulication (which you should), you really need to at least skim these two articles
If you want more information, I found the following quite useful: