Quantcast
Channel: File Services and Storage forum
Viewing all articles
Browse latest Browse all 13565

De-Duplication on 3TB of data behaving weird

$
0
0

Hi All,

So have been playing with the lovely new Storage Spaces and De-duplication features in Server 2012 and have come across some strange behaviour.  The test volume in question contains just over 3.2TB of data, stored as a mirrored Virtual Disk on a Storage Pool consisting of 4x 2TB sata drives.  The data is a mixture of user home folders containing anything and everything you can imagine.  

I enabled the de-duplication feature on the volume about 3 days ago, manually set the file age to 0 and then triggered a new optimization job using powershell.  Whilst the job was running I kept check on the process using the Get-DedupJob and Get-DedupStatus powershell commands.  The job is still running now and has taken about 3 days so far.  (The server is low on ram and this has caused some lovely paging issues, hence the massive run time.)

During the first 0-50% or so the savings reported kept climbing and gradually ended up as high as 1.4TB which is about what I was expecting (lots of duplicate audio/video files and documents)  

BUT - somewhere around 50-70% the savings have completely plummeted back to zero and started growing again. At 86% it now lists only 129GB saved.  Its almost as if it has forgotten about the earlier savings it had made, even though it still lists almost 100% of files optimized.

PS C:\Windows\system32> Get-DedupStatus

FreeSpace    SavedSpace   OptimizedFiles     InPolicyFiles      
---------    ----------   --------------     -------------     
7.06 TB      129.45 GB    114973             116706             

Can anyone explain this behaviour?  Was it simply a rounding error and should of displayed GB instead of TB?  (I highly doubt this)  Is there something else going on that I should know off?

My next step will be to un-optimize the data and then optimize it again, setting up a script to log the output of get-dedupstatus every hour and see if I can log and reproduce the same issue.

I have had similar results with other smaller data sets where the savings climb up very well to around the 50% mark, then seemingly disappear!




Viewing all articles
Browse latest Browse all 13565

Trending Articles



<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>