processing files on data deduplication enabled SMB share

Hello community!

We are using Matlab to process large numbers of JPG files stored in a folder structure on several SMB network shares. The file servers are running Windows Server 2012R2 and have the data deduplication feature enabled, ie. the servers search for duplicate blocks at nighttime and free up disk space by only storing identical blocks once. This process is supposed to be transparent to users and applications.

Now execution of the Matlab "dir" command on directories/files returns a list of files, but only if these files have just been written to the share and are thus not deduplicated yet. After a couple of days, executing the very same "dir" command on the very same directory does not output the file names anymore. However the files do exist, and opening them in any other program works fine. The only difference is that according to the Windows Explorer file properties dialog, the size on disk has become much smaller than the file size. That's why I suspect this to be connected to the server's deduplication feature. Strangely, the "dir" command works as expected if an asterisk is added to the path:

K>> dir('\\FILESERVER\Share\2015\20150629040500_0.jpg')
'\\FILESERVER\Share\2015\20150629040500_0.jpg' not found.
K>> dir('\\FILESERVER\Share\2015\20150629040500_0.jpg*')

20150629040500_0.jpg  

K>>

Also, the "open" command fails on these files.

Our IT guy did disable the deduplication feature for some folders, however this seems to be a very tedious task, and it still rather is a workaround than a solution to the problem.

We are using Matlab R2014a. Are there any known issues when using Matlab for processing data stored on SMB shares with data deduplication enabled? Do you have any suggestions on what can be done to resolve this? Thanks in advance for any hints on this!