Present CIFS volume as LUN for local access

[darraghos]

I am looking to migrate some of our file shares to SharePoint. I am looking to run the Jam Software Tree size tool against the shares to see the breakdown of files/directories etc. The problem is that the tooling is running on a Windows Server and access/trawling the shares over CIFS and is very slow given the amount of files/directories it has to example. The Jam tool when accessing a local NTFS file system however can use the $MFT table for the scan and is thus lightening fast.

So, is there any way to achieve this? If a took a snapshot of the CIFS volume could I present that out in read-only mode and attach as a LUN (block storage) to a Windows Server so I could access the $MFT ? I'm thinking not possible but a definitive answer would be great. 

[SpindleNinja · Accepted solution]

If i'm understanding this correctly.   You want to take a volume that's NTFS and being shared out as CIFS.  and map it to a Windows Host as a LUN?    

Not possible though, can't convert file to block of block to file. 

[parisi]

If you want the share to be a LUN, here's your steps:

1. Create a new volume and a LUN (16TB is your LUN size limit)

2. Map the LUN to the server

3. Copy the data from CIFS to the LUN - there's no way to convert to a LUN in place.

Keep in mind that if you do this, only that server gets fast, local-ish access to the LUN. Everyone else that accesses the data will need to use the Windows server as the connection point; you can't map a LUN to multiple hosts. That Windows server would need to act as an SMB server to that LUN.

If you're looking for faster scanning of the share, I'd suggest using something like XCP and NFS to scan.

And if you can, leverage file analytics to get near instant access to file info.

https://blog.netapp.com/xcp-data-migration-software

Also, the next release of ONTAP will have some on-box tools for this as well.

[darraghos]

Yuk! I'll have to make do with trawling it over CIFS (1TB of data)

[darraghos]

I guess I misunderstood how the data is stored at the NetApp end. For efficiency this is a filesystem on the cluster rather than a filesystem on a LUN presented to the clutter itself 🙂

[SpindleNinja]

SAN vs NAS can be looked at "who owns the data (blocks)".   

In the SAN world (iSCSI/FC/NVMe) the system hands out a chunk to a host and the host puts it's own file system on it, and does what it wants. 

NAS (CIFS/SMB/NFS/AFP) the system serving out the data owns the data.   

Any storage array does this,  not just NetApp.  The only thing that NetApp does kinda differently is that a "LUN" can kinda look like a "file"  because it's inside of a flex vol.   If that makes sense.   But people put multiple luns inside of a flexvol to act as a consistency group sometimes.  Or they want better storage effs.