Mounting NTFS data

[andypm1982]

Hi there.

I have an issue where I have a 12disk array on a fresh FreeNAS install and 4 sas disks that contain my NTFS data that I backed up before installing FreeNAS. What’s the recommended way to move it all back.

I’m trying to avoid a network transfer as it will require me buying another HBA and sas cables for one of my other systems.

thanks.

 

[Samuel Tai]

If these are straight NTFS simple volumes, and not part of a stripe/mirror/spanned volume, then you could use the import procedure in section 10.6 of the Guide, assuming you have slots available on your backplane. If they are part of a larger volume, then you'll have to transfer the data over the network.

 

[andypm1982]

I did try this but how do I access the data once the disk is mounted??

thanks.

 

[Samuel Tai]

The import process copies everything to a new dataset under your data zpool, and then unmounts the import disk when complete.

 

[andypm1982]

So the contents of the NTFS disk will still be safe on the NTFS disk?

 

[Samuel Tai]

If the Guide is correct, yes. I also took a look at the code for the import process in /usr/local/lib/python3.7/site-packages/middlewared/plugins/pool.py. The relevant subroutine is import_disk, at line 1990. Within this procedure, there's first a read-only mount of the disk to be imported.

 

[andypm1982]

yes I appear to have figured it out although it seems to be a slow process but I imagine that's just because its a single disk to an array on the same supermicro backplane, I have found before that having multiple arrays on the same backplane tends to make things drag when copying between arrays.

im planning on moving to a dell r730xd I have acquired and expanding to an md1200 for an additional 12 disks once im happy with how freenas works, its been a good few years since I toyed with freenas and back then it was a couple of disks connected directly to the mainboard.

 

[andypm1982]

I do have one question, do I need to keep sync enabled on the array?? I have a ups and have heard that sync slows down the performance of the array??

 

[Samuel Tai]

You can configure sync on a per-dataset/per-zvol basis. Unless you have a SSD-backed log vdev in your data zpool, I recommend leaving sync at the default setting of standard.

 

[andypm1982]

Is the import process a quick or slow process?? At the moment it’s taken an hour to do 100gb, that’s like 6 days of importing at that rate. Not sure if it’s related but services has chewed up over 52gb of memory leaving currently half a gb of free memory.

 

[Samuel Tai]

Unfortunately, imports are slow. Partially, that's due to the NTFS mount going through FUSE.