If I understand what you are trying to achieve here, it is neither simple nor obvious. And I would not try it myself.
- In Unix/Linux users have names, but every user also have an ID number. The filesystem (files and directories) contain the user ID (not the name) of the owner of the given file / directory.
When you create a user on the DNS it gets a username/ userID. Files created using that user will have their 'owner ID' set to the 'user ID' of that user.
When you hook up the drive to a Linux box and copy some files there, the 'owner ID' of those files will be set by the Linux box to the 'user ID' of the Linux user you are using on that box. Which - numerically - might or might not be the same as the userID of some user used by the DNS. Probably will not be the same. Hence, when you plug the drive back to the DNS, you end up with a bunch of files / directories owned by an 'unknown user'. Not good (TM).
- The owner of the file(s) can also specify 'permissions', i.e. if he wants to let other users to be able to read and/or write certain files / directories. You could create the files as anybody and then set the permissions to let anyone to be able to read and write the files. (See Linux command chmod)
- If you know your Linux, you could even go back and change the owner ID of those files. Yes, there are Linux commands to do that. (chown) It is easiest doing these as root - see the fun-plug system.
- All this is Linux/Unix user management and filesystem stuff, not specific to the DNS. You would face the same issues would you move a HDD between any two Linux boxes.
- Add into the picture that you are not even going to access these files as a native Linux user on the DNS, but via a file server (Samba) running on the DNS, which introduces its own user management / user translation and file access management mechanisms...
- I think you are better off biting the bullet and copy the files slowly.
Side notes:
- If you have 2x1TB data and you don't want to use mirroring, you don't need 2 DNS-321. You could just put 2 drives into a single DNS, configure them as separate drives (no raid/mirroring).
- If you have 2 TB valuable data that you are concerned about to lose, you might be better off using some mirroring to defend against drive failures. Then you will need 2 DNS. You put two fresh, empty, clean, IDENTICAL (!) 1 TB drives into one, set them up as a mirror, load it with the 1TB data. (Will take days, probably best to do it in chunks. Just creating the 1TB mirror will take an overnight anyway. For me it took an overnight to load 300GB. Copy the data, and verify before deleting the original. You can sleep, work, meet girlfriend, have life while the machine works.) Then you repeat this with the other DNS and the other 1TB data.
- I don't know if the DNS supports mirroring 2TB drives, but you might be better off using 2Tb drives if you can. The price difference between the 1TB and 2TB drives would be partially made up by the difference between buying 1 or 2 DNS-321. (Now I am ruining D-Link's business because I am telling you to buy more drives instead of buying more DNS boxes...) Of course, if you already have the two DNS boxes, your options are limited.
- If you value your data, you should consider mirroring. Yes, it will cost you quite a bit more in the form of having to buy more drives. Without mirroring you can get away just buying a single new drive, with mirroring you have to get 4 new drives. But when one of your drives 'go', you will appreciate that you have the mirroring.
- You can NOT put a drive into a DNS-321, load it with data and THEN add another drive an setup mirroring. You have to first add both drives, create the mirror, before you load any data. Sorry. The DNS-321 probably uses the 'md' mechanism provided by the Linux kernel (I did not verify this, just speculating) to create that mirror. This limitation comes from that, so don't curse the D-Link guys. They just use what is available for them.