I'm trying to understand the difference between an iSCSI and a network share. So I've got a NAS device on the network.

When i use a network share, the computer knows that the location is remote. While, when i have an iSCSI target connected, the computer thinks that the target is another Hard Disk that is physically connected to itself. kewl - i get all that.

What I don't understand is the next bit. When i connect the iSCSI target.. my computer sees the new Hard Disk and wants me to format it. er.... waa???????? format it? But its already formatted! and i have data on it already .. just connect to the share and have a look in there.

So that's what I don't understand. Why do we need to format the disk when it's already formatted? If i format it, what will happen with the size of the other stuff? and the existing data?

I understand that, when u create an iSCSI target, u specify the size. So what happens if i specify more size than is free? (eg. we have existing data, remember).....


FWIW, the NAS is a Synology DS211.


First, the word format is a holdover from ancient times. What it means these days is to initialize the file system (file systems are just data on the disk for keeping track of your files). Modern disks are permanently formatted, you can't actually change the on disk format as you could (and had to do) 20+ years ago.

The NAS device almost certainly runs some proprietary or Linux OS; so the file system is likely EXT3 or some other FS that Windows doesn't read/recognize.

When you attach the disk via iSCSI Windows doesn't recognize the file system and deems it necessary to format the disk. Also, the "disk" your attaching to doesn't necessarily have to be a phystical disk. It might be a file in the existing file system; it could be a reserved block section of the disk pool (ZFS can do this notably); or it might be some logical/physical disk(s).

NAS devices typically export file systems over IP; such as CIFS/SMB, AFS, or NFS. SAN devices typically export block storage over the network, such as FC/FCoE, iSCSI, or ATAoE.

  • To clarify, when he attaches via iSCSI, he has created a LUN inside the filesystem of the NAS. – SpacemanSpiff Apr 7 '11 at 3:01
  • @SpacemanSpiff - a LUN? .. Also, when u say windows doesn't recognize the FS ... and wants to format it .. does this mean it's really formatting the FS of the NAS or just a virtual 'fake' disk .. which could translate to a single file on the NAS of a particular size.. ?? – Pure.Krome Apr 7 '11 at 3:17
  • download.synology.com/download/ds/userguide/x11-Series/… specifically pages 52-54. Create an iSCSI LUN and attach it to a taget. Connect to the target. Format the target. Honestly I don't know why I spent 20 minutes looking for that... – Holocryptic Apr 7 '11 at 3:36
  • Cheers @Holocryptic - but as I said, I already know how to use and setup an iSCSI device (suggested, when I said that I believe I know the diff between a share and iSCSI). What I didn't understand is the relationship between them --- ie. can a share see the data in an iSCSI drive? etc... – Pure.Krome Apr 7 '11 at 3:49
  • I'm not sure what you're asking anymore. Specifically regarding if a share can see the data in an iSCSI drive. I'm pretty tired right now, but I can't wrap my head around that. The NAS device already has a file system on it. It's EXT4 according to the docs I saw. When you manage the device, you create a LUN (a place to store data) and attach a target to it. That target (and the associated LUN) for all intents and purposes is now a disk. You can do whatever you want to it. Including format it. If you're connecting to a target from (sounds like windows) and it's asking to format... – Holocryptic Apr 7 '11 at 4:04

They both provide access to storage over the network. The difference is that a standard network share (SMB, CIFS, AFS) typically share at the file level, whereas iSCSI shares storage at the block level.

You can think of iSCSI-attached storage as if you're adding another hard drive to your system.

Because iSCSI is raw storage, you will need to create a filesystem on it before using it.

When you create an iSCSI share, the NAS is likely creating a disk image on its filesystem that will be associated with the iSCSI target. All IO bound for this specific target will be constrained to within that disk image.

  • Damn, you beat me to it! – HTTP500 Apr 7 '11 at 3:01
  • Crap, both of you guys posted as I was typing... – Holocryptic Apr 7 '11 at 3:02
  • yep.. that I all understand (and said so in the OP) .. but what I didn't understand was this formatting and how that impacts existing data on the NAS ??? – Pure.Krome Apr 7 '11 at 3:18
  • @Pure.Krome - details of how the NAS associates network shares and iSCSI targets with disk will vary depending on each vendor's implementation. I would hope that it's just creating a disk image for the iSCSI target, but I can't be sure about how this specific NAS works. You really just need to read your vendor's documentation and/or talk with their support to verify their behavior. – EEAA Apr 7 '11 at 3:19

If you want to edit 100GB file in network file storage, when you click to open it, first, your machine downloads file from file server then opens it. You edit and when you save, it uploads the file.

When you use ISCSI you can open a part of a file , edit it and upload only the part you edited back to iscsi disk. You dont need to download whole file to open a portion of it.

That is what differs in file level and block level access.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.