Disk Spanning Vs JBOD

HI Everyone

I’ve got My Cloud Ex2 Ultra and now confused that which configuration should I use between JBOD or SPANNING. If I configure spanning and my one drive fails, am I able to access data available in other drive or not ?

1 Like

No, if you use spanning then you’ll lose everything on all of the drives involved, as bits of files can end up anywhere all across all the disks spanned.

JBOD (“just a bunch of disks”) works each disk independently, so if one drive fails (as opposed to the controller or something like that) then all the others should still be accessible.

if you use spanning then you’ll lose everything on all of the drives involved, as bits of files can end up anywhere all across all the disks spanned.

I thought this is RAID 0? Data are randomly written to any of disks.

For Spanning, it simply fills disk 1, then disk 2, disk 3…

Raid 0 should stripe the data across both, whereas spanning fills the disks on order as you say.

But in both cases you’ve got one virtual disk across multiple physical ones, and if the disks physically fail you lose everything in either case.

DarrenHill

Feb '18

Raid 0 should stripe the data across both, whereas spanning fills the disks on order as you say.

But in both cases you’ve got one virtual disk across multiple physical ones, and if the disks physically fail you lose everything in either case. DarrenHill

I refer to this post for my question.

  • Like sh2sg, I thought that “Spanning” mode was filling disk 1 then disk 2, disk 3… and then, in case that one drive fails, it would allow to access data available in other drive.
  • But DarrenHill seems to say that “Raid 0” and “Spanning” mode leads to the same result, which is the loss of everything (all data) in case one disk physically fail.
  • So my question is why one would choose “Spanning” mode since it is less performant and equally unsecured as “Raid 0”?
  • Would it be only because Spanning have no size limitation in case of use of different disk size while with Raid 0 the storage space added to the array by each disk is limited to the size of the smallest disk?
    Thank you for your help.

Spanning is a bit “pot luck” in the case of losing a disk.

As already noted, it basically daisy-chains the disks together, starts at one end of the chain and writes stuff to the first disk, then the second, then the third etc until the whole lot is full.

In theory if you or the OS never delete anything, then things would just “pile up” and if you lose a disk, you would only lose the files (or parts of files) which are on that failed disk. However in practice most OS’s will try and fill up the gaps when files are deleted, and so you can get parts or all of files stored in earlier disks than the “top of the pile” one. In that case if even part of the file is on the disk that fails, the file is gone.

In the case of RAID0, as different parts of the file are written in parallel across multiple disks (for the increase in speed that this gives), you will tend to lose everything as you’re almost certain to have bits of all the files missing (the bits that were on the failed disk).

Got it. Thanks DarrenHill