You can make sure that the 146GB hot spare is associated with the RAID 1, by limiting the size of the RAID 1 to 146GB during the creation of the virtual disk (volume). Selection of the disk group is not based on the size of the drive, but the actual capacity on the drive that is configured for a virtual disk (volume).

So, say I want to create a single raid 5 array (disk group) mixing:

-2 x 300GB HDs

-4 x 146GB HDS

I create a disk group selecting all those disks, and then I create a single RAID 5 virtual disk ending up with a size of (6-1) x 146 GB = 730GB

Then I can assign an existing 146GB  global hot spare to that virtual disk

At the end of the day:

-if a 146GB physical HD will fail, the 146GB hot spare will take over

-if a 300GB pysical HD will fail, the 146GB hot spare will take over anyway, because only 146GB out of 300GB of it are used by the virtual disk.

Is it correct?

Thanks a lot

Guido

I'm not sure how exactly the MD3000 firmware looks at available hotspares;

- can the hotspare hold the data needing to be rebuilt

- is the hotspare the same size or larger than the failed drive

However, the part about "Then I can assign an existing 146GB  global hot spare to that virtual disk" isn't possible; hotspares on this system are global; you make it a hotspare and it'll take over for any physical disk that fails anywhere in the enclosure (or any daisy chained MD1000). You can't specify which virtual disk/disk group the hotspare is for.

Right, "global" hot-spare means "for every disk in every array".

That's OK for us.

Now I need to understand if a 146GB global-hot spare can take over a 300GB physical disk which is used only for 146GB...

Yes, a 146GB hot spare can take over for a physical disk in a disk group that has a combination of 300GB and 146GB disks.

You can verify this by looking in the Storage Array Profile ("All" tab and look for HOT SPARE COVERAGE).

Thank you!

Guido