D-Link Forums

The Graveyard - Products No Longer Supported => D-Link Storage => DNS-323 => Topic started by: PhillB on May 21, 2009, 11:05:28 AM

Title: Volume_2 after rebuilding raid array
Post by: PhillB on May 21, 2009, 11:05:28 AM

Hi,

I had a DNS-323 configured as Raid1 with two Seagate 1TB drives.  All showed up as a single Volume_1 in Windows.

However, one drive failed so I replaced it with the same Seagate 1TB drive (refurb from Seagate under warantee) and rebuilt the Raid1 array.  Now I have two physical drives showing under hard drive info on the DNS-323 Info page.  One is Volume_1 Raid1 the other is Volume_2 JBOD.  The Volume_2 drive is only 3189MB.

Any idea why I now have two volumes showing up or how to get rid of it?  The Volume_1 size is identical to the size before the drive failed, that is 981858 MB.

Thanks for any thoughts,

Phill.

Title: Re: Volume_2 after rebuilding raid array
Post by: ECF on May 21, 2009, 11:52:04 AM

Are you sure the Drive is the same exact model? Somehow it seems to have 3189MB more space.The System creates a JBOD volume with the remaining space when creating a RAID1 Volume if the two drives are not the same size.

Title: Re: Volume_2 after rebuilding raid array
Post by: PhillB on May 21, 2009, 02:28:03 PM

Thanks for this.

Yes, the two drives are exactly the same model:
Seagate Barracuda 7200.11
1TB
ST31000340AS
P/N 9Bx158-303 

The only difference is the serial numbers between the two drives and that they have different firmware versions.

Odd then that one appears to be around 3GB larger.

Thanks,

Phill.

Title: Re: Volume_2 after rebuilding raid array
Post by: fordem on May 22, 2009, 06:19:01 AM

Quote from: PhillB on May 21, 2009, 02:28:03 PM

Thanks for this.

Yes, the two drives are exactly the same model:
Seagate Barracuda 7200.11
1TB
ST31000340AS
P/N 9Bx158-303 

The only difference is the serial numbers between the two drives and that they have different firmware versions.

Odd then that one appears to be around 3GB larger.

Thanks,

Phill.

It's not at all unusual.

In the early days of magnetic disks, the manufacturer would ship the disk with a defect list that outlined the defective blocks found during initialization and we would have to manually enter those bad blocks into a table to prevent the system from writing to them.  You could have a 5MB disk (stop laughing, my first disk was a 14" monster with a 5MB per platter capacity, one platter fixed, the other removable) with half a dozen or so bad blocks whilst another might have more or less.  Each block had a capacity of 512 bytes, so you'd lose a few KB or so - no big deal - but it was rare to find two disks with exactly the same capacity.

Modern disks have much greater capacities and it is now the custom for the manufacturer to reserve an area into which bad blocks are 'remapped', you no longer get a defect list, but the disk is smart enough to know that instead of writing to block that has been marked bad, it should instead write that data into another block in the reserved area.

It is of course entirely possible for the disk to also have a few GB less, and had this happened to you, your DNS-323 would have told you that the disk was too small to be used as a replacement - I've seen this happen, not with my DNS-323, but with an IBM server.

Title: Re: Volume_2 after rebuilding raid array
Post by: PhillB on May 23, 2009, 08:01:29 AM

Thanks for the explanation Fordem, makes sense.  I hadn't really thought about the fact that 3GB these days is only a small percentage of the total drive size.  I've been more used to low GB SCSI drives till recently so 3GB seemed a huge difference in size but its only 3% I suppose on a 1TB drive.

Cheers,

Phill.