Understanding How the LinkStation Duo Raid 1 Works

[depeche]

Hi, I searched, but could not find this fully addressed on the forum.  I just purchased the LinkStation Duo LS-WX4 Firmware 1.6 and set it up with Raid 1.
I consider myself somewhat computer literate, but the Raid process is new to me, so I would like to summerize my understanding, so I don't have any surprises in the future.
Please let me know if I am correct with all of this.  I have backed up my data to one of the drives.  This drive functions as a normal external drive.  Raid 1 is automatically
backing up/mirroring that data to the second drive.  That second drive is XFS file format.  There is no way to view or access this mirrored drive, so the only way to know if
something is wrong is with an error code the NASNavigator2 will give me or with an e-mail notification from the LinkStation.  Is this correct?
If the drive with the XFS mirror fails, all I need to do is replace it and the data on the good drive will automatically be backed up/mirrored again on the new drive.
What happens if the drive that is functioning as a normal external driver fails?  I assume I will still not be able to view my data on the mirrired drive, since it is in XFS format.
Once I replace that failed drive, will the LinkStation then copy the data from the mirrored XFS drive back to the new drive, where I can view it like a normal external drive again?
Thanks

[depeche]

Thanks, I appreciate the clarification.

[PaulDriver]

First, all drives are configured with a GPT partition table, then they are allocated as various pieces of raw disk to MD, the Linux Multi Disk Software RAID Array System via the mdadm command.

(under Ubuntu enter sudo apt-get install mdadm xfs xfsprogs in a terminal to work with the Linkstation drives)

 

 

If the Linkstation created the drive from a bare partitionless drive, then the layout is as follows.

if, on the otherhand, you do, like I did, and put a drive with an GPT partition table, you will see a single empty partition.

 

 

In a Raid1 box. the raw partitions on /dev/sdx will be laid out like this:

/dev/sdx raw GPT partition 1, 1.0 GB@ /dev/md0.0 EFS4 Filesystem

/dev/sdx raw GPT partition 2, 5.0 GB @ /dev/md1 (don't remember the filesystem type)

/dev/sdx raw GPT partition 3, 1.0 MB @ n/a

/dev/sdx raw GPT partition 4, 1.0 MB @ n/a

/dev/sdx raw GPT partition 5, 1.0 GB @ /dev/md10 (don't remember the filesystem type)

/dev/sdx raw GPT partition 6, Remainder of drive @ /dev/md22 XFS filesystem

 

Partition 6 is the peice that most of you are going to be interested in, it's where the money (data) is.

 

In a bigger box. partition 6 is a chunk that is used as part of the array, and gets labeled differently

(AFAIK, I have no emperical knowledge of a Buffalo Raid5 box)

 

Filesystems are laid down ON TOP of the MD Software RAID subsystem, and are not accessable without MD running.

 

(edits:fixed typos, grammer and the calling the xfsprogs package as xfs-tools )