Rebuild RAID on LS-WVL with Debian

# Last thing we have to do, get the RAID array back together.

# Power down the NAS
# Put back the drive you pulled before
# Start netcat on your target pc (Just to be sure)
# Power on the NAS
# In a few minutes, you should be able to login to the NAS (if not, check the bootlog in netcat)

# --------------------------
# Update: 17-05-2013
# If you need to reconstruct a drive with the same partition layout
# Login as root:
parted

# Disk selection in parted
# The LS-WVL has two drive bays: left (A) is /dev/sda and the right drive bay (B) is /dev/sdb
# Parted selects the lowest possible drive at start.
# First we need to get the partitions on the drive which contains your data. So if your data
# is on the right partition, run
select /dev/sdb
# If it's on the left partition, this drive is already selected. You can select it either way with
select /dev/sda
# You will get feedback like
# (parted) select /dev/sdb
# Using /dev/sdb

# Now we get the partition table
unit s              # Select sectors as unit, so you have the most exact way of replicating the drive
print               # Print the partition table

# Mine for example is this:

# (parted) print
# Model: ATA ST3000DM001-9YN1 (scsi)
# Disk /dev/sda: 5860533168s
# Sector size (logical/physical): 512B/4096B
# Partition Table: gpt              # <<< HERE IS YOUR PARTITION TABLE

# Number  Start      End          Size         File system  Name     Flags
#  1      2048s      2002943s     2000896s     ext3         primary
#  2      2002944s   12003327s    10000384s                 primary
#  3      12003328s  12005375s    2048s                     primary
#  4      12005376s  12007423s    2048s                     primary
#  5      12007424s  14008319s    2000896s                  primary
#  6      14008320s  5843758319s  5829750000s               primary

# Now select the other drive (empty drive)
select /dev/sdb                     # Select /dev/sda if the LEFT drive is the empty drive

# Create the partition table. The output of your previous 'print' command gave a partition label.
mklabel gpt

# Recreate every partition using mkpart, in the same order.
# mkpart [name] [start] [end]          # See here: http://www.gnu.org/software/parted/manual/html_node/mkpart.html
# For me, the commands were
mkpart primary 2048 2002943
mkpart primary 2002944 12003327
mkpart primary 12003328 12005375
mkpart primary 12005376 12007423
mkpart primary 12007424 14008319
mkpart primary 14008320 5843758319
# USE YOUR OWN VALUES. ONLY FOR DEMONSTRATION

# Now you can use 'print' again. Compare the two outputs. All three columns (Start, End and Size) should
# be EXACTLY the same. File System and Name will come later.
print

# If everything suits you, exit parted with
quit

# -------------------------------
# Rebuild RAID
# -------------------------------
# Log in as root, if you're not already are.

# You have 4 partitions: boot, root, data and swap.
# partition     device          drive
# boot          /dev/md0        /dev/sda1
# root          /dev/md1        /dev/sda2
# data          /dev/md2        /dev/sda6
# swap          /dev/md10       /dev/sda5
# (if you pulled/replaced the left drive, you will have sdbX instead of sdaX)

# Cross check this table with your own NAS:
cat /proc/mdstat

# Example output:
# md1 : active raid1 sda2[0]
#       4999156 blocks super 1.2 [2/1] [U_]

# md1           - device
# sda2          - drive
# [2/1]         - 1 of 2 drives are working in the RAID array

# Normally:
# sda1 -> sdb1
# sda2 -> sdb2
# etc etc...

mdadm --manage /dev/md0 --add /dev/sdb1

# /dev/md0      - the device we are managing
# /dev/sdb1     - the drive we are adding. Note that we are ADDing sdB1
# Do this for all 4 partitions:
# mdadm --manage /dev/md0 --add /dev/sdb1
# mdadm --manage /dev/md1 --add /dev/sdb2
# etc etc...

# You can check the progress of rebuilding the RAID array with:
cat /proc/mdstat

# For me, rebuilding 3TB took ~400 minutes.