Now that we have a raid array created, we are going to create some virtual drives. Why virtual drives? Well, we have one large terabyte drive right now, and if we just put folders in there, it can quickly get cumbersome. You have no control on size of those folders, and permissions can be more difficult. And if you try to share this drive, you can’t assign different drive letters in windows to different folders, just one to the drive. One other issue is file system. I am going to use just a basic file system here, but in the past I have created different file systems based on the type of use ie. large video files, small text files, etc. Making those decisions is beyond the scope of this tutorial, and I have decided that for my purposes now, it doesn’t matter that much. You can decided differently.

What we will discuss is using LVM. This enables you to set up virtual drives that can contain different file systems, and that can be grown and shrunk (usually) to fit the space needs of the system. We will look at maintenance of these file systems at a later tutorial. Here we will create a backup, and a pictures virtual drive. We will not use the full terabyte of space, so that we can grow these as needed, or add another for say music at another time.

I will once again be doing this on an ubuntu system, but the use of these tools is fairly standard across linux distributions.

First we will need to make sure that you have LVM installed.

sudo apt-get install lvm2

We need to create what is called a physical volume. We need to tell LVM what actual physical drives are involved. Lets check what drives we have again.

$ sudo fdisk -l

Disk /dev/sda: 1000.2 GB, 1000204886016 bytes
255 heads, 63 sectors/track, 121601 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0x00000000

   Device Boot      Start         End      Blocks   Id  System
/dev/sda1               1      121601   976760001   fd  Linux raid autodetect

Disk /dev/sdb: 1000.2 GB, 1000204886016 bytes
255 heads, 63 sectors/track, 121601 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0xfe76bc93

   Device Boot      Start         End      Blocks   Id  System
/dev/sdb1               1      121601   976760001   fd  Linux raid autodetect

This is just a partial list of the drives. See that the /dev/sda1 and /dev/sdb1 drives have a system type of fd. In the last part of teh tutorial we set up software raid. If you are NOT using raid, you will need to add these devices to your physical volume. For us, we are going to add /dev/md0 that we created last time. Lets just verify that the array is there.

$ sudo mdadm --detail --scan
ARRAY /dev/md0 level=raid1 num-devices=2 metadata=00.90 UUID=****

Note: all UUID’s will be replaced with **** in this doc. They will be different on every machine anyway.

As you can see, array /dev/md0 is ready to go. Lets create that physical volume finally.

sudo pvcreate /dev/md0

You can verify that it is created with the pvdisplay command.

$ sudo pvdisplay
  "/dev/md0" is a new physical volume of "931.51 GB"
  --- NEW Physical volume ---
  PV Name               /dev/md0
  VG Name
  PV Size               931.51 GB
  Allocatable           NO
  PE Size (KByte)       0
  Total PE              0
  Free PE               0
  Allocated PE          0
  PV UUID               ****

Now we need to create our volume group. That last step identified what hardware LVM has available. This next step groups that hardware into volume groups. I will call my “nas”.

sudo vgcreate nas /dev/md0

We can then verify our volume group.

$ sudo vgdisplay
  --- Volume group ---
  VG Name               nas
  System ID
  Format                lvm2
  Metadata Areas        1
  Metadata Sequence No  1
  VG Access             read/write
  VG Status             resizable
  MAX LV                0
  Cur LV                0
  Open LV               0
  Max PV                0
  Cur PV                1
  Act PV                1
  VG Size               931.51 GB
  PE Size               4.00 MB
  Total PE              238466
  Alloc PE / Size       0 / 0
  Free  PE / Size       238466 / 931.51 GB
  VG UUID               ****

Next, we can finally create our virtual drives. I am going to create a backup drive of 100 gig, and a pictures drive of 400 gig, both which are added to the “nas” volume group. You can create drives of what ever size you would like.

sudo lvcreate --name backup --size 100G nas
sudo lvcreate --name pictures --size 400G nas

You can see information about these virtual drives with another display command. Guess what it is.

$ sudo lvdisplay
  --- Logical volume ---
  LV Name                /dev/nas/backup
  VG Name                nas
  LV UUID                gqrIVE-YtS4-cBne-QyPK-oMO1-7V7V-UaBoUL
  LV Write Access        read/write
  LV Status              available
  # open                 0
  LV Size                100.00 GB
  Current LE             25600
  Segments               1
  Allocation             inherit
  Read ahead sectors     auto
  - currently set to     256
  Block device           254:0

  --- Logical volume ---
  LV Name                /dev/nas/pictures
  VG Name                nas
  LV UUID                oDxJJi-lxsG-9RHw-RBME-oJah-jOJJ-J8I4p3
  LV Write Access        read/write
  LV Status              available
  # open                 0
  LV Size                400.00 GB
  Current LE             102400
  Segments               1
  Allocation             inherit
  Read ahead sectors     auto
  - currently set to     256
  Block device           254:1

Now that we have our virtual drives, can we start filling them up? Not quite. We need partitions on them. These drives are just like a bare drive from the factory. We can put whatever file system we want on them. I am just going to use ext3. This may not be the best file system out there, but it is proven, and works good enough for me.

sudo mkfs.ext3 /dev/nas/backup
sudo mkfs.ext3 /dev/nas/pictures

That takes a little bit of time, just as would without LVM. As far as your programs, and bash, and Gnome are concerned, /dev/nas/backup is just another drive. It does not care, and does no know that underneath all this is a raid array and virtual drives.

Now we need to get them mounted so that they are useable. First off, I create a couple of mount points. I like to keep them organized.

sudo mkdir /var/nas
sudo mkdir /var/nas/backup /var/nas/pictures

sudo mount /dev/nas/backup /var/nas/backup
sudo mount /dev/nas/pictures /var/nas/pictures

Now, do a “df” to see the drives on the system and the space used.

$ df
Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/sdd1            186693520   2469660 174740292   2% /
/dev/mapper/nas-backup
                     103212320    192248  97777192   1% /var/nas/backup
/dev/mapper/nas-pictures
                     412849328    203156 391674652   1% /var/nas/pictures

Note: I removed other system file systems that we are not talking about here.

You can see that our two drives are there. They are listed at /var/mapper/nas-backup and /var/mapper/nas-pictures, but they are the /dev/nas/backup and /dev/nas/pictures we have been setting up.

Now you can copy your data into these drives.

I am not the only one that uses these drives, and they will be shared on samba (later tutorial) so I like to change the permissions on these folders. I will create a “nas” group, add myself (insert your user id for *user*), and change the permissions.

sudo groupadd nas
sudo adduser *user* nas
sudo chown -R *user*:nas /var/nas/backup
sudo chown -R *user*:nas /var/nas/pictures

The file permissions are set up as writable by owner and readable by group and other. I will leave them this way for now.

Things look good to go, and they would for awhile, until you reboot. Then your drives would disappear. We need to make them a little more permanent. We do this by putting some entries in the /etc/fstab file. This is the list of all mountable file systems that the OS uses to connect things up when booting.

This file in Ubuntu is using UUID’s, so we will to. First thing you need to do is find out what yours are. So…

$ sudo blkid
/dev/sdc1: UUID="****1" SEC_TYPE="ext2" TYPE="ext3"
/dev/sdd1: UUID="****2" TYPE="ext3"
/dev/sdd5: TYPE="swap" UUID="****3"
/dev/sda1: UUID="****4" TYPE="mdraid"
/dev/sdb1: UUID="****5" TYPE="mdraid"
/dev/md0: UUID="****6" TYPE="lvm2pv"
/dev/mapper/nas-backup: UUID="****7" TYPE="ext3"
/dev/mapper/nas-pictures: UUID="****8" TYPE="ext3"

You can see the unique ID’s that are assigned to the drives in the array, the raid array its self, the LVM physical volume, and finally, the two logical volumes. Use these UUID’s that you get when you execute blkid. Now lets edit /etc/fstab

$ sudo nano -w /etc/fstab
# /etc/fstab: static file system information.
#
#     
       

proc            /proc           proc    defaults        0       0
# /dev/sdd1
UUID=****2 /               ext3    relatime,errors=remount-ro 0  $
# /dev/sdd5
UUID=****3 none            swap    sw              0       0
/dev/scd0       /media/cdrom0   udf,iso9660 user,noauto,exec,utf8 0       0

#/dev/nas/backup
UUID=****7 /var/nas/backup ext3    relatime,noexec 0 2

#/dev/nas/pictures
UUID=****8 /var/nas/pictures ext3  relatime,noexec 0 2

You shouldn’t have to change anything except add the lines at the bottom for the backup and pictures drives (or whatever you created) . Please use the UUID’s you found from blkid.

Press ctrl-x to exit nano and save the file.

Try a reboot and see how it works. Once you bring the system back up, open a terminal and type df. This will show you all mounted drives, and your new drives should be there.

We are assuming that at this point you have Ubuntu installed. We are gong to install software raid in a moment. The first thing I want to do is make sure we have a static ip address. Usually Ubuntu will start up with a dynamic ip address. This can make it more difficult to transfer files to this box, so lets make it static.

Setting a static IP address

Type ifconfig eth0, and you will see your current address.

Lets make it static. Note, typing sudo in front of commands lets you assume the privileges of the administrator, or super user.

sudo nano /etc/network/interfaces

We will replace the text there with this text, which will assign 10.0.0.50 to our server, and assume the gateway is at 10.0.0.1:

auto eth0
iface eth0 inet static
address 10.0.0.50
netmask 255.255.255.0
network 10.0.0.0
broadcast 10.0.0.255
gateway 10.0.0.1

restart the network to get the new settings :

sudo /etc/init.d/networking restart

ifconfig eth0 and you should see the new settings.

If you want to see a whole lot more networking commands, see here.

Finding out drives

Great. Now before we can put two drives into a RAID 1 array, we need to tell linux what drives we are going to use for this.

How do we know what disks? I just type dmesg. This will scroll lots of stuff down the screen. If you are in the GUI at a terminal you can scroll. Otherwise type dmesg | less which will let you move through the text with the cursor keys.

This is all the hardware that was found and initialized. You will first see the ataX with X being the number of the drive. You need to look for a section after this that has letters like hda or sda. The drives I am using are SATA drives and they are sda and sdb. I could tell by the size of the drives. (press Q to exit less if you need to)

Partition for Raid 1

Ok, lets tell Linux that we want those drives to be used for raid.

sudo fdisk /dev/sda

[p] will list the partitions. If you had one previously, you will see it. Remove them. Press d for delete, then give a partition number. (no need if you only had one)

Now add a [n]ew [p]rimary partition number [1] with default start and end (the whole disk)

n [enter] p [enter] 1 [enter][enter][enter]

Now [p] again to verify that the partition is there. Note the type: 83. Type [l] to see all the types. We need to set the type to fd for Linux software raid.

t [enter] fd [enter]

Now write the partition table and exit: w [enter]

Now we repeat the process with the other drive.

sudo fdisk /dev/sdb

And repeat the fdisk commands.

Setting up mdadm for Raid1

Ok, now for Linux software raid. First we need to install it.

sudo apt-get install mdadm

When I did this, I was prompted that citadel-server was being installed. This is what is used by default as the mail server. Mdadm is dependant on a mail server so it can mail you of failures. Pick an admin user and say no to external auth.

Once installed you can build an array. I am going to use two drives in a RAID 1, mirrored setup. This duplicates all the data on both drives. They are mirrored. This is the command with my drives used. You will substitute for your devices:

sudo mdadm --create --verbose /dev/md0 --level=1 --raid-devices=2 /dev/sda1 /dev/sdb1

This should tell you that the array /dev/md0 has started.

If you would like to see when the array is built:

cat /proc/mdstat

This will tell you how long it will take, or if it is done.

Ok, that’s it for now.¬† We will look at commands to manipulate the array in a later article in the series. Next up will be creating the resizable volumes on the array using LVM.

rsync --exclude=".*" -ave ssh /my/masters userId@x.x.x.y:/media/pictures

This will copy everything new from /my/masters to the backup machine at /media/pictures without copying any “dot” files. This also does not delete anything from the new machine.

Anyway, I was doing the rsync, but ran out of space on the backup box. A ctrl-z, and now I need to make more space. I am running ext3, so it should be just a few commands. First I unmounted (umount) the filesystems just to be sure. Then :

lvextend -L +40G /dev/vg/pictures

This will grow the logical volume by 40 gig. Next we need to resize the filesystem to fill up the space.

resize2fs /dev/vg/pictures

I have done this before, but this time I got a message to run e2fsck first.

e2fsck -f /dev/vg/pictures

This finished without issue, and I then ran the resize2fs command and it finished after a couple of minutes. Remount the filesystem, a “df” and it confirms that there is more space available.

I removed the last file copied to make sure it wasn’t wrecked somehow, and then kicked off the rsync again. It just picked up where it left off.

This time I want something that doesn’t consume as much power, can be powered on and off easily, is fairly small, I can swap drives on it, does Raid 1 (mirroring), and doesn’t fry eggs on the top of the case.

Options. Not many. I looked at the Buffalo Linkstation Pro Duo, but it only ships in a dual drive configuration, and the 2 TB version seems way too expensive (Amazon $510). You get soaked on the drive cost here. The 1 TB version is two 500GB drives for $265. That is $245 to add another TB. (Without subtracting the cost of the other two drives).

I like the looks of the Netgear ReadyNAS Duo, and it ships with one drive and one open bay. For 500GB, Amazon wants $344, and I would have to replace 2 drives. For the 1 TB version the unit jumps to $617. Yikes. I am still 1 TB short of a mirror. Netgear has other 4 bay versions too, but they get even more expensive. Way more.

I suppose there is always the Drobo. It is a 4 bay solution. Not exactly a mirror, but some sort of wacky internal proprietary solution that provides redundancy. Amazon has this for $349 with no drives, but I would still need to add the Drobo Share to make it a NAS for $195. While a lot of photographers use this solution tied to one computer, I am not sure I like it as a NAS.

Lets try to add this up. First, drives. Newegg and Amazon’s stock of 1 TB drives seems to vary quite a bit. You can get Seagate and Western Digital drives from $179 to $240. I will just use $200 for now. Millage will vary. End price is a NAS with 2 TB total.

• Linkstation 1TB : $265 + 2x$200 = $665
• Linkstation 2TB : $510
• ReadyNAS 500GB : $344 + 2x$200 = $744
• ReadyNAS 1TB: $617 + $200 = $817
• Drobo : $349 + $195 + 2x$200 = $944

So, it looks like out of the box the Linkstation 2TB was the cheapest after all. I do have some concerns about the Linkstation. The unit needs to be disassembled to replace a drive, and I hear that it is not the fastest device.  $500 plus right now sounds like a whole lot of cabbage.

I suppose that trying to find a new motherboard that my existing processor and memory could fit into that would have onboard SATA ports would be another idea to explore. Without the external card, maybe the machine could boot up properly. Still, I think that it uses too much power. Maybe someday I will get a Kill-A-Watt and find out for sure.