Tuesday, March 27, 2012

Using ATA Over Ethernet (AoE) On Debian Squeeze (Initiator And Target)


This guide explains how you can set up an AoE target and an AoE initiator (client), both running Debian Squeeze. AoE stands for "ATA over Ethernet" and is a storage area network (SAN) protocol which allows AoE initiators to use storage devices on the (remote) AoE target using normal ethernet cabling. "Remote" in this case means "inside the same LAN" because AoE is not routable outside a LAN (this is a major difference compared to iSCSI). To the AoE initiator, the remote storage looks like a normal, locally-attached hard drive.
I do not issue any guarantee that this will work for you!

1 Preliminary Note

I'm using two Debian Squeeze servers here:
  • server1.example.com (Initiator): IP address 192.168.0.100
  • server2.example.com (Target): IP address 192.168.0.101

2 Loading The aoe Kernel Module On Both Systems

server1/server2:
Before we start, we must make sure that the the kernel supports AoE:
grep ATA_OVER /boot/config-`uname -r`
This should display something like this:
root@server2:~# grep ATA_OVER /boot/config-`uname -r`
CONFIG_ATA_OVER_ETH=m
root@server2:~#
This means that AoE was built as a kernel module. Let's check if the module is already loaded:
lsmod | grep aoe
If you get nothing back, this means it's not loaded. In this case we can load it as follows:
modprobe aoe
Let's check again if the module is loaded:
lsmod | grep aoe
root@server2:~# lsmod | grep aoe
aoe                    21380  0
root@server2:~#
To have the module loaded automatically when the system boots, we add the aoe module to /etc/modules:
vi /etc/modules
# /etc/modules: kernel modules to load at boot time.
#
# This file contains the names of kernel modules that should be loaded
# at boot time, one per line. Lines beginning with "#" are ignored.
# Parameters can be specified after the module name.

aoe
loop
 

3 Setting Up The Target (server2)

server2:
First we set up the target (server2):
apt-get install vblade
We can use unused logical volumes, image files, hard drives (e.g. /dev/sdb), hard drive partitions (e.g. /dev/sdb1) or RAID devices (e.g. /dev/md0) for the storage. In this example I will create a logical volume of 20GB named storage1 in the volume group vg0:
lvcreate -L20G -n storage1 vg0
(If you want to use an image file, you can create it as follows:
mkdir /storage
dd if=/dev/zero of=/storage/storage1.img bs=1024k count=20000
This creates the image file /storage/storage1.img with a size of 20GB.
)
Now we export our storage device as follows:
vbladed 0 1 eth0 /dev/vg0/storage1
The first number (0) is the shelf number (major), the second (1) the slot number (minor), change these numbers to your liking. Each AoE device is identified by a couple major/minor which must be unique (if you are exporting multiple devices), with major between 0-65535 and minor between 0-255. The eth0 part tells vbladed which ethernet device to use (if you ethernet device is eth1, then use eth1 - you can find out about your ethernet devices by running
ifconfig
).
To start the export automatically whenever you boot the target, open /etc/rc.local...
vi /etc/rc.local
... and add the following line to it (before the exit 0 line):
[...]
/usr/sbin/vbladed 0 1 eth0 /dev/vg0/storage1
[...]

4 Setting Up The Initiator (server1)

 
server1:
On server1, we install the initiator:
apt-get install aoetools
Now we check what AoE storage devices are available:
aoe-discover
The command
aoe-stat
should now show the storage devices:
root@server1:~# aoe-stat
      e0.1        21.474GB   eth0 up
root@server1:~#
At this point we have a new block device available on the client box named /dev/etherd/e0.1. If we have a look at the /dev tree a new node appears:
ls -la /dev/etherd/
root@server1:~# ls -la /dev/etherd/
total 0
drwxr-xr-x  2 root root     160 Mar 16 13:34 .
drwxr-xr-x 14 root root    3000 Mar 16 13:33 ..
c-w--w----  1 root disk 152,  3 Mar 16 13:33 discover
brw-rw----  1 root disk 152, 16 Mar 16 13:34 e0.1
cr--r-----  1 root disk 152,  2 Mar 16 13:33 err
c-w--w----  1 root disk 152,  6 Mar 16 13:33 flush
c-w--w----  1 root disk 152,  4 Mar 16 13:33 interfaces
c-w--w----  1 root disk 152,  5 Mar 16 13:33 revalidate
root@server1:~#
In the output of
fdisk -l
you should now also find the new hard drive:
root@server1:~# fdisk -l

Disk /dev/sda: 32.2 GB, 32212254720 bytes
255 heads, 63 sectors/track, 3916 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x00029d5c

   Device Boot      Start         End      Blocks   Id  System
/dev/sda1   *           1        3793    30461952   83  Linux
/dev/sda2            3793        3917      992257    5  Extended
/dev/sda5            3793        3917      992256   82  Linux swap / Solaris

Disk /dev/etherd/e0.1: 21.5 GB, 21474836480 bytes
255 heads, 63 sectors/track, 2610 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x00000000

Disk /dev/etherd/e0.1 doesn't contain a valid partition table
root@server1:~#
To use that device, we must format it:
fdisk /dev/etherd/e0.1
root@server1:~# fdisk /dev/etherd/e0.1
Device contains neither a valid DOS partition table, nor Sun, SGI or OSF disklabel
Building a new DOS disklabel with disk identifier 0xa00b110d.
Changes will remain in memory only, until you decide to write them.
After that, of course, the previous content won't be recoverable.


The number of cylinders for this disk is set to 2610.
There is nothing wrong with that, but this is larger than 1024,
and could in certain setups cause problems with:
1) software that runs at boot time (e.g., old versions of LILO)
2) booting and partitioning software from other OSs
   (e.g., DOS FDISK, OS/2 FDISK)
Warning: invalid flag 0x0000 of partition table 4 will be corrected by w(rite)

Command (m for help):
 <-- n
Command action
   e   extended
   p   primary partition (1-4)

<-- p
Partition number (1-4): <-- 1
First cylinder (1-2610, default 1): <-- ENTER
Using default value 1
Last cylinder or +size or +sizeM or +sizeK (1-2610, default 2610):
<-- ENTER
Using default value 2610

Command (m for help):
 <-- t
Selected partition 1
Hex code (type L to list codes):
 <-- 83

Command (m for help): <-- w
The partition table has been altered!

Calling ioctl() to re-read partition table.
Syncing disks.
root@server1:~#

Afterwards, the output of
fdisk -l
should look as follows:
root@server1:~# fdisk -l

Disk /dev/sda: 32.2 GB, 32212254720 bytes
255 heads, 63 sectors/track, 3916 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x00029d5c

   Device Boot      Start         End      Blocks   Id  System
/dev/sda1   *           1        3793    30461952   83  Linux
/dev/sda2            3793        3917      992257    5  Extended
/dev/sda5            3793        3917      992256   82  Linux swap / Solaris

Disk /dev/etherd/e0.1: 21.5 GB, 21474836480 bytes
255 heads, 63 sectors/track, 2610 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x0cf58b15

            Device Boot      Start         End      Blocks   Id  System
/dev/etherd/e0.1p1               1        2610    20964793+  83  Linux
root@server1:~#
Now we create a filesystem on /dev/etherd/e0.1p1...
mkfs.ext4 /dev/etherd/e0.1p1
... and mount it for test purposes:
mount /dev/etherd/e0.1p1 /mnt
You should now see the new device in the outputs of...
mount
root@server1:~# mount
/dev/sda1 on / type ext3 (rw,errors=remount-ro)
tmpfs on /lib/init/rw type tmpfs (rw,nosuid,mode=0755)
proc on /proc type proc (rw,noexec,nosuid,nodev)
sysfs on /sys type sysfs (rw,noexec,nosuid,nodev)
udev on /dev type tmpfs (rw,mode=0755)
tmpfs on /dev/shm type tmpfs (rw,nosuid,nodev)
devpts on /dev/pts type devpts (rw,noexec,nosuid,gid=5,mode=620)
/dev/etherd/e0.1p1 on /mnt type ext4 (rw)
root@server1:~#
... and
df -h
root@server1:~# df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/sda1              29G  806M   27G   3% /
tmpfs                 249M     0  249M   0% /lib/init/rw
udev                  244M  108K  244M   1% /dev
tmpfs                 249M     0  249M   0% /dev/shm
/dev/etherd/e0.1p1     20G  172M   19G   1% /mnt
root@server1:~#
You can unmount it like this:
umount /mnt
To have the device mounted automatically at boot time, e.g. in the directory /storage, we create that directory...
mkdir /storage
... and add the following line to /etc/fstab:
vi /etc/fstab
[...]
/dev/etherd/e0.1p1       /storage        ext4    defaults,auto,_netdev 0 0
This alone isn't enough to have the device mounted at boot time because the AoE stuff gets loaded after /etc/fstab is read. Therefore we open /etc/rc.local...
vi /etc/rc.local
... and add the following lines to it (before the exit 0 line):
[...]
aoe-discover
sleep 5
mount -a
[...]
For test purposes, you can now reboot the system:
reboot
After the reboot, the device should be mounted:
mount
root@server1:~# mount
/dev/sda1 on / type ext3 (rw,errors=remount-ro)
tmpfs on /lib/init/rw type tmpfs (rw,nosuid,mode=0755)
proc on /proc type proc (rw,noexec,nosuid,nodev)
sysfs on /sys type sysfs (rw,noexec,nosuid,nodev)
udev on /dev type tmpfs (rw,mode=0755)
tmpfs on /dev/shm type tmpfs (rw,nosuid,nodev)
devpts on /dev/pts type devpts (rw,noexec,nosuid,gid=5,mode=620)
/dev/etherd/e0.1p1 on /storage type ext4 (rw,_netdev)
root@server1:~#
df -h
root@server1:~# df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/sda1              29G  806M   27G   3% /
tmpfs                 249M     0  249M   0% /lib/init/rw
udev                  244M  108K  244M   1% /dev
tmpfs                 249M     0  249M   0% /dev/shm
/dev/etherd/e0.1p1     20G  172M   19G   1% /storage
root@server1:~#

5 Links

No comments:

Post a Comment