1. Introduction to UEC and its components
2. Installation & Configuration
3. Web Interface
4. Image Management
5. Instance Management
6. Storage Management
7. Network Management
8. Security
9. Troubleshooting
10. Euca Commands
11. Hacks

This is a live book and will be updated on an ongoing basis based on your feedback and comments.

Lo!!! there ends the similarity….

I wasn’t able to access the instance. No network connectivity….. neither PING nor RDP works. And this was frustrating!!!

As mentioned in my earlier post about UEC/Eucalyptus – Debugging instances, I tweaked the /usr/share/gen_kvm_libvirt_xml file on the Node Controller and added the VNC display option.

The outcome of this was surprising!!!!

UEC: Windows instance on Lucid Lynx [HACK]

It is not safe to leave this option enabled on a production setup.

Bringing up eucalyptus instance with VNC attached

The Storage Controller is a block storage in Eucalyptus which is compatible with Amazon’s Elastic Block Store. euca2ools or EC2 commands can be used to control these volumes. Below we specify on using euca2ools commands for volume creation, attaching a volume to a running instance and how-to use the attached volume in the instance.

Interacting with Storage Controller

Linux Instances

First create a volume of the required size using euca-create-volume

$ euca-create-volume -s 10 -z mycloud
VOLUME  vol-333C04B8    10       creating        2010-03-26T05:20:56.383Z

Check the available volumes and their status with euca-describe-volumes

$ euca-describe-volumes
VOLUME  vol-333C04B8    10   mycloud   <strong>available</strong> 2010-03-26T05:20:56.383Z

Now attach the volume to a running instance using euca-attach-volume

$ euca-attach-volume -i i-41620887 -d /dev/sdb vol-333C04B8

The volumes status changes from “available” to “in-use” if it is attached properly.

$ euca-describe-volumes
VOLUME  vol-333C04B8    10  mycloud <strong>in-use</strong>  2010-03-26T05:41:28.019Z
ATTACHMENT      vol-331F04B2    i-41620887      unknown,requested:/dev/sdb  2010-03-26T05:43:51.343Z

Once you login to the instance through ssh, you should see the new volume.

$ sudo fdisk -l
Disk /dev/sda: 2195 MB, 2195718144 bytes
4 heads, 32 sectors/track, 33504 cylinders
Units = cylinders of 128 * 512 = 65536 bytes
Disk identifier: 0x000adc1c
Device Boot      Start         End      Blocks   Id  System
/dev/sda1               1       32769     2097152+  83  Linux
/dev/sda3           32769       33504       47071   82  Linux swap / Solaris
Disk /dev/sdb: 10.7 GB, 10737418240 bytes
64 heads, 32 sectors/track, 10240 cylinders
Units = cylinders of 2048 * 512 = 1048576 bytes
Disk identifier: 0x00000000
Disk /dev/sdb doesn't contain a valid partition table

Now /dev/sdb is ready to be formatted and mounted so that it can be used as a permanent storage. Unmount the volume before bringing down the instance, to avoid any data corruption.

Windows Instances

For windows instances, creating the volume and attaching the volume would be the same.

NOTE: Even while attaching to windows instances, the device has to be specified as /dev/sdb.

The only change is the way this disk is accessed. Once the disk is attached to a running windows instance, perform the following steps:

• Connect to windows instance via RDP.
• Go to Start -> Run and type diskmgmt.msc without the quotes. Select OK. This opens the Disk Management console.
• Right click on the new disk at the bottom of the list and choose Initialize Disk.
• A pop up will appear, choose OK.
• Next right click on the volume and select New Partition.
• A wizard appears and accept the defaults, other than change the disk letter as per need.
• Finish up the wizard, and windows will now format the partition.

After completion the disk management tool shows a healthy partition on the new drive.

To detach a volume from an instance use euca-detach-command command:

$ euca-detach-volume vol-333C04B8

Note that the volume has to unmounted before detaching it from the instance. Volume can be deleted using euca-delete-volume command:

$ euca-delete-volume vol-333C04B8

WS3 Storage Controller

Walrus is a storage service in Eucalyptus which is compatible with Amazon’s S3. Using Walrus users can store persistent data, which is organized as buckets and objects. WS3 is a file level storage system, as compared to the block level storage system of Storage Controller.

See Amazon’s S3 Getting Started Guide for more information http://docs.amazonwebservices.com/AmazonS3/2006-03-01/gsg/.

Walrus Controller options can be modified from the Web UI, on the “Configuration” page under “Walrus Configuration” section.

Third party tools for interacting with Walrus

• s3curl S3 Curl is a command line tool that is a wrapper around curl. http://open.eucalyptus.com/wiki/s3curl
• s3cmd is a tool that allows command line access to storage that supports the S3 API. http://open.eucalyptus.com/wiki/s3cmd
• s3fs is a tool that allows users to access S3 buckets as local directories. http://open.eucalyptus.com/wiki/s3fs

DISCLAIMER: I haven’t yet used any of the above tools to interact with Walrus Storage, so can’t authenticate how well they gel with Walrus Controller. Need to check the tools.

Snapshots

The volumes that have been created on the Block Storage can be used to create point-in-time snapshots of volumes, which are stored on WS3. These snapshots can be used as the starting point for new Block Storage volumes and protect data for long-term durability. The same snapshot can be used to create many volumes as per requirement.

Snapshots can be created by using euca-create-snapshot command:

$ euca-create-snapshot vol-333C04B8

euca-describe-snapshots lists the available snapshots.

$ euca-describe-snapshots
SNAPSHOT        snap-32A804A2   vol-333C04B8    <strong>completed</strong>       2010-04-15T13:48:32.01Z      100%

Volumes from these snapshots can be created by using the snapshot id.

$ euca-create-volume -s 10 --snapshot snap-32A804A2 --zone mycloud

Please note that the volume can only be created after the snapshot status is reflected as “completed” as shown by euca-describe-snapshots command. Similarly, snapshots from a volume can only be created after the volume status is “created”, but not when it is “creating”. The volume status is obtained from euca-describe-volumes command.

To delete a snapshot:

$ euca-delete-snapshot snap-32A804A2

Installation

The following sections describe how to install and configure a basic UEC setup spanning 3 servers. 2 Servers(Server1 and Server2) will run Lucid 64-bit server version and server will run Lucid Desktop 64-bit version(Client1). We have decided on installing the Desktop version on Client1 so that we can use Firefox or other browsers to access the web interface of UEC and also to use RDP/VNC clients during the image creation process.

Server1 – To run Cloud Controller, Cluster Controller, Walrus Storage service and Storage Controller

Hardware       Minimum            Suggested
CPU            1GHz               2 x 2GHz
Memory         1GB                2GB
Disk           5400rpm IDE        7200rpm SATA
Disk Space     40GB               200GB
Networking     100Mbps            1000Mbps

Server2 – To run node-controller for running Virtual Machine (VM) instances

Hardware       Minimum            Suggested
CPU            VT extensions      VT, 64-bit, Multicore
Memory         1GB                4GB
Disk           5400rpm IDE        7200rpm SATA or SCSI
Disk Space     40GB               100GB
Networking     100Mbps            1000Mbps

Client1 – To function as the client for the cloud and also to be used for bundling images etc.

Hardware       Minimum            Suggested
CPU            VT extensions      VT, 64-bit, Multicore
Memory         1GB                4GB
Disk           5400rpm IDE        7200rpm SATA or SCSI
Disk Space     40GB               100GB
Networking     100Mbps            1000Mbps

For ease of understanding, the instructions here assume the following sample setup. Please make necessary modifications to suit your setup.

• Server1 has 2 NICs. eth0 is connected to the enterprise network and eth1 is connected to a dedicated switch to be used for connecting various components of the UEC.
• Server1 will have the following IP addresses:  eth0 – 192.168.10.121 (To be setup during the installation process) eth1 – 192.168.20.1 (To be setup after the first reboot after installation)
• Server1 has the hostname server1.example.com
• Server2 has 2 NICs, eth0 is connected to the dedicated switch mentioned above and eth1 is connected to the enterprise network to have internet connectivity during the install process. This interface can later be deactivated.
• Server2 will have the following IP address: eth0 – 192.168.20.2 (To be setup during the installation process) eth1 – 192.168.10.122 (To be setup after the first reboot after installation)
• Server2 has the hostname server2.example.com
• Client1 has the hostname client1.example.com
• Server1 is a 64-bit server and Server2 is a 64-bit VT-enabled server.
• The enterprise network runs on a class ‘C’ private network – 192.168.10.0/255.255.255.0
• Name servers of the enterprise – 192.168.10.2 and 192.168.10.3
• The following IP addresses are allocated for the cloud instances as public IP addresses( from the enterprise nework range used) 192.168.10.200-192.168.10.220
• The UEC components will use a dedicated class ‘C’ private network for their interconnections – 192.168.20.0/255.255.255.0
• Cluster Name – “myueccluster”
• The first user name created on Server1, Server2 and Client1 is “uecadmin” and we perform all the operations while logged in as “uecadmin”

Server1

Installation

• Boot the server off the Ubuntu Server 10.04 CD. At the graphical boot menu, select “Install Ubuntu Enterprise Cloud” and proceed with the basic installation steps.
• Installation only lets you set up the IP address details for one interface. Please do that for eth0.
• You will need to choose certain configuration options for your UEC, during the course of the install:

1. Cloud Controller Address – Leave this blank as server1 is the Cloud Controller in this setup.
2. Cloud Installation Mode – Select ‘Cloud controller’, ‘Walrus storage service’, ‘Cluster controller’ and ‘Storage controller’.
3. Network interface for communication with nodes – eth1
4. Eucalyptus cluster name – myueccluster
5. Eucalyptus IP range – 192.168.10.200-192.168.10.220

Post Install Steps & Configuration

• Set up static IP for the second network interface eth1 by adding the following to /etc/network/interfaces

auto eth1
iface eth1 inet static
address 192.168.20.1
netmask  255.255.255.0
network 192.168.20.0
broadcast 192.168.20.255

• After making the above changes, run the following command to restart the networking.

uecadmin@server1:~$sudo /etc/init.d/networking restart

• Make sure that you have the latest version of EUCALYPTUS.

uecadmin@server1:~$sudo apt-get update
uecadmin@server1:~$sudo apt-get upgrade eucalyptus

• Install the NTP package, this machine is going to act as an NTP server for the nodes. The time on all components of UEC will have to be in sync. Since CLC is connected to Internet, we can run ntp server on it and have other components sync to it.

uecadmin@server1:~$sudo apt-get install ntp

• Open the file /etc/ntp.conf and add the following two lines to make sure the server serves time even when its connectivity to the Internet is down. The following settings makes sure that the ntp server uses its own clock as the clock source.

server 127.127.1.0
fudge  127.127.1.0 stratum 10

• Restart ntp service to make the changes effective.

uecadmin@server1:~$sudo /etc/init.d/ntp restart

• Save the changes and restart CC.

uecadmin@server1:~$sudo start eucalyptus-cc CLEAN=1

Server2

Installation

• Boot the server off the Ubuntu Server 10.04 CD. At the graphical boot menu, select “Install Ubuntu Enterprise Cloud” and proceed with the basic installation steps.
• Installation only lets you set up the IP address details for one interface. Please do that for eth0.
• You will need to choose certain configuration options for your UEC, during the course of the install. You can ignore all the settings, except the following:

1. Cloud Controller Address – 192.168.20.1
2. Cloud Installation Mode – Select ‘Node Controller’

Post Install Steps & Configuration

• Set up static IP for the second network interface eth1 by adding the following to /etc/network/interfaces

auto eth1
iface eth1 inet static
address 192.168.10.122
netmask  255.255.255.0
network 192.168.10.0
broadcast 192.168.10.255

• After making the above changes, run the following command to restart the networking.

uecadmin@server2:~$sudo /etc/init.d/networking restart

• Add your DNS server details to /etc/resolv.conf

nameserver 192.168.10.2
nameserver 192.168.10.3

• Make sure that you have the latest version of EUCALYPTUS.

uecadmin@server2:~$sudo apt-get update
uecadmin@server2:~$sudo apt-get upgrade eucalyptus

• Install the NTP package to synchronize with the NTP server

uecadmin@server2:~$sudo apt-get install ntp

• Open the file /etc/ntp.conf and add the line

server 192.168.20.1

• Restart ntp service to make the changes effective.

uecadmin@server2:~$sudo /etc/init.d/ntp restart

• Open the file /etc/eucalyptus/eucalyptus.conf and make the following changes

VNET_PUBINTERFACE="br0"
VNET_PRIVINTERFACE="br0"
VNET_BRIDGE="br0"
VNET_DHCPDAEMON="/usr/sbin/dhcpd3"
VNET_DHCPUSER="dhcpd"
VNET_MODE="MANAGED-NOVLAN"

• After making the above changes, run the following commands

uecadmin@server2:~$sudo restart eucalyptus-nc-publication
uecadmin@server2:~$sudo restart eucalyptus-nc

Install CC’s ssh public key to NC

Next we need to install the Cluster Controller’s eucalyptus user’s public ssh key into the Node Controller’s eucalyptus user’s authorized_keys file.

• On the Node Controller, temporarily set a password for the eucalyptus user:

uecadmin@server2:~$sudo passwd eucalyptus

• On the Cluster Controller:

uecadmin@server2:~$sudo -u eucalyptus ssh-copy-id -i ~eucalyptus/.ssh/id_rsa.pub eucalyptus@192.168.20.2

• You can now remove the password of the eucalyptus account on the Node:

uecadmin@server2:~$sudo passwd -d eucalyptus

Client1

The purpose of Client1 machine is to interact with the Cloud setup, for bundling and registering new Eucalyptus Machine Image (EMI)

Installation

• Boot the Desktop off the Ubuntu Desktop 10.04 CD and install. The Desktop would be on the enterprise network and obtaining an IP addresses through DHCP.
• Install KVM to help us install images on KVM platform and bundle them.

uecadmin@client1:~$apt-get install qemu-kvm

Post Install Configuration

• Install euca2ools to be able to manage the cloud from it.

uecadmin@client1:~$sudo apt-get install euca2ools

• Login to the web interface of CLC by using the following link https://192.168.10.121:8443. Username is admin and password is admin.

• Note that the installation of UEC installs a self signed certificate for the web server. The browser will warn you about the certificate not being signed by a trusted certifying authority. You can authorize the browser to access the server with the self signed certificate. Installing a valid certificate signed by a trusted certifying authorities is beyond the scope of this guide.

• When you login for the first time, the web interface prompts you to change the password and provide the email ID of the admin. After completing this mandatory step, download the credentials archive from https://192.168.10.121:8443/#credentials and save to new location in ~/.euca directory.
• Extract the credentials archive.

cd .euca
unzip mycreds.zip

• To validate that the euca2ools are able to communicate with the UEC, get the local cluster availability details. Source eucarc script to make sure that the environmental variables used by euca2ools are set properly.

$ . ~/.euca/eucarc
$ euca-describe-availability-zones verbose

You should see something like the following if it works properly:

$ euca-describe-availability-zones verbose
AVAILABILITYZONE        myueccluster    192.168.10.121
AVAILABILITYZONE        |- vm types     free / max   cpu   ram  disk
AVAILABILITYZONE        |- m1.small     0016 / 0016   1    128     2
AVAILABILITYZONE        |- c1.medium    0014 / 0014   1    256     5
AVAILABILITYZONE        |- m1.large     0007 / 0007   2    512    10
AVAILABILITYZONE        |- m1.xlarge    0003 / 0003   2   1024    20
AVAILABILITYZONE        |- c1.xlarge    0001 / 0001   4   2048    20

If you see the free/max VCPUs as 0 in the above list, it means the node did not get registered automatically. Please use the following on Server1 and approve when prompted to add 192.168.20.2 as the Node Controller.

uecadmin@client1:~$sudo euca_conf --discover-nodes

Please see the following posts on how to bundle Linux and Windows images:

• UEC: Bundling Linux Image
• UEC: Bundling Windows Image

Cloud Computing is a computing model, where resources such as computing power, storage, network and software are abstracted and provided as services on the Internet in a remotely accessible fashion. Billing models for these services are generally similar to the ones adopted for public utilities. On-demand availability, ease of provisioning, dynamic and virtually infinite scalability are some of the key attributes of Cloud Computing.

An infrastructure setup using the cloud computing model is generally referred to as ‘Cloud’. The following are the broad categories of services available on the ‘Cloud’:

1. Infrastructure As A Services (IAAS)
2. Platform As A Service (PAAS)
3. Software As A Service (SAAS)

This ‘Cloud’ is generally available to anyone on the Internet willing to pay. However, a variant called ‘Private Cloud’ is increasingly becoming popular for private infrastructure that has some of the attributes of the ‘Cloud’ as mentioned above.

Eucalyptus

Eucalyptus is a software available under GPL that helps in creating and managing a private or even a publicly accessible cloud. It provides an EC2 compatible cloud computing platform and S3 compatible cloud storage platform. Eucalyptus has become very popular and is seen as one of the key open source cloud platforms. Since Eucalyptus makes its services available through EC2/S3 compatible APIs, the client tools written for AWS can be used with Eucalyptus as well.

Ubuntu Enterprise Cloud (UEC)

Ubuntu Enterprise Cloud, UEC for short, is a stack of applications from Canonical included with Ubuntu Server Edition. UEC includes Eucalyptus along with a number of other open source software. UEC makes it very easy to install and configure the Cloud. Canonical also provides commercial technical support for UEC.

UEC Components

1. Node Controller (NC)
2. Cluster Controller (CC)
3. Walrus Controller (WS3)
4. Storage Controller (SC)
5. Cloud Controller (CLC)

Node Controller (NC)

A UEC node is a VT enabled server capable of running KVM as the hypervisor. UEC automatically installs KVM when the user chooses to install the UEC node. The VMs running on the hypervisor and controlled by UEC are called instances.

Node Controller runs on each node and controls the life cycle of instances running on the node. The NC interacts with the OS and hypervisor running on the node on one side and the CC on the other side.

NC queries the Operating System running on the node to discover the node’s physical resources – the number of cores, the size of memory, the available disk space and also to learn about the state of VM instances running on the node and propagates this data up to the CC.

Functions:

1. Collection of data related to the resource availability and utilization on the Node and reporting the data to CC.
2. Instance life cycle management.

Cluster Controller (CC)

CC manages certain number of Node Controllers and deploys instances using them. CC also manages the networking for the instances running on the Nodes under certain types of networking modes of Eucalyptus. More on this in later chapters.

CC communicates with CLC on one side and NCs on the other side.

Functions:

1. To receive requests from CLC to deploy instances.
2. To decide which NCs to use for deploying the instances on.
3. To control the virtual network available to the instances.
4. To collect information about the NCs registered with it and report it to the CLC.

Walrus Storage Controller (WS3)

WS3 provides a simple storage service using REST and SOAP APIs compatible with S3 APIs from Amazon Web Services.

Functions:

1. Storing the the machine images.
2. Storing snapshots.
3. Storing and serving files (either from a running instance or from anywhere on the web) using S3 API

WS3 should be considered as a simple file storage system.

Storage Controller (SC)

SC provides persistent block storage for use by the instances. This is similar to the EBS service from AWS.

SC runs on the same machine as the Cluster Controller and is configured automatically when the Cluster Controller is installed.

Functions:

1. Creation of persistent Elastic Block Storage (EBS) devices.
2. Providing the block storage over AoE or iSCSI protocol to the instances.
3. Allowing creation of snapshots of volumes.

Cloud Controller (CLC)

The Cloud Controller (CLC) is the front end to the entire cloud infrastrcuture. CLC provides an EC2/S3 compliant web services interface to the client tools on one side and interacts with the rest of the components of the Eucalyptus infrastructure on the other side. CLC also provides a web interface to the users for managing certain aspects of the UEC infrastructure.

Functions:

1. Monitor the availability of resources on various components of the cloud infrastructure, including hypervisor nodes that are used to actually provision the instances and the cluster controllers that manage the hypervisor nodes.
2. Resource arbitration – Decide which clusters will be used for provisioning the instances.
3. Monitoring the running instances.

In short, CLC has a comprehensive knowledge of the availability and usage of resources in the cloud and the state of the cloud.

On CC

axis2c.log – This file logs messages specific to Axis2c web services.

cc.log – This file logs messages specific to Cluster controller.

cc-registration.log – This file logs the success or failure of Cluster Controller registration. For example,

SUCCESS: new cluster 'mycloud' on host '192.168.1.1' successfully registered.

cloud-debug.log, cloud-error.log, cloud-output.log – These files log messages specific to Cloud Controller.

sc-registration.log – This file logs the success or failure of Storage Controller registration. For example,

SUCCESS: new SC for cluster 'mycloud' on host '192.168.1.1' successfully registered.
Adding SC 192.168.1.1 to cluster mycloud

sc-stats.log – This file logs statistics specific to Storage Controller.

walrus-registration.log – This file logs the success or failure of Walrus Storage Controller registration. For example,

SUCCESS: new walrus on host '192.168.1.1' successfully registered.
Adding WALRUS host 192.168.1.1

walrus-stats.log – This file logs statistics specific to Walrus Storage Controller.

httpd-cc_error_log

On NC

axis2c.log – This file logs messages specific to Axis2c web services.

nc.log – This file logs messages specific to Node controller.

euca_test_nc.log
httpd-nc_error_log

The logging level of these files is controlled by the LOGLEVEL macro in eucalyptus.conf. This macro controls the level of logging output that appears in various eucalyptus log files. The options are DEBUG, INFO, WARN, ERROR, and FATAL (in descending order of verbosity). The default is DEBUG (everything).

This will be the main HDD in the virtual image, so make sure to give it as much space as you’ll need. Since we’re building a kvm image, we can use a qcow2 format for disk images. Qcow2 is an expandable image format, so it’ll only take as much storage space as it’s actually used within the image.

$ kvm-img create -f qcow2 win-2k3.img 20G

OS Installation

Get the iso file of the Windows distribution to be installed in the image. And start the installation process.

$sudo kvm -m 1024 -cdrom Win2003_cd1.iso -drive file=win-2k3.img,if=scsi,boot=on -nographic -vnc :0

After finishing the installation and in case the installation needs 2nd CD as well, reboot the VM and launch the VM by the following command.

$ sudo kvm -m 1024 -boot c -cdrom Win2003_cd2.iso -drive file=win-2k3.img,if=scsi,boot=on -nographic -vnc :0

Once installation is complete, create the boot disk needed for Windows 2003 server.

Copying the boot files

Re-launch the VM using the following command, so that the first CD can be accessed in order to copy some files from it.

$ sudo kvm -m 1024 -boot c -cdrom Win2003_cd1.iso -drive file=win-2k3.img,if=scsi,boot=on -nographic -vnc :0

Copy ntldr, ntdetect.com and sym_hi.sys files from CD1, to a specific location to create the boot disk, say C:\bootfiles

Rename sym_hi.sys to Ntbootdd.sys (this is the SCSI driver needed to boot the hard disk). In case, sym_hi.sys is not present in the CD, copy it from the following location of the installation

C:\WINDOWS\system32\drivers

Write the following lines to a new file and save it as Boot.ini in the same location, i.e, C:\bootfiles

[boot loader]
timeout=30
Default= multi(0)disk(0)rdisk(0)partition(1)\Windows

[operating systems]
multi(0)disk(0)rdisk(0)partition(1)\WINDOWS="Windows Server 2003"

Shut down the VM and proceed for the boot disk creation.

Boot disk creation

On CC, execute the following command to create the boot disk.

$ dd bs=512 count=2880 if=/dev/zero of=win-boot.img

Now attach this disk as a floppy disk and start the Windows VM.

$ sudo kvm -m 1024 -boot c -drive file=win-2k3.img,if=scsi,boot=on,index=1 -fda win-boot.img -nographic -vnc :0

On the Windows VM, format the floppy disk and copy all the files from C:\bootfiles to the floppy.
Shutdown the VM and test the boot disk created with the following command

$ sudo kvm -m 1024 -boot a -drive file=win-2k3.img,if=scsi,boot=on,index=1 -fda win-boot.img -nographic -vnc :0

This should boot into windows, if the boot floppy is properly created.

Network Configuration

Now download the e1000 drivers, since e1000 is the default interface that is supported by Eucalyptus. Download the drivers from the following location:
e1000 Drivers Download

For 32-bit VM, download PRO2KXP.exe
For 64-bit VM, download PROEM64T .exe

Once the download is finished, shutdown the VM and restart the VM with the following command.

$ sudo kvm -m 1024 -boot a -drive file=win-2k3.img,if=scsi,boot=on,index=1 -fda win-boot.img -net nic,model=e1000 -net tap-nographic -vnc :0

The above command tell the VM to make use of e1000 model as its NIC. After executing the above command, the following warning is seen on the console which can be ignored.

Warning: vlan 0 is not connected to host network

Once the VM boots, install the network drivers.

Enable remote desktop for all users on the VM. This is needed, as this is the only way, without installing third party tools, Windows machines can be accessed remotely. Enable remote desktop by navigating to Start → Settings → Control Panel → System → Remote and select “Remote Desktop” option for all users. This completes the Windows disk setup. Shut down the VM.

Memdisk

The kernel memdisk can be used to load Windows. The default syslinux memdisk can be used for this purpose. memdisk is present in /usr/lib/syslinux

If it is not present, install syslinux by the following command.

$ sudo apt apt-get install syslinux

After installing, memdisk will be found in /usr/lib/syslinux

$ cd /usr/lib/syslinux
$ ls -l memdisk
-rw-r--r-- 1 root root 20068 2008-07-15 20:25 memdisk

Copy it to the location where all windows files are present, i.e., win-2k3.img and win-boot.img

Registering with UEC

The last step would be to upload the images to UEC. The files that need to be uploaded are:

• memdisk
• win-boot.img
• win-2k3.img

Before starting the upload process shut down the VM.

Registering memdisk

Execute the following commands to bundle and register the memdisk.

$ euca-bundle-image -i memdisk --kernel true
$ euca-upload-bundle -b win2k3_bucket -m /tmp/memdisk.manifest.xml
$ euca-register win2k3_bucket/memdisk.manifest.xml

Save the output produced by the last command above (eki-XXXXXXXX), which will be needed while registering the disk image.

Registering boot disk

Execute the following commands to bundle and register the boot disk (win-boot.img).

$ euca-bundle-image -i win-boot.img --ramdisk true
$ euca-upload-bundle -b win2k3_bucket -m /tmp/win-boot.img.manifest.xml
$ euca-register win2k3_bucket/win-boot.img.manifest.xml

Save the output produced by the last command above (eri-XXXXXXXX), which will be needed while registering the disk image.

Registering disk image

Execute the following commands to bundle and register the disk image (win-2k3.img)

$ euca-bundle-image -i win-2k3.img --kernel eki-XXXXXXXX --ramdisk eri-XXXXXXXX
$ euca-upload-bundle -b win2k3_bucket -m /tmp/win-2k3.img.manifest.xml
$ euca-register win2k3_bucket/win-2k3.img.manifest.xml

Replace eki-XXXXXXXX and eri-XXXXXXXX with the exact values as noted down earlier.

Image Listing

The new images that have been uploaded can be seen by using euca-describe-images command.
For Ex:

$ euca-describe-images
IMAGE   emi-55470DE2    win-2k3/win-2k3.img.manifest.xml        admin   available    public  x86_64  machine
IMAGE   eki-2BC30D1B    win-2k3/memdisk.manifest.xml    admin   available   public   x86_64  kernel
IMAGE   eri-76C20ED5    win-2k3/win-boot.img.manifest.xml       admin   available    public  x86_64  ramdisk

Security Group

Verify that the security group to be used with this instance is allowing RDP traffic.

$ euca-describe-groups
GROUP   admin   default default group
PERMISSION      admin   default ALLOWS  tcp     3389    3389    FROM    CIDR0.0.0.0/0

If the above entry is not present, allow RDP with the following command.

$ euca-authorize default -P tcp -p 22 -s 0.0.0.0/0

Running the instance

To launch a new instance of the custom image, execute euca-run-instances command.

$ euca-run-instances  emi-55470DE2 -k mykey -t c1.xlarge

Since this is a windows instance the VM type is specified as c1.xlarge, which has 20GB of hard disk space. Also specifying the key is not mandatory, as the access to the instance would be through RDP.

Accessing the instance

From the remote machine, navigate to Start → Programs → Accessories → Remote Desktop Connection.
Specify the public ip address that the instance gets and click Connect.
The desktop of the windows instance should be accessible after providing the credentials.

UPDATE: For Windows XP bundling

While bundling Windows XP, the OS doesn’t recognize SCSI driver during installation process.

The following workaround is suggested:
1. Attach the disk as IDE disk.

$ sudo kvm -m 1024 -cdrom winxpcd.iso -drive file=winxp.img,boot=on -nographic -vnc :0

2. Finish the installation of Windows XP.
3. Now create another disk image using kemu-img

$ kvm-img create -f qcow2 newdisk.img 5G

4. Start the instance by attaching the second disk as SCSI disk.

$ sudo kvm -m 1024 -drive file=winxp.img,boot=on -drive file=newdisk.img,if=scsi -nographic -vnc :0

5. After Windows XP boots, it detects and installs the driver for the SCSI disk (sys_hi.sys)

The remaining procedure is same as that of bundling Windows 2003 Server.

First create a volume of the required size using euca-create-volume

$ euca-create-volume -s 10 -z mycloud
VOLUME  vol-333C04B8    10       creating        2010-03-26T05:20:56.383Z

Check the available volumes and their status with euca-describe-volumes

$ euca-describe-volumes
VOLUME  vol-333C04B8    10   mycloud   available     2010-03-26T05:20:56.383Z

Now attach the volume to a running instance using euca-attach-volume

$ euca-attach-volume -i i-41620887 -d /dev/sdb vol-333C04B8

The volumes status changes from “available” to “in-use” if it is attached properly.

$ euca-describe-volumes
VOLUME  vol-333C04B8    10  mycloud in-use  2010-03-26T05:41:28.019Z
ATTACHMENT      vol-331F04B2    i-41620887      unknown,requested:/dev/sdb  2010-03-26T05:43:51.343Z

Once you login to the instance through ssh, you should see the new volume.

$ sudo fdisk -l

Disk /dev/sda: 2195 MB, 2195718144 bytes
4 heads, 32 sectors/track, 33504 cylinders
Units = cylinders of 128 * 512 = 65536 bytes
Disk identifier: 0x000adc1c

 Device Boot      Start         End      Blocks   Id  System
/dev/sda1               1       32769     2097152+  83  Linux
/dev/sda3           32769       33504       47071   82  Linux swap / Solaris

Disk /dev/sdb: 10.7 GB, 10737418240 bytes
64 heads, 32 sectors/track, 10240 cylinders
Units = cylinders of 2048 * 512 = 1048576 bytes
Disk identifier: 0x00000000

Disk /dev/sdb doesn't contain a valid partition table

Now /dev/sdb is ready to be formatted and mounted so that it can be used as a permanent storage. Unmount the volume before bringing down the instance, to avoid any data corruption.

Windows Instances

For windows instances, creating the volume and attaching the volume would be the same.

NOTE: Even while attaching the device need to specified as /dev/sdb

The only change is the way this disk is accessed. Once the disk is attached to a running windows instance, perform the following steps:

• Connect to windows instance via RDP
• Go to Start -> Run and type diskmgmt.msc without the quotes. Select OK. This opens the Disk Management console.
• Right click on the new disk at the bottom of the list and choose Initialize Disk.
• A pop up will appear, choose OK.
• Next right click on the volume and select New Partition.
• A wizard appears and accept the defaults, other than change the disk letter as per need.
• Finish up the wizard, and windows will now format the partition.

After completion the disk management tool shows a healthy partition on the new drive.

Troubleshooting

If the network configuration on CC and NC has changed, say either the interfaces changed, due to which the ip addresses have also changed. In such cases the following errors appear in nc.log

[EUCAERROR ] libvirt: operation failed: adding scsi disk failed: pci_add pci_addr=auto storage file=/dev/etherd/e0.8,if=scsi
failed to add file=/dev/etherd/e0.8,if=scsi
 (code=9)
[Fri Mar 26 10:57:39 2010][001621][EUCAERROR ] virDomainAttachDevice() failed (err=-1) XML=
[Fri Mar 26 10:57:39 2010][001621][EUCAERROR ] ERROR: doAttachVolume() failed error=1

In order to overcome this, the configuration has to be changed accordingly from the webUI. Access the webUI of UEC using https://:8443. On Configuration page, under Clusters section, replace the correct ip and interface name for Storage Controller.

See below screenshot.

This will be the main HDD in the virtual image, so make sure to give it as much space as youll need. Since were building a kvm image, we can use a qcow2 format for disk images. Qcow2 is an expandable image format, so itll only take as much storage space as its actually used within the image.

$ kvm-img create -f qcow2 image.img 5G

OS Installation

Download the iso file of the Linux distribution you want installed in the image.

$ wget http://releases.ubuntu.com/karmic/ubuntu-9.04-server-amd64.iso

And start the installation process:

$ sudo kvm -m 256 -cdrom ubuntu-9.04-server-amd64.iso -drive file=image.img,if=scsi,index=0 \
-boot d -net nic,vlan=0,model=e1000,macaddr= 00:16:3e:de:ad:01 -net tap -nographic -vnc :0

If your installation process requires more than 256MB of RAM change the -m option, and if you need more processors available, you can use the -c option.

The command above will boot a new KVM instance, with the disk image youve created as the primary HDD and the iso as the first bootable device. Also the -nographic option will not display any graphical output. You can connect to the instance through VNC (use display number :0 ) and finish the installation.

For Ex: vncviewer A.B.C.D :0, where A.B.C.D is the IP address of the CC.

After finishing the installation, relaunch the VM by executing the following command.

$ sudo kvm -m 256 -drive file=image.img,if=scsi,index=0,boot=on -boot c \
-net nic,vlan=0,model=e1000,macaddr=00:16:e3:de:ad:01 -net tap -nographic -vnc :0

At this point you can add all the packages you want to have installed, update the installation, add users and any settings that need to be present in your new UEC instances.

$ sudo apt-get update
$ sudo apt-get upgrade
$ sudo apt-get install mediawiki

Integrating with UEC

The new image needs to know what IP it has when started in UEC and also, it needs to have the public key of the user allowed to do a passwordless access through SSH. The way its done in UEC is via a restful interface provided by the cloud. The interface is available under this URL: http://169.254.169.254/latest/meta-data.

First install curl on the VM.

$ sudo apt-get install curl

And add the following lines to /etc/rc.local of the image.

depmod -a
modprobe acpiphp

# simple attempt to get the user ssh key using the meta-data service
# assuming user is the username of an account that has been created
mkdir -p /home/user/.ssh
echo >> /home/user/.ssh/authorized_keys
curl -m 10 -s http://169.254.169.254/latest/meta-data/public-keys/0/openssh-key | grep 'ssh-rsa' >>  /home/user/.ssh/authorized_keys
echo "AUTHORIZED_KEYS:"
echo "************************"
cat /home/user/.ssh/authorized_keys
echo "************************"

Add the above lines before the exit 0 in /etc/rc.local.

Also remove persistent udev rules from /etc/udev/rules.d. This is needed because the interface name will change whenever the instance reboots and Eucalyptus expects the interface to be as eth0.

$ sudo rm -rf /etc/udev/rules.d/70-persistent-*

Registering with UEC

The last step would be to upload the images to UEC. The files that need to be uploaded are:

• vmlinuz-2.6.28-11-server
• initrd.img-2.6.28-11-server
• image.img

Copy the kernel and the initrd image from the VM image to some place outside. They will be used later for creating and uploading a complete virtual image to UEC.

 $ scp /boot/initrd.img-2.6.28-11-server user@A.B.C.D:
 $ scp /boot/vmlinuz-2.6.28-11-server user@A.B.C.D:

Before starting the upload process shut down the VM.

Registering kernel image

Execute the following commands to bundle and register the kernel image (vmlinuz-2.6.28-11-server)

$ euca-bundle-image -i vmlinuz-2.6.28-11-server --kernel true
 $ euca-upload-bundle -b mybucket -m /tmp/vmlinuz-2.6.28-11-server.manifest.xml
 $ euca-register mybucket/vmlinuz-2.6.28-11-server.manifest.xml

Save the output produced by the last command above (eki-XXXXXXXX), which will be needed while registering the disk image.

Registering ramdisk image

Execute the following commands to bundle and register the ramdisk image (initrd.img-2.6.28-11-server)

$ euca-bundle-image -i  initrd.img-2.6.28-11-server
 $ euca-upload-bundle -b mybucket -m /tmp/initrd.img-2.6.28-11-server.manifest.xml
 $ euca-register mybucket/initrd.img-2.6.28-11-server.manifest.xml

Save the output produced by the last command above (eri-XXXXXXXX), which will be needed while registering the disk image.

Registering disk image

Execute the following commands to bundle and register the ramdisk image (image.img)

$ euca-bundle-image -i image.img --kernel eki-XXXXXXXX --ramdisk eri-XXXXXXXX
 $ euca-upload-bundle -b mybucket -m /tmp/image.img.manifest.xml
 $ euca-register mybucket/image.img.manifest.xml

Replace eki-XXXXXXXX and eri-XXXXXXXX with the exact values as noted down earlier.

Image Listing

The new images that have been uploaded can be seen by using euca-describe-images command.

$ euca-describe-images
 IMAGE   emi-70B70EC0    mybucket/image.img.manifest.xml admin   available   public   x86_64  machine
 IMAGE   eri-A2BE13EC    mybucket/initrd.img-2.6.28-11-server.manifest.xml   admin    available       public  x86_64  ramdisk
 IMAGE   eki-685F1306    mybucket/vmlinuz-2.6.28-11-server.manifest.xml  adminavailable       public  x86_64  kernel

Running instance of custom image

To launch a new instance of the custom image, execute euca-run-instances command.

$ euca-run-instances  emi-70B70EC0 -k mykey -t c1.xlarge

Credits: This post draws inspiration from Sam for all his endeavors on “Technology from Up there to Down here”