The Argo-server is the glue which ties the clients to the Xen virtualisation software. It listens upon a socket and processes commands from connecting clients.

By default the server binds upon port 20203, this can be changed via the configuration file /etc/default/argo-server.

Once the software has been installed you should be able to start the server by running:

/etc/init.d/argo-server start

Failing that you can use the argo-server command to launch an instance.

Once the server is running you may connect to it remotely, either using telnet, or using one of the available clients.

To configure the allowed clients please edit the file /etc/argo-server/argo-server.conf - the comments within that file should make the format clear. If they do not please let me know.

Available Xen Instances

The server will allow different clients to work with any virtual Xen instances running upon your host. In order to detect Xen instances it looks for files inside the /etc/xen directory with a .cfg suffix.

For example it will assume that the file /etc/xen/test.my.flat.cfg is a Xen instance - and this file will be parsed so that the name will be displayed when a client attempts to list available instances.

The actual name of the configuration file is irrelevent - since the configuration file will be assumed to contain a name inside it, eg:

#
# /etc/xen/vm1.steve.org.uk.cfg
#
kernel = "/boot/xen-linux-2.6.12.6-xen"
memory = 128
name   = "vm1.steve.org.uk"
root   = "/dev/sda1 ro"
vif    = ...

In this example the instance will be called vm1.steve.org.uk since that is the value of the name = parameter.

To detect the IP address of this instance the server will run:

nslookup vm1.steve.org.uk

So if your clients fail to detect the IP address of static hosts please ensure either DNS resolution succeeds - or there is a static entry inside /etc/osts.

Notes

• The server must be started as root to invoke the xm commands.
• Adding support for sudo would be simple, but this might not be desirable in all cases.