FileDriverSpec

Summary

Implementation of a file based volume driver to manage images created by and creating with qemu-img, for example on a mounted NFS export.

storage is already mounted on the system, it should not be mounted by the driver
storage is readable and writable by the nova-volume process
storage is mounted, readable and writable on all compute nodes
- for example by using NFS or OCFS2
it's possible to use more than one mountpoint to storage the images
it's possible to specify the used format

check if the specified storage is correctly mounted and writable before creating a new image
create a new images with qemu-img
remove images with rm
- mabye better to use a Python library here
check needed free storage on the storage with df
- mabye better to use a Python library here
add a new FLAG to set the default format
add a new FLAG to specify one or more mountpoints to be used to store new images
add the possibility to import already created images as a new volume
- useful for migrating and using existing images into/in the cloud
add the possibility to choose the used format per volume
add the possibility to choose the used mountpoint per volume
add the possibility to move one image from a mountpoint to an other one

add a new list in the dashboard to choose the format to use by newly created volume

create new method FileDriver() in nova/volume/driver.py implementing the points described unter Implementation

I don't know if it's possible at the momentan to use 2 volume drivers in one Nova setup
- if it's possible there has to be implemented a method to migrate from and to migrate to a file image

This need not be added or completed until the specification is nearing beta.