The Storage system allows OpenNebula administrators and users to set up Images, which can be operative systems or data, to be used in Virtual Machines easily. These Images can be used by several Virtual Machines simultaneously, and also shared with other users.
If you want to customize the Storage in your system, visit the Storage subsystem documentation.
There are six different types of Images. Using the command
oneimage chtype, you can change the type of an existing Image.
For Virtual Machine disks:
OS: An bootable disk Image. Every VM template must define one DISK referring to an Image of this type.
CDROM: These Images are read-only data. Only one Image of this type can be used in each VM template.
DATABLOCK: A datablock Image is a storage for data. These Images can be created from previous existing data, or as an empty drive.
“File” types. Images of these types cannot be used as VM disks, and are listed in Sunstone under the Files tab:
KERNEL: A plain file to be used as kernel (VM attribute OS/KERNEL_DS).
RAMDISK: A plain file to be used as ramdisk (VM attribute OS/INITRD_DS).
CONTEXT: A plain file to be included in the context CD-ROM (VM attribute CONTEXT/FILES_DS).
KERNEL, RAMDISK and CONTEXT file Images can be registered only in File Datastores.
Some of the operations described below do not apply to KERNEL, RAMDISK and CONTEXT Images, in particular: clone and persistent.
||The Image file is being copied or created in the Datastore.|
||Image file is being copied or created in the Datastore, with VMs waiting for the operation to finish.|
||Image ready to be used.|
||Non-persistent Image used by at least one VM. It can still be used by other VMs.|
||Persistent Image is use by a VM. It cannot be used by new VMs.|
||Image disabled by the owner, it cannot be used by new VMs.|
||Error state, a FS operation failed. See the Image information with
||The Image is being deleted from the Datastore.|
||The Image is being cloned.|
This is the state diagram for persistent Images:
And the following one is the state diagram for non-persistent Images:
Users can manage their Images using the command line interface command
oneimage. The complete reference is here.
You can also manage your Images using Sunstone, selecting the Images tab. By default this tab is available in the
admin view, but not in the
The three types of Images can be created from an existing file, but for datablock Images you can specify a size and let OpenNebula create an empty Image in the Datastore.
If you want to create an OS Image, you need to prepare a contextualized virtual machine, and extract its disk.
Please read first the documentation about VM contextualization here.
Once you have a disk you want to register, you can upload it directly using Sunstone:
To register it from the command line you need to create a new image template, and submit it using the
oneimage create command.
The complete reference for the image template is here. This is how a sample template looks like:
cat ubuntu_img.one NAME = "Ubuntu" PATH = "/home/cloud/images/ubuntu-desktop/disk.0" TYPE = "OS" DESCRIPTION = "Ubuntu desktop for students."
You need to choose the Datastore where to register the new Image. To know the available datastores, use the
onedatastore list command. In a clean installation you will only have one datastores with type
onedatastore list ID NAME SIZE AVAIL CLUSTERS IMAGES TYPE DS TM STAT 0 system 145.2G 56% 0 0 sys - shared on 1 default 145.2G 56% 0 3 img fs shared on 2 files 145.2G 56% 0 0 fil fs ssh on
To submit the template, you just have to issue the command
oneimage create ubuntu_img.one --datastore default ID: 0
You can also create Images using just parameters in the
oneimage create call. The parameters to generate the Image are as follows:
||Name of the new Image|
||Description for the new Image|
||Type of the new Image: OS, CDROM, DATABLOCK, KERNEL, RAMDISK, CONTEXT|
||Tells if the Image will be persistent|
||Device prefix for the disk (eg. hd, sd, xvd or vd)|
||Device the disk will be attached to|
||Path of the Image file|
||Driver to use (raw, qcow2, tap:aio:...)|
||Type of the Image (BLOCK, CDROM or FILE)|
||Source to be used. Useful for not file-based Images|
||Size in MB. Used for DATABLOCK type|
To create the previous example Image you can do it like this:
oneimage create --datastore default --name Ubuntu --path /home/cloud/images/ubuntu-desktop/disk.0 \ --description "Ubuntu desktop for students."
You can use gz compressed image files when registering them in OpenNebula.
Limitations when Uploading Images from Sunstone¶
Image file upload to the server via the client browser is possible. The process is as follow:
- Step 1: The client uploads the whole image file to the server in a temporal file in the
tmpdirfolder specified in the configuration.
- Step 2: OpenNebula registers an Image setting the PATH to that temporal file.
- Step 3: OpenNebula copies the image file to the datastore.
- Step 4: The temporal file is deleted and the request returns successfully to the user (a message pops up indicating that Image was uploaded correctly).
Note that when file sizes become big (normally over 1GB), and depending on your hardware, it may take long to complete the copying in step 3. Since the upload request needs to stay pending until copying is successful (so it can delete the temp file safely), there might be Ajax timeouts and/or lack of response from the server. This may cause errors, or trigger re-uploads (which re-initiate the loading progress bar).
Existing Images can be cloned to a new one. This is useful to make a backup of an Image before you modify it, or to get a private persistent copy of an Image shared by other user. Note that persistent Images with snapshots cannot be cloned. In order to do so, the user would need to flatten it first, see the snapshots section for more information.
To clone an Image, execute
oneimage clone Ubuntu new_image
You can optionally clone the Image to a different Datastore. The new Datastore must be compatible with the current one, i.e. have the same DS_MAD drivers.
oneimage clone Ubuntu new_image --datastore new_img_ds
The Sunstone Images tab also contains a dialog for the clone operation:
Listing Available Images¶
You can use the
oneimage list command to check the available images in the repository.
oneimage list ID USER GROUP NAME DATASTORE SIZE TYPE PER STAT RVMS 0 oneadmin oneadmin ttylinux-vd default 200M OS No used 8 1 johndoe users my-ubuntu-disk- default 200M OS Yes used 1 2 alice testgroup customized-ubun default 200M OS Yes used 1
To get complete information about an Image, use
oneimage show, or list Images continuously with
Orphan images (i.e images not referenced by any template) can be shown with
oneimage orphans command.
Making Images Persistent¶
oneimage persistent and
oneimage nonpersistent commands to make your Images persistent or not.
A persistent Image saves back to the datastore the changes made inside the VM after it is shut down.
oneimage list ID USER GROUP NAME DATASTORE SIZE TYPE PER STAT RVMS 0 oneadmin oneadmin Ubuntu default 10G OS No rdy 0 oneimage persistent Ubuntu oneimage list ID USER GROUP NAME DATASTORE SIZE TYPE PER STAT RVMS 0 oneadmin oneadmin Ubuntu default 10G OS Yes rdy 0 oneimage nonpersistent 0 oneimage list ID USER GROUP NAME DATASTORE SIZE TYPE PER STAT RVMS 0 oneadmin oneadmin Ubuntu default 10G OS No rdy 0
Note that persistent Images with snapshots cannot be made non-persistent. In order to do so, the user would need to flatten it first, see the snapshots section for more information.
Managing Snapshots in Persistent Images¶
Persistent Images can have associated snapshots if the user created them during the life-cycle of VM that used the persistent Image. The following are operations that allow the user to manage these snapshots directly:
oneimage snapshot-revert <image_id> <snapshot_id>: The active state of the Image is overwritten by the specified snapshot. Note that this operation discards any unsaved data of the disk state.
oneimage snapshot-delete <image_id> <snapshot_id>: Deletes a snapshot. This operation is only allowed if the snapshot is not the active snapshot and if it has no children.
oneimage snapshot-flatten <image_id> <snapshot_id>: This operation effectively converts the Image to an Image without snapshots. The saved disk state of the Image is the state of the specified snapshot. It’s an operation similar to running
snapshot-revertand then deleting all the snapshots.
Images with snapshots cannot be cloned or made non-persistent. To run either of these operations the user would need to flatten the Image first.
How to Use Images in Virtual Machines¶
Assuming you have an OS Image called Ubuntu desktop with ID 1, you can use it in your virtual machine template as a DISK. When this machine is deployed, the first disk will be taken from the Datastore.
Images can be referred in a DISK in two different ways:
IMAGE_ID, using its ID as returned by the create operation
IMAGE, using its name. In this case the name refers to one of the Images owned by the user (names can not be repeated for the same user). If you want to refer to an IMAGE of other user you can specify that with
IMAGE_UID(by the uid of the user) or
IMAGE_UNAME(by the name of the user).
CPU = 1 MEMORY = 3.08 DISK = [ IMAGE_ID = 7 ] DISK = [ IMAGE = "Ubuntu", IMAGE_UNAME = "oneadmin" ] DISK = [ type = swap, size = 1024 ] NIC = [ NETWORK_ID = 1 ] NIC = [ NETWORK_ID = 0 ] # FEATURES=[ acpi="no" ] GRAPHICS = [ type = "vnc", listen = "220.127.116.11", port = "5902" ]
Once the VM is deployed and changes are made to its disk, you can save those changes in two different ways:
- Disk snapshots, a snapshot of the disk state is saved, you can later revert to this saved state.
- Disk save_as, the disk is copied to a new Image in the datastore. A new virtual machine can be started from it. The disk must be in a consistent state during the save_as operation (e.g. by unmounting the disk from the VM).
A detailed description of this process is described in section Virtual Machine Instances
How to Use File Images in Virtual Machines¶
KERNEL and RAMDISK¶
KERNEL and RAMDISK type Images can be used in the OS/KERNEL_DS and OS/INITRD_DS attributes of the VM template. See the complete reference for more information.
OS = [ KERNEL_DS = "$FILE[IMAGE=kernel3.6]", INITRD_DS = "$FILE[IMAGE_ID=23]", ROOT = "sda1", KERNEL_CMD = "ro console=tty1" ]