OpenNebula Hybrid Driver
Cloud bursting is a model in which the local resources of a Private Cloud are combined with resources from remote Cloud providers. This driver offers the possibility to implement Cloud bursting by deploying Virtual Machines seamlessly on a remote OpenNebula installation.
- Cloud Bursting to a remote OpenNebula.
- Manage remote instances importing them as 'wild VMs'.
This add-on is compatible with:
|5.0||Works from the CLI and advanced Sunstone dialogs|
|5.2||Fully integrated in Sunstone|
To install this add-on, clone it and run install.sh:
sudo ./install.sh -u oneadmin -g oneadmin
You will need to add the following to
/etc/one/oned.conf: oned.conf for hybrid OpenNebula
There are two configuration files:
/etc/one/opennebula_driver.conf: Configures access to the remote OpenNebula: username, password, endpoint
/etc/one/opennebula_driver.default: Default values for templates deployed in a remote OpenNebula.
Configure the remote OpenNebula
Ask the remote OpenNebula administrator to create a new user for you. This user should have access to VM Templates ready to be instantiated.
Configure the local OpenNebula
/etc/one/opennebula_driver.conf to add a new region for the remote OpenNebula. You may also use the default section.
regions: remotehostname: user: remoteuser password: abc endpoint: http://remotehostname:2633/RPC2 # Capacity is taken from the user and group quotas of the remote # OpenNebula user. Alternatively, you can set a hard limit here capacity: # In percentage. 800 means 8 cpu cores. 0 to use remote quotas cpu: 0 # In MB. 0 to use remote quotas memory: 0
Then create a new Host with
vm drivers set to
Create a local hybrid VM Template
Your hybrid VM Template must contain this section. Set TEMPLATE_ID to the target VM Template ID in the remote OpenNebula.
PUBLIC_CLOUD=[ TEMPLATE_ID="123", TYPE="opennebula" ]
If this Template does not define a local disk and must be deployed only in the remote OpenNebula instance, add this requirement:
SCHED_REQUIREMENTS = "PUBLIC_CLOUD = YES"
To match the reported allocated Host resources with the actual usage in the remote OpenNebula, set the same CPU and MEMORY as the remote Template.
To contribute bug patches or new features, you can use the github Pull Request model. It is assumed that code and documentation are contributed under the Apache License 2.0.