In the following list you can check the highlights of 4.0 organized by component (a detailed list of changes can be found here):

OpenNebula 4.0 brings multiple new features to manage virtual machines:

• VM System Snapshots, i.e. including the complete VM state (memory+disk). Snapshots can be listed and reverted freely for running VMs. These feature is supported for KVM and VMware (see more info here).
• Disk Snapshots, similar to the 3.8 save_as operation but to be performed live, so applications must guarantee FS integrity (more here).
• Capacity re sizing, to change the CPUs and memory allocated to a VM. These operation is only for off-line re-sizing (check this out).
• Programmable VM operations, so you want to suspend your VM Friday afternoon and resume it first thing on Monday. You can do this without intervention by programming the suspend and resume operations, see a description here.
• NIC hotplugging, need another network card in your VM, hot-plug a new one.
• IPv6 Support, extending OpenNebula's contextualization process you can now easily define site and global unicast addresses. For the sake of completeness the link local address is also generated, (more here).
• Plain Files Datastore, This is an special datastore to store kernels, ramdisks and files to be used in context. The whole set of OpenNebula features applies, e.g. ACLs, ownership… more information here.
• User defined VM tags now you can tag VMs with any information (key-value), this is useful to manage your VMs but also to store application specific data (VMs can call-back home), more in the VM guide
• Powerful CLI, the shell is also important for us, with OpenNebula 4.0 you can easily control your infrastructure, read this.
• Better VM operations. Now you can undeploy a running VM keeping the instance defined (holding any resource, IP or image) but without consuming any host capacity. The action names are now stable and in their final form. Note that we have change back vm destroy to vm delete to keep the compatibility and coherence with the other resources.
• And much more…: one-step template update & instantation, rename virtual resources, user umask, capacity checks for deployment operations…

There has been also several improvements for administrators and new features to enhance the robustness and scalability of OpenNebula core:

• Syslog support, for the oned daemon and the scheduler, read this to configure it.
• ACLs for clusters, now you can define fine-grain access to the resources of a cluster.
• Default Quotas, so you can limit resource usage by default, see this.
• New persistency model, VM disk images can now be “inmutable” to implement advanced transfer mechanisms.
• Efficient monitoring, we have greatly reduced the latency and overhead of monitoring, read at the end of this post.
• VM cancellation and cleanup synchronization, the previous destroy mechanism was to aggressive and sometimes images were not properly cleaned-up. The cancellation and clean-up procedure are now sync.
• Better concurrency for network drivers, to prevent race conditions when instantiating multiple VMs in the same host.

The UI interface has been redesigned with a fresher and modern look. You can see it in action in these screencasts, or check this screen-shot:

This is a new feature that allows you to customize the UI for different users or user groups, so the interface implements the provisioning model for each role. Sunstone views replaces the 3.x self-service portal. Sunstone views does not require any additional service (apart from the sunstone server). Sunstone ships with two predefined views admin and user, we may add a couple of more pre-defined views for the final release.

The back-end of OpenNebula has been also improved in several areas, as described below:

• CRL support, read this to add revocation list.
• Better DN management, there is no limitation now to the DN characters, also the authentication process has been fixed, read this if you are interested.

• Support for Xen 3 & Xen 4, better defaults are now for both Xen versions without the need of adapting them.
• New attributes: KVM includes a couple of new attributes, CACHE can be now set in disk attachments and the IO attribute can be defined for DISKS.
• More defaults: a default can be now defined for the raw attribute. Also VM supports a default NIC model.
• Custom attributes for VMware VMX files: VMware users need to define the guestOS attribute, this and any other data can be defined in the VMX files. See an example here.

• Improved iSCSI support, to improve the concurrency of the TGT server.
• Ceph Datastore, a brand new Datastore and Transfer Manager that uses the Ceph FS, read here if you are a Ceph user.

• Cluster attributes for VM placement, you can include cluster and host attributes in your requirements and rank expressions. This is a very easy way to tag all the hosts of a cluster with a given property, see here examples
• Affinity policies, There is a special variable, CURRENT_VMS, that can be used to deploy VMs in a Host where other VMs are (not) running, see here examples
• Any host attribute for scheduling expressions, if you need an even more flexible requirement or rank mechanism you can use Xpath to refer to any host attribute.
• Scheduler feedback through VM tags, the VM is not deployed?, now the scheduler will send you feedback through the VM properties.

Although OpenNebula 4.0 is a major release, we have kept compatibility at the API level. Legacy method names can be used at the XML-RPC and Ruby and JAVA bindings (it'd be a good idea to start using the new ones, though).

• Retrieve OpenNebula configuration: New methods to retrieve the running configuration of the OpenNebula daemon.

• Ruby gems: The Ruby bindings are now also packed as a gem, see here and here.

• Rackup and memcache: If you need to scale your public Cloud servers, they now support rack and memcache. If you are interested read this

There have been minor changes in the OpenNebula DB schema to solve bug related with large deployments, snapshots & generic data. These changes are automatically managed by the migration process, but if you have developed any customization you may need to update it.

A detailed upgrade process can be found in the documentation. For a complete set of changes to migrate from a 3.x installation please refer to the Compatibility Guide. You should also read this document if you are an OpenNebula 3.x user.

In few days we will release the first Service Pack of the OpenNebula 3 Series. After its first year, and for two additional years, the Project distributes a Service Pack (SP) for each major release that combines all the patches to easily install or deploy its last update. There may be up to three maintenance releases every year for each Service Pack. C12G Labs is committed to supporting all users that install latest Services Packs.

In few days OpenNebulaApps 4.0 will be also available for download including AppStage, AppFlow and AppMarket for your OpenNebula 4.0 cloud, please stay tuned!. These tools will be incorporated into the main distribution of OpenNebula from version 4.2.

OpenNebula is released under the Apache 2.0 open source license. The complete source tree and binary packages for OpenNebula can be downloaded here.

Please report any bug or send feedback at the development portal or at the mailing list.

The documentation of OpenNebula 4.0 can be found here.

Because OpenNebula is inherently portable to different operating systems and virtualization platforms, most Linux distributions and Hypervisors are supported. However, not all platform configurations and combinations exhibit a similar functionality, performance and stability. You can contact us if you need advise about the best platform configurations and environments for functionality and performance. Please read our Certification Policy for more information.

This is the list of the individual platform components that have been through the complete OpenNebula Quality Assurance and Certification Process.

The OpenNebula project would like to thank the community members and users who have contributed to this software release by being active with the discussions, answering user questions, or providing patches for bugfixes, features and documentation, and especially to and especially to Produban - Banco Santander, China Mobile, Blackberry, GlobalRap, RedIris, Schneider Electric - Telvent, Qingye Jiang (Eucalyptus), Iftikhar Rathore (CISCO), Joel Merrick (BBC), Bill Campbell (Axcess Finantial), Patrick McGarry (Ceph), and Karanbir Singh (CentOS), Hyun Woo Kim (FermiLab), Márk Gergely, Mihály Héder and Tamás Marlok (MTA SZTAKI), Simon Boulet, Vlastimil Holer, Jon A., Jean-Philippe Garcia Ballester, Matthew Richardson and Dmitri Chebotarov. OpenNebula's Development Portal provides more details about the specific contributions. We would also like to thank all the people that have contributed translations for Sunstone.

More information about the project can be found at the project web page. You may be also interested in checking the OpenNebula Ecosystem that includes many interesting projects contributed by the community to enhance or add new features to OpenNebula.