What’s new in RPCO r13.0 Mitaka#

Major new releases of OpenStack such as Mitaka typically include many changes, enhancements, and new features. RPCO is a tested configuration of a subset of all available OpenStack services.

This page lists some of the significant upstream OpenStack changes and is provided for your awareness. It is not a statement of support. For more information about supported features and configurations, contact your Rackspace sales team or support specialist.

Block Storage service (cinder)#

  • Numerous new storage drivers have been added, and existing HP drivers have been rebranded to HPE.

    Note

    Existing configurations will continue to work with the legacy name.

  • Previously, the only way to remove volumes in error states from a consistency-group was to delete the consistency group and recreate it. Now you can delete the volumes in error and error_deleting states.

  • Cinder services now automatically downgrade RPCO messages so that they are understood by the oldest version in a deployment. Disabled and dead services are also considered.

  • Support for new granularity for policies has been added. This support enables a cloud operator to correctly manage policy for volume type operations.

Image Service (glance)#

  • Reauthentication with trusts has been implemented when the image status is updated in the registry after the image is uploaded. When an image upload takes a long time (more than the token expiration time), glance uses trusts to receive a new token and update the image status in the registry. This enables users to upload large images without increasing the token expiration time.

Orchestration (heat)#

  • A stack can now be searched for resources based on a resource’s name, status, type, action, ID, and physical resource ID. This feature is enabled in the REST API and CLI.

  • Template validation has been improved to ignore the given set of error codes. For example, Orchestration reports a template as invalid if it does not find any required OpenStack services in the cloud deployment.

  • A new immutable Boolean field has been added to the parameters section in a Heat Orchestration Template (HOT) template. This field gives template authors the ability to mark template parameters as immutable to restrict updating parameters that can have destructive effects on the application.

  • Support for the Networking service (neutron) LBaaS v2 has been added. These are unique for v2 and do not support or mix with existing v1 resources.

Dashboard (horizon)#

  • The experience of setting up a port when setting up an instance has been improved.

  • The Object Storage (swift) panel has been moved, rendering logic to the client-side by using AngularJS.

  • Various additions have been made to the Dashboard UI to support new functionality in Mitaka, for example, support for managing domains and projects when using the Identity service (keystone) v3.

  • You can now edit the metadata of an instance.

  • The launch instance wizard now follows the new workflow by default.

Identity service (keystone)#

  • Roles can now be optionally defined as domain specific. A domain-specific role can be assigned to a domain or to a project within its domain, and any subset of global roles that it implies will appear in a token scoped to the respective domain or project.

  • The Identity service now supports creating implied roles. Role inference rules can now be added to indicate when the assignment of one role implies the assignment of another.

  • Two configuration options have been added to the [resource] section to specify a special admin project: admin_project_domain_name and admin_project_name.

  • The Identity service now supports authenticating via Time-based One-time Password (TOTP).

Networking service (neutron)#

  • External networks can now be controlled by using the RBAC framework that was added in the Liberty release. This feature allows networks to be made available to specific tenants (as opposed to all tenants) to be used as an external gateway for routers and floating IPs.

  • Integration with an external DNS service has been added:

    • Floating IPs can have dns_name and dns_domain attributes.

    • Ports can have a dns_name attribute.

    • The network where a port is created can have a dns_domain attribute.

    • Floating IPs and ports will be published in an external DNS service if they have dns_name and dns_domain attributes.

  • The segment_mtu option of the ML2 configuration has been deprecated and replaced with the global_physnet_mtu option. The Networking service should be able to calculate the path_mtu value automatically including encapsulation overhead like VXLAN.

  • The SriovNicSwitchMechanismDriver driver now exposes a new VIF type hostdev_physical for ports with the vNIC type direct-physical (used for SR-IOV PF passthrough). This enables nova to provision PFs as neutron ports.

  • New description fields have been added to security group rules, nets, ports, routers, floating IP addresses, and subnet pools.

  • QoS bandwidth limiting has been added.

  • ML2 provisioned ports can now recover from a binding failed state.

  • DHCP and L3 Agent scheduling is now availability zone aware.

    • If a DHCP agent is assigned to an availability zone, the network will be hosted by the DHCP agent with the availability zone specified by the user.

    • If an L3 agent is assigned to an availability zone, the router will be hosted by the L3 agent with the availability zone specified by the user.

  • Border Gateway Protocol (BGP) announcements for public networks have been added.

    • Announcement of tenant prefixes and host routes for floating IPs via BGP

    • Announcement of tenant subnets via BGP using centralized neutron router gateway port as the next-hop

    • Announcement of floating IP host routes via BGP using the centralized neutron router gateway port as the next-hop

  • Neutron services now respond to SIGUSR2 signal by dumping valuable debug information to standard error output.

Compute service (nova)#

  • The Compute service now supports UEFI booting of instances.

  • A disk space scheduling filter is now available, which prefers Compute nodes with the most available disk space. By default, free disk space is given equal importance to available RAM. To increase the priority of free disk space in scheduling, increase the value of the disk_weight_multiplier option.

  • Two new list/show API methods for server migration have been added (in microversion 2.23).

    • The list method returns the in progress live-migraton information of a server.

    • The show method returns information about a specified in progress live migration of a server.

  • All code and tests for the Compute service’s EC2 and ObjectStore API support which was deprecated in Kilo have been completely removed in Mitaka and have been replaced by the new ec2-api project.

  • Libvirt qemu/kvm instances can now be attached via MacVtap in bridge mode to a network. VLAN and FLAT attachments are supported. Other attachments than compute are not supported.