11 KiB
22.04 (Draft version in progress)
The 22.04 OpenStack Charms release includes updates for the charms
described on the ../project/openstack-charms page. As of this release,
the project consists of 62 (...CHANGE IF REQUIRED...) stable charms.
For the list of bugs resolved in this release refer to the 22.04 milestone in Launchpad.
For scheduling information of past and future releases see the ../project/release-schedule.
Note
Release Notes contents is superseded by updated information published in the OpenStack Charm Guide (this guide) after the release of any given OpenStack Charms version.
Important
Always upgrade to the latest stable charms before making any major changes to your cloud and before filing bug reports. Note that charm upgrades and OpenStack upgrades are functionally different. For instructions on performing the different upgrade types see Upgrades overview in the OpenStack Charms Deployment Guide.
Summary of changes:
New stable charms
<TITLE>
New stable charm features
With each new feature, there is a corresponding example bundle in the
form of a test bundle, and/or a section in the OpenStack
Charms Deployment Guide, that details its usage. Test bundles are
located in the src/tests/bundles directory of the relevant
charm repository (see all charm
repositories).
cinder
charm: new option to configure
scheduler_default_filters
The cinder charm has a new configuration option:
scheduler-default-filters. This option can be used to set
the value of scheduler_default_filters in the cinder
configuration.
glance charm: add support for Cinder iSCSI/FC backend storage
The glance charm adds support for Cinder iSCSI/FC backend storage through three new configuration options:
cinder-volume-types
Comma separated list of cinder volume_types that can be configured as cinder storage backends. The first one in the list will be configured as the default backend.
cinder-http-retries
Number of cinderclient retries on failed http calls.
cinder-state-transition-timeout
Time period of time in seconds to wait for a cinder volume transition to complete.
ironic-conductor charm: new option to configure use-ipmitool-retries
The ironic-conductor charm has a new configuration option:
use-ipmitool-retries. This option can be used to set the
value of use_ipmitool_retries in the ironic
configuration.
keystone charm: add get-admin-password action
The keystone charm has a new get-admin-password action.
This action retrieves the admin password of keystone service via juju
action.
nova-cloud-controller
charm: new option to configure scheduler-max-attempts
The nova-cloud-controller charm has a new configuration option:
scheduler-max-attempts. This option will set the
scheduler.max_attempts in the nova configuration.
This flag allows to increase the number of retries and hence hosts to schedule on and therefore allowing the successful scheduling of instances.
nova-compute charm: new option to add extra-repositories
The nova-compute charm has a new configuration option for adding
extra apt repositories to the compute node:
extra-repositories. This configuration option takes a
comma-delimited list of apt source repository spec entries to add as apt
package repositories. The valid values are those accepted by the
add-apt-repository command.
nova-compute charm: vTPM Enablement
The nova-compute charm now allows for installing and configuring for
virtual TPM (vTPM) support. The nova-compute charm has a new option
enable-vtpm that when set to True, will configure the nova
and libvirt services to be able to provide vTPM devices via swtpm.
This requires swtpm, swtpm-tools, and libtpm0 packages to be available for installation. These are available in the Ubuntu 22.04 release and are expected to be backported to Ubuntu 20.04. The OpenStack Charmers team has included a backport of these packages in the ppa:openstack-charmers/swtpm ppa.
neutron-api
charm has a new configuration option:
uplink-status-propagation
This option allows the PF (physical function) link status for an
OpenStack SR-IOV port on the host to be propagated to the VF (virtual
function) link status on a cloud instance. To enable this, set the
option to 'true' and assign attribute 'propagate_uplink_status' to the
SR-IOV port during its creation (via the flag
--enable-uplink-status-propagation).
Increase KeepAlive timeout for OpenStack API endpoints and dashboard
The HTTP KeepAlive timeout for OpenStack API endpoints and the Dashboard has been changed to 75 seconds. The previous timeout of 5 seconds (Apache's default) was causing unnecessary termination of client TCP connections, which was also affecting inter-service (OpenStack) communication.
neutron-api charm: new option to enable network-segment-range support
The neutron-api charm has a new option for enabling Neutron's
'Network Segment Range' service plugin:
enable-network-segment-range. It allows cloud operators to
dynamically manage network segment ranges through the Neutron API. For
more details, refer to Network
segment ranges in the upstream documentation.
nova-compute charm: new options to set block device timeout
The nova-compute charm has two new options:
block-device-allocate-retriesblock-device-allocate-retries-interval
These options configure the block device allocation timeout. The default values have been set at 300 and 3 respectively, resulting in an overall timeout of 15 minutes. The previous (inherited upstream) timeout of 3 minutes resulted in failures when dealing with large guest images (e.g. Windows or customised Linux).
nova-cloud-controller charm: new option to expose max_local_block_devices
The nova-cloud-controller charm has a new option:
max-local-block-devices. It exposes the upstream
max_local_block_devices flag. In particular, setting it to '0' will
forbid local block devices, effectively compelling users to request
volumes instead. For more information, refer to block
device mapping FAQ section in the upstream documentation.
nova-cloud-controller charm: new options to filter hosts by isolating aggregates
The nova-cloud-controller charm has three new options:
limit-tenants-to-placement-aggregateplacement-aggregate-required-for-tenantsenable-isolated-aggregate-filtering
These options are useful for limiting host aggregates to specific tenants. For more information, refer to Filtering hosts by isolating aggregates in the upstream documentation.
nova-compute charm: new option to expose reserved_host_disk_mb
The nova-compute charm has a new option:
reserved-host-disk. It takes into account available host
disk space when scheduling instances. It is similar to existing options
reserved-host-memory and
reserved-huge-pages.
openstack-dashboard charm: new options to allow for branding
The openstack-dashboard charm has three new options:
site-brandingsite-branding-linkhelp-url
These options are used to set Dashboard parameters that reflect the local environment.
Documentation updates
<TITLE>
New tech-preview charms
cinder-nimblestorage
The cinder-nimblestorage charm provides NimbleStorage storage backend support for the Openstack Cinder service. It is a subordinate charm used in conjunction with the cinder principal charm.
cinder-solidfire
The cinder-solidfire charm provides Solidfire storage backend support for the Openstack Cinder service. It is a subordinate charm used in conjunction with the cinder principal charm.
nova-compute-nvidia-vgpu
The nova-compute-nvidia-vgpu charm provides Nvidia vGPU support to the OpenStack Nova Compute service. It is a subordinate charm used in conjunction with the nova-compute principal charm.
New tech-preview charm features
<TITLE>
Informational notices
nova-cloud-controller: cpu-allocation-ratio default changed to 2
The default value for cpu-allocation-ratio has been reduced from 16 to 2. The old default was more appropriate for dev, test or lab type environments but is rarely suitable for clouds running production workloads. If you were relying on the previous default of 16 and start to see VM scheduling failures after upgrade of this charm, you can opt back into a higher contention ratio by running:
juju config nova-cloud-controller cpu-allocation-ratio=16