Ussuri Series Release Notes

21.1.1-5

Upgrade Notes

  • The default for [glance] num_retries has changed from 0 to 3. The option controls how many times to retry a Glance API call in response to a HTTP connection failure. When deploying Glance behind HAproxy it is possible for a response to arrive just after the HAproxy idle time. As a result, an exception will be raised when the connection is closed resulting in a failed request. By increasing the default value, Nova can be more resilient to this scenario were HAproxy is misconfigured by retrying the request.

Bug Fixes

  • Add support for the hw:hide_hypervisor_id extra spec. This is an alias for the hide_hypervisor_id extra spec, which was not compatible with the AggregateInstanceExtraSpecsFilter scheduler filter. See bug 1841932 for more details.

21.1.1

Bug Fixes

  • Since Libvirt v.1.12.0 and the introduction of the libvirt issue , there is a fact that if we set cache mode whose write semantic is not O_DIRECT (i.e. “unsafe”, “writeback” or “writethrough”), there will be a problem with the volume drivers (i.e. LibvirtISCSIVolumeDriver, LibvirtNFSVolumeDriver and so on), which designate native io explicitly.

    When the driver_cache (default is none) has been configured as neither “none” nor “directsync”, the libvirt driver will ensure the driver_io to be “threads” to avoid an instance spawning failure.

  • An issue that could result in instances with the isolate thread policy (hw:cpu_thread_policy=isolate) being scheduled to hosts with SMT (HyperThreading) and consuming VCPU instead of PCPU has been resolved. See bug #1889633 for more information.

  • Resolve a race condition that may occur during concurrent interface detach/attach, resulting in an interface accidentally unbind after attached. See bug 1892870 for more details.

  • Addressed an issue that prevented instances using multiqueue feature from being created successfully when their vif_type is TAP.

  • Resolved an issue whereby providing an empty list for the policies field in the request body of the POST /os-server-groups API would result in a server error. This only affects the 2.1 to 2.63 microversions, as the 2.64 microversion replaces the policies list field with a policy string field. See bug #1894966 for more information.

21.1.0

Upgrade Notes

  • When using file-backed memory, the nova-compute service will now fail to start if the amount of reserved memory configured using [DEFAULT] reserved_host_memory_mb is equal to or greater than the total amount of memory configured using [libvirt] file_backed_memory. Where reserved memory is less than the total amount of memory configured, a warning will be raised. This warning will become an error in a future release.

    The former combination is invalid as it would suggest reserved memory is greater than total memory available, while the latter is considered incorrect behavior as reserving of file-backed memory can and should be achieved by reducing the filespace allocated as memory by modifying [libvirt] file_backed_memory.

Bug Fixes

  • A new [workarounds]/reserve_disk_resource_for_image_cache config option was added to fix the bug 1878024 where the images in the compute image cache overallocate the local disk. If this new config is set then the libvirt driver will reserve DISK_GB resources in placement based on the actual disk usage of the image cache.

  • Previously, attempting to configure an instance with the e1000e or legacy VirtualE1000e VIF types on a host using the QEMU/KVM driver would result in an incorrect UnsupportedHardware exception. These interfaces are now correctly marked as supported.

21.0.0

Prelude

The 21.0.0 release includes many new features and bug fixes. Please be sure to read the upgrade section which describes the required actions to upgrade your cloud from 20.0.0 (Train) to 21.0.0 (Ussuri).

There are a few major changes worth mentioning. This is not an exhaustive list:

New Features

  • Handling accelerator requests for an instance is now supported (where supported by the underlying virt driver) as of microversion 2.82. The Cyborg service generates an event for the binding completion for each accelerator request (ARQ) for an instance. Adds a new event accelerator_request_bound for this to the API POST /os-server-external-events

    The lists of operations that are supported or unsupported for instances with accelerators are listed in accelerator operation guide

  • The libvirt driver now supports live migration with virtual persistent memory (vPMEM), which requires QEMU as hypervisor. In virtualization layer, QEMU will copy vpmem over the network like volatile memory, due to the typical large capacity of vPMEM, it may takes longer time for live migration.

  • Allow the following filter parameters for GET /servers/detail and GET /servers for non-admin in microversion 2.83:

    • availability_zone

    • config_drive

    • key_name

    • created_at

    • launched_at

    • terminated_at

    • power_state

    • task_state

    • vm_state

    • progress

    • user_id

  • With microversion 2.84 the GET /servers/{server_id}/os-instance-actions/{request_id} API returns a details parameter for each failed event with a fault message, similar to the server fault.message parameter in GET /servers/{server_id} for a server with status ERROR.

  • Microversion 2.80 changes the list migrations APIs and the os-migrations API.

    In this microversion, expose the user_id and project_id fields in the following APIs:

    • GET /os-migrations

    • GET /servers/{server_id}/migrations

    • GET /servers/{server_id}/migrations/{migration_id}

    The GET /os-migrations API will also have optional user_id and project_id query parameters for filtering migrations by user and/or project, for example:

    • GET /os-migrations?user_id=ef9d34b4-45d0-4530-871b-3fb535988394

    • GET /os-migrations?project_id=011ee9f4-8f16-4c38-8633-a254d420fd54

    • GET /os-migrations?user_id=ef9d34b4-45d0-4530-871b-3fb535988394&project_id=011ee9f4-8f16-4c38-8633-a254d420fd54

  • With microversion 2.85 add new API PUT /servers/{server_id}/os-volume_attachments/{volume_id} which support for specifying delete_on_termination field in the request body to re-config the attached volume whether to delete when the instance is deleted.

  • The Nova policies implemented the scope concept and new default roles (admin, member, and reader) provided by keystone.

  • LXC instances now support cloud-init.

  • The 2.86 microversion adds support for flavor extra spec validation when creating or updating flavor extra specs. Use of an unrecognized or invalid flavor extra spec in the following namespaces will result in a HTTP 400 response.

    • accel

    • aggregate_instance_extra_specs

    • capabilities

    • hw

    • hw_rng

    • hw_video

    • os

    • pci_passthrough

    • powervm

    • quota

    • resources (including _{group} suffixes)

    • trait (including _{group} suffixes)

    • vmware

  • A new policy rule os_compute_api:servers:show:host_status:unknown-only has been added to control whether a user can view a server host status of UNKNOWN in the following APIs:

    • GET /servers/{server_id} if using API microversion >= 2.16

    • GET /servers/detail if using API microversion >= 2.16

    • PUT /servers/{server_id} if using API microversion >= 2.75

    • POST /servers/{server_id}/action (rebuild) if using API microversion >= 2.75

    This is different than the os_compute_api:servers:show:host_status policy rule which controls whether a user can view all possible host status in the aforementioned APIs including UP, DOWN, MAINTENANCE, and UNKNOWN.

  • A new image metadata prefilter has been added to allow translation of hypervisor-specific device model requests to standard traits. When this feature is enabled, nova is able to utilize placement to select hosts that are capable of emulating the requested devices, avoiding hosts that could not support the request. This feature is currently supported by the libvirt driver and can be enabled by configuring the [scheduler]/image_metadata_prefilter to True in the controller nova.conf.

  • Image pre-caching on hosts by aggregate is now supported (where supported by the underlying virt driver) as of microversion 2.81. A group of hosts within an aggregate can be compelled to fetch and cache a list of images to reduce time-to-boot latency. Adds the new API:

    • POST /os-aggregates/{aggregate_id}/images

    which is controlled by the policy compute:aggregates:images rule.

    See the [image_cache]/precache_concurrency config option for more information about throttling this operation.

  • Add --cell option to the nova-manage placement heal_allocations command. This option allows healing instance allocations within a specific cell.

  • With microversion 2.85, existing policy os-volumes-attachments:update is used for updating the resources with the change in its default value from SYSTEM_ADMIN to PROJECT_MEMBER_OR_SYSTEM_ADMIN. New policy os-volumes-attachments:swap is introduced for swapping the attachment of servers with default to SYSTEM_ADMIN.

  • Added support for instance-level PCI NUMA policies using the hw:pci_numa_affinity_policy flavor extra spec and hw_pci_numa_affinity_policy image metadata property. These apply to both PCI passthrough and SR-IOV devices, unlike host-level PCI NUMA policies configured via the alias key of the [pci] alias config option. See the VM Scoped SR-IOV NUMA Affinity spec for more info.

  • Stable device rescue for boot from volume instances is now supported through the use of the 2.87 microversion when the compute hosting the instance also reports the COMPUTE_RESCUE_BFV trait such as the libvirt driver.

    No changes have been made to the request or reponse parameters of the rescue API itself.

  • The server evacuate, os-migrateLive and unshelve action APIs now support servers with neutron ports having resource requests, e.g. ports that have QoS minimum bandwidth rules attached.

  • When using the libvirt driver, Nova instances will now get a VirtIO-RNG (Random Number Generator) device by default. This is to ensure guests are not starved of entropy during boot time. In case you want to disallow setting an RNG device for some reason, it can be done by setting the flavor Extra Spec property hw_rng:allowed to False.

Known Issues

  • In prior releases, an attempt to boot an instance directly from an image that was created by the Block Storage Service from an encrypted volume resulted in the instance going ACTIVE but being unusable. If a user then performed the image-create action on such an instance, the new image would inherit the cinder_encryption_key_id and, beginning with the 20.0.0 (Train) release, the cinder_encryption_key_deletion_policy image properties, assuming these were not included in the non_inheritable_image_properties configuration option. (The default setting for that option does not include these.) Beginning with 20.0.0 (Train), when the new image was deleted, the encryption key for the original image would be deleted, thereby rendering it unusable for the normal workflow of creating a volume from the image and booting an instance from the volume. Beginning with this release:

    • The Compute API will return a 400 (Bad Request) response to a request to directly boot an image created from an encrypted volume.

    • The image properties cinder_encryption_key_id and cinder_encryption_key_deletion_policy are absolutely non-inheritable regardless of the non_inheritable_image_properties setting.

Upgrade Notes

  • The non_inheritable_image_properties configuration option inhibits the transfer of image properties from the image an instance was created from to images created from that instance. There are, however, image properties (for example, the properties used for image signature validation) that should never be transferred to an instance snapshot. Prior to this release, such properties were included in the default setting for this configuration option, but this allowed the possibility that they might be removed by mistake, thereby resulting in a poor user experience. To prevent that from happening, nova now maintains an internal list of image properties that are absolutely non-inheritable regardless of the setting of the configuration option. See the help text for non_inheritable_image_properties in the sample nova configuration file for details.

  • All the policies except the deprecated APIs policy have been changed to implement the scope_type and new defaults. Deprecated APIs policy will be moved to scope_type and new defaults in the next release.

    Please refer Policy New Defaults for detail about policy new defaults and migration plan.

    • Scope

      Each policy is protected with appropriate scope_type. Nova support two types of sope_type with their combination. ['system'], ['project'] and ['system', 'project'].

      To know each policy scope_type, please refer the Policy Reference

      This feature is disabled by default can be enabled via config option [oslo_policy]enforce_scope in nova.conf

    • New Defaults(Admin, Member and Reader)

      Policies are default to Admin, Member and Reader roles. Old roles are also supproted. You can switch to new defaults via config option [oslo_policy]enforce_new_defaults in nova.conf file.

    • Policies granularity

      To implement the reader roles, Below policies are made more granular

      • os_compute_api:os-agents is made granular to

        • os_compute_api:os-agents:create

        • os_compute_api:os-agents:update

        • os_compute_api:os-agents:delete

        • os_compute_api:os-agents:list

      • os_compute_api:os-attach-interfaces is made granular to

        • os_compute_api:os-attach-interfaces:create

        • os_compute_api:os-attach-interfaces:delete

        • os_compute_api:os-attach-interfaces:show

        • os_compute_api:os-attach-interfaces:list

      • os_compute_api:os-deferred-delete is made granular to

        • os_compute_api:os-deferred-delete:restore

        • os_compute_api:os-deferred-delete:force

      • os_compute_api:os-hypervisors is made granular to

        • os_compute_api:os-hypervisors:list

        • os_compute_api:os-hypervisors:list-detail

        • os_compute_api:os-hypervisors:statistics

        • os_compute_api:os-hypervisors:show

        • os_compute_api:os-hypervisors:uptime

        • os_compute_api:os-hypervisors:search

        • os_compute_api:os-hypervisors:servers

      • os_compute_api:os-security-groups is made granular to

        • os_compute_api:os-security-groups:add

        • os_compute_api:os-security-groups:remove

        • os_compute_api:os-security-groups:list

      • os_compute_api:os-instance-usage-audit-log is made granular to

        • os_compute_api:os-instance-usage-audit-log:list

        • os_compute_api:os-instance-usage-audit-log:show

      • os_compute_api:os-instance-actions is made granular to

        • os_compute_api:os-instance-actions:list

        • os_compute_api:os-instance-actions:show

      • os_compute_api:os-server-password is made granular to

        • os_compute_api:os-server-password:show

        • os_compute_api:os-server-password:clear

      • os_compute_api:os-rescue is made granular to

        • os_compute_api:os-rescue

        • os_compute_api:os-unrescue

      • os_compute_api:os-used-limits is renamed to

        • os_compute_api:limits:other_project

      • os_compute_api:os-services is made granular to

        • os_compute_api:os-services:list

        • os_compute_api:os-services:update

        • os_compute_api:os-services:delete

  • Nova policies implemented the scope_type and new defaults provided by keystone. Old defaults are deprecated and still work if rules are not overridden in the policy file. If you don’t override any policies at all, then you don’t need to do anything different until the W release when old deprecated rules are removed and tokens need to be scoped to work with new defaults and scope of policies. For migration to new policies you can refer to this document.

    If you are overwriting the policy rules (all or some of them) in the policy file with new default values or any new value that requires scoped tokens, then non-scoped tokens will not work. Also if you generate the policy file with ‘oslopolicy-sample-generator’ json format or any other tool, you will get rules defaulted in the new format, which examines the token scope. Unless you turn on oslo_policy.enforce_scope, scope-checking rules will fail. Thus, be sure to enable oslo_policy.enforce_scope and educate end users on how to request scoped tokens from Keystone, or use a pre-existing sample config file from the Train release until you are ready to migrate to scoped policies. Another way is to generate the policy file in yaml format as described here and update the policy.yaml location in oslo_policy.policy_file.

    For more background about the possible problem, check this bug. A upgrade check has been added to the nova-status upgrade check command for this.

  • Python 2.7 support has been dropped. The minimum version of Python now supported by nova is Python 3.6.

  • Upgrading to Train on a deployment with a large database may hit bug 1862205, which results in instance records left in a bad state, and manifests as instances not being shown in list operations. Users upgrading to Train for the first time will definitely want to apply a version which includes this fix. Users already on Train should upgrade to a version including this fix to ensure the problem is addressed.

  • Starting in the Ussuri release, compute node resource providers are automatically marked with the COMPUTE_NODE trait. This allows them to be distinguished easily from other providers, including sharing and nested providers, as well as other non-compute-related providers in a deployment. To make effective use of this trait (e.g. for scheduling purposes), all compute nodes must be upgrade to Ussuri. Alternatively, you can manually add the trait to pre-Ussuri compute node providers via openstack resource provider trait set

  • The [osapi_v21]/project_id_regex configuration option which has been deprecated since the Mitaka 13.0.0 release has now been removed.

  • The nova-console service has been deprecated since the 19.0.0 Stein release and has now been removed. The following configuration options are therefore removed.

    • [upgrade_levels] console

    In addition, the following APIs have been removed. Calling these APIs will now result in a 410 HTTPGone error response:

    • POST /servers/{server_id}/consoles

    • GET /servers/{server_id}/consoles

    • GET /servers/{server_id}/consoles/{console_id}

    • DELETE /servers/{server_id}/consoles/{console_id}

    Finally, the following policies are removed. These were related to the removed APIs listed above and no longer had any effect:

    • os_compute_api:os-consoles:index

    • os_compute_api:os-consoles:create

    • os_compute_api:os-consoles:delete

    • os_compute_api:os-consoles:show

  • The nova-network feature has been deprecated since the 14.0.0 (Newton) release and has now been removed. The remaining nova-network specific REST APIs have been removed along with their related policy rules. Calling these APIs will now result in a 410 (Gone) error response.

    • GET /os-security-group-default-rules

    • POST /os-security-group-default-rules

    • GET /os-security-group-default-rules/{id}

    • DELETE /os-security-group-default-rules/{id}

    • POST /os-networks

    • DELETE /os-networks

    • POST /os-networks/add

    • POST /os-networks/{id} (associate_host)

    • POST /os-networks/{id} (disassociate)

    • POST /os-networks/{id} (disassociate_host)

    • POST /os-networks/{id} (disassociate_project)

    • POST /os-tenant-networks

    • DELETE /os-tenant-networks

    The following policies have also been removed.

    • os_compute_api:os-security-group-default-rules

    • os_compute_api:os-networks

    • os_compute_api:os-networks-associate

  • The networks quota, which was only enabled if the enabled_network_quota config option was enabled and only useful with nova-network, is removed. It will not longer be present in the responses for the APIs while attempts to update the quota will be rejected.

    • GET /os-quota-sets

    • GET /os-quota-sets/{project_id}

    • GET /os-quota-sets/{project_id}/defaults

    • GET /os-quota-sets/{project_id}/detail

    • PUT /os-quota-sets/{project_id}

    • GET /os-quota-class-sets/{id}

    • PUT /os-quota-class-sets/{id}

    The following related config options have been removed.

    • enable_network_quota

    • quota_networks

  • The following nova-manage commands have been removed.

    • network

    • floating

    These were only useful for the now-removed nova-network service and have been deprecated since the 15.0.0 (Ocata) release.

  • The nova-dhcpbridge service has been removed. This was only used with the now-removed nova-network service.

  • The following config options only applied when using the nova-network network driver which has now been removed. The config options have therefore been removed also.

    • [DEFAULT] firewall_driver

    • [DEFAULT] allow_same_net_traffic

    • [DEFAULT] flat_network_bridge

    • [DEFAULT] flat_network_dns

    • [DEFAULT] flat_interface

    • [DEFAULT] vlan_interface

    • [DEFAULT] vlan_start

    • [DEFAULT] num_networks

    • [DEFAULT] vpn_ip

    • [DEFAULT] vpn_start

    • [DEFAULT] network_size

    • [DEFAULT] fixed_range_v6

    • [DEFAULT] gateway

    • [DEFAULT] gateway_v6

    • [DEFAULT] cnt_vpn_clients

    • [DEFAULT] fixed_ip_disassociate_timeout

    • [DEFAULT] create_unique_mac_address_attempts

    • [DEFAULT] teardown_unused_network_gateway

    • [DEFAULT] l3_lib

    • [DEFAULT] network_driver

    • [DEFAULT] network_manager

    • [DEFAULT] multi_host

    • [DEFAULT] force_dhcp_release

    • [DEFAULT] update_dns_entries

    • [DEFAULT] dns_update_periodic_interval

    • [DEFAULT] dhcp_domain

    • [DEFAULT] use_neutron

    • [DEFAULT] auto_assign_floating_ip

    • [DEFAULT] floating_ip_dns_manager

    • [DEFAULT] instance_dns_manager

    • [DEFAULT] instance_dns_domain

    • [DEFAULT] default_floating_pool

    • [DEFAULT] ipv6_backend

    • [DEFAULT] metadata_host

    • [DEFAULT] metadata_port

    • [DEFAULT] iptables_top_regex

    • [DEFAULT] iptables_bottom_regex

    • [DEFAULT] iptables_drop_action

    • [DEFAULT] ldap_dns_url

    • [DEFAULT] ldap_dns_user

    • [DEFAULT] ldap_dns_password

    • [DEFAULT] ldap_dns_soa_hostmaster

    • [DEFAULT] ldap_dns_servers

    • [DEFAULT] ldap_dns_base_dn

    • [DEFAULT] ldap_dns_soa_refresh

    • [DEFAULT] ldap_dns_soa_retry

    • [DEFAULT] ldap_dns_soa_expiry

    • [DEFAULT] ldap_dns_soa_minimum

    • [DEFAULT] dhcpbridge_flagfile

    • [DEFAULT] dhcpbridge

    • [DEFAULT] dhcp_lease_time

    • [DEFAULT] dns_server

    • [DEFAULT] use_network_dns_servers

    • [DEFAULT] dnsmasq_config_file

    • [DEFAULT] ebtables_exec_attempts

    • [DEFAULT] ebtables_retry_interval

    • [DEFAULT] fake_network

    • [DEFAULT] send_arp_for_ha

    • [DEFAULT] send_arp_for_ha_count

    • [DEFAULT] dmz_cidr

    • [DEFAULT] force_snat_range

    • [DEFAULT] linuxnet_interface_driver

    • [DEFAULT] linuxnet_ovs_integration_bridge

    • [DEFAULT] use_single_default_gateway

    • [DEFAULT] forward_bridge_interface

    • [DEFAULT] ovs_vsctl_timeout

    • [DEFAULT] networks_path

    • [DEFAULT] public_interface

    • [DEFAULT] routing_source_ip

    • [DEFAULT] use_ipv6

    • [DEFAULT] allow_same_net_traffic

    • [DEFAULT] defer_iptables_apply

    • [DEFAULT] share_dhcp_address

    • [upgrade_levels] network

    • [vmware] vlan_interface

  • The nova-xvpvncproxy service has been deprecated since the 19.0.0 Stein release and has now been removed. The following configuration options have also been removed:

    • [vnc] xvpvncproxy_base_url

    • [vnc] xvpvncproxy_host

    • [vnc] xvpvncproxy_port

    • [xvp] console_xvp_conf_template

    • [xvp] console_xvp_conf

    • [xvp] console_xvp_log

    • [xvp] console_xvp_multiplex_port

    • [xvp] console_xvp_pid

  • New paste pipelines and middleware have been created to allow API version discovery to be performed without authentication or redirects. Because this involves an api-paste.ini change, you will need to manually update your api-paste.ini with the one from the release to get this functionality.

  • Compatibility code for compute drivers that do not implement the update_provider_tree interface has been removed. All compute drivers must now implement update_provider_tree.

Deprecation Notes

  • During Policy new defaults, below policies are deprecated and will be removed in 23.0.0 release. These are replaced by the new granular policies listed in feature section.

    • os_compute_api:os-agents

    • os_compute_api:os-attach-interfaces

    • os_compute_api:os-deferred-delete

    • os_compute_api:os-hypervisors

    • os_compute_api:os-security-groups

    • os_compute_api:os-instance-usage-audit-log

    • os_compute_api:os-instance-actions

    • os_compute_api:os-server-password

    • os_compute_api:os-used-limits

    • os_compute_api:os-services

  • The [api]auth_strategy conf option and the corresponding test-only noauth2 pipeline in api-paste.ini are deprecated and will be removed in a future release. The only supported auth_strategy is keystone, the default.

  • The [glance]api_servers configuration option is deprecated and will be removed in a future release. Deployments should use standard keystoneauth1 options to configure communication with a single image service endpoint. Any load balancing or high availability requirements should be satisfied outside of nova.

  • The [scheduler] driver config option has been deprecated. This was previously used to switch between different scheduler drivers including custom, out-of-tree ones. However, only the FilterScheduler has been supported in-tree since 19.0.0 (Stein) and nova increasingly relies on placement for basic functionality, meaning developing and maintaining out-of-tree drivers is increasingly difficult. Users who still rely on a custom scheduler driver should migrate to the filter scheduler, using custom filters and weighters where necessary.

  • The vmwareapi driver is deprecated in this release and may be removed in a future one. The driver is not tested by the OpenStack Nova project and does not have a clear maintainer.

  • The following conf options have been moved to the [image_cache] group and renamed accordingly. The old option paths are deprecated and will be removed in a future release.

    Deprecated Option

    New Option

    [DEFAULT]image_cache_manager_interval

    [image_cache]manager_interval

    [DEFAULT]image_cache_subdirectory_name

    [image_cache]subdirectory_name

    [DEFAULT]remove_unused_base_images

    [image_cache]remove_unused_base_images

    [DEFAULT]remove_unused_original_minimum_age_seconds

    [image_cache]remove_unused_original_minimum_age_seconds

    [libvirt]remove_unused_resized_minimum_age_seconds

    [image_cache]remove_unused_resized_minimum_age_seconds

Bug Fixes

  • Long standing bug 1694844 is now fixed where the following conditions would lead to a 400 error response during server create:

    • [cinder]/cross_az_attach=False

    • [DEFAULT]/default_schedule_zone=None

    • A server is created without an availability zone specified but with pre-existing volume block device mappings

    Before the bug was fixed, users would have to specify an availability zone that matches the zone that the volume(s) were in. With the fix, the compute API will implicitly create the server in the zone that the volume(s) are in as long as the volume zone is not the same as the [DEFAULT]/default_availability_zone value (defaults to nova).

  • This release contains a fix for bug 1748697 which distinguishes between resize and cold migrate such that the allow_resize_to_same_host config option is no longer used to determine if a server can be cold migrated to the same compute service host. Now when requesting a cold migration the API will check if the source compute node resource provider has the COMPUTE_SAME_HOST_COLD_MIGRATE trait and if so the scheduler can select the source host. Note that the only in-tree compute driver that supports cold migration to the same host is VMwareVCDriver.

    To support rolling upgrades with N-1 computes if a node does not report the trait and is old the API will fallback to the allow_resize_to_same_host config option value. That compatibility will be removed in a subsequent release.

  • Prior to this release Nova determined if UEFI support should be enable solely by checking host support as reported in bug 1845628.

    Nova now correctly check for guest os support via the hw_firmware_type image metadata property when spawning new instance and only enables UEFI if the guest and host support it. Guest deletion has also been updated to correctly clean up based on the UEFI or BIOS configuration of the vm.

  • Bug 1845986 has been fixed by adding iommu driver when the following metadata options are used with AMD SEV:

    • hw_scsi_model=virtio-scsi and either hw_disk_bus=scsi or hw_cdrom_bus=scsi

    • hw_video_model=virtio

    Also a virtio-serial controller is created when hw_qemu_guest_agent=yes option is used, together with iommu driver for it.

  • This release contains a fix for a regression introduced in 15.0.0 (Ocata) where server create failing during scheduling would not result in an instance action record being created in the cell0 database. Now when creating a server fails during scheduling and is “buried” in cell0 a create action will be created with an event named conductor_schedule_and_build_instances.

  • The DELETE /os-services/{service_id} compute API will now return a 409 HTTPConflict response when trying to delete a nova-compute service which is involved in in-progress migrations. This is because doing so would not only orphan the compute node resource provider in the placement service on which those instances have resource allocations but can also break the ability to confirm/revert a pending resize properly. See https://bugs.launchpad.net/nova/+bug/1852610 for more details.

  • This release contains a fix for bug 1856925 such that resize and migrate server actions will be rejected with a 409 HTTPConflict response if the source compute service is down.

  • For AArch64 Nova now sets max as the default CPU model. It does the right thing in context of both QEMU TCG (plain emulation) and for KVM (hardware acceleration).

  • The Compute service has never supported direct booting of an instance from an image that was created by the Block Storage service from an encrypted volume. Previously, this operation would result in an ACTIVE instance that was unusable. Beginning with this release, an attempt to boot from such an image will result in the Compute API returning a 400 (Bad Request) response.

  • A fix for serious bug 1862205 is provided which addresses both the performance aspect of schema migration 399, as well as the potential fallout for cases where this migration silently fails and leaves large numbers of instances hidden from view from the API.

  • A new config option [neutron]http_retries is added which defaults to 3. It controls how many times to retry a Neutron API call in response to a HTTP connection failure. An example scenario where it will help is when a deployment is using HAProxy and connections get closed after idle time. If an incoming request tries to re-use a connection that is simultaneously being torn down, a HTTP connection failure will occur and previously Nova would fail the entire request. With retries, Nova can be more resilient in this scenario and continue the request if a retry succeeds. Refer to https://launchpad.net/bugs/1866937 for more details.

  • An instance can be rebuilt in-place with the original image or a new image. Instance resource usage cannot be altered during a rebuild. Previously Nova would have ignored the NUMA topology of the new image continuing to use the NUMA topology of the existing instance until a move operation was performed. As Nova did not explicitly guard against inadvertent changes to resource requests contained in a new image, it was possible to rebuild with an image that would violate this requirement; see bug #1763766 for details. This resulted in an inconsistent state as the instance that was running did not match the instance that was requested. Nova now explicitly checks if a rebuild would alter the requested NUMA topology of an instance and rejects the rebuild if so.

  • With the changes introduced to address bug #1763766, Nova now guards against NUMA constraint changes on rebuild. As a result the NUMATopologyFilter is no longer required to run on rebuild since we already know the topology will not change and therefore the existing resource claim is still valid. As such it is now possible to do an in-place rebuild of an instance with a NUMA topology even if the image changes provided the new image does not alter the topology which addresses bug #1804502.

  • When using the api-paste.ini from the release, version discovery requests without a trailing slash will no longer receive a 302 redirect to the corresponding URL with a trailing slash (e.g. a request for /v2.1 will no longer redirect to /v2.1/). Instead, such requests will respond with the version discovery document regardless of the presence of the trailing slash. See bug 1728732 for details.

  • When using the api-paste.ini from the release, requests to the versioned discovery endpoints (/v2.1 and /v2) no longer require authentication. When using the compute API through certain clients, such as openstacksdk, this eliminates an unnecessary additional query. See bug 1845530 for details.

Other Notes

  • Nova now has a config option called [workarounds]/never_download_image_if_on_rbd which helps to avoid pathological storage behavior with multiple ceph clusters. Currently, Nova does not support multiple ceph clusters properly, but Glance can be configured with them. If an instance is booted from an image residing in a ceph cluster other than the one Nova knows about, it will silently download it from Glance and re-upload the image to the local ceph privately for that instance. Unlike the behavior you expect when configuring Nova and Glance for ceph, Nova will continue to do this over and over for the same image when subsequent instances are booted, consuming a large amount of storage unexpectedly. The new workaround option will cause Nova to refuse to do this download/upload behavior and instead fail the instance boot. It is simply a stop-gap effort to allow unsupported deployments with multiple ceph clusters from silently consuming large amounts of disk space.

  • A new pair of ssl_ciphers and ssl_minimum_version configuration options have been introduced for use by the nova-novncproxy, nova-serialproxy, and nova-spicehtml5proxy services. These new options allow one to configure the allowed TLS ciphers and minimum protocol version to enforce for incoming client connections to the proxy services.

    This aims to address the issues reported in bug 1842149, where it describes that the proxy services can inherit insecure TLS ciphers and protocol versions from the compiled-in defaults of the OpenSSL library on the underlying system. The proxy services provided no way to override such insecure defaults with current day generally accepted secure TLS settings.

  • A new nova-manage command, placement audit, has been added. This can be used to identify and optionally remove compute allocations in placement that are no longer referenced by existing instances or migrations. These orphaned allocations typically occur due to race conditions during instance migration or removal and will result in capacity issues if not addressed. For more details on CLI usage, see the man page entry: https://docs.openstack.org/nova/latest/cli/nova-manage.html#placement

  • The resize and migrate server action APIs used to synchronously block until a destination host is selected by the scheduler. Those APIs now asynchronously return a response to the user before scheduling. The response code remains 202 and users can monitor the operation via the status and OS-EXT-STS:task_state fields on the server resource and also by using the os-instance-actions API. The most notable change is NoValidHost will not be returned in a 400 error response from the API if scheduling fails but that information is available via the instance actions API interface.

  • The nova libvirt virt driver supports creating instances with multi-queue virtio network interfaces. In previous releases nova has based the maximum number of virtio queue pairs that can be allocated on the reported kernel major version. It has been reported in bug #1847367 that some distros have backported changes from later major versions that make major version number no longer suitable to determine the maximum virtio queue pair count. A new config option has been added to the libvirt section of the nova.conf. When defined nova will now use the [libvirt]/max_queues option to define the max queues that can be configured, if undefined it will fallback to the previous kernel version approach.

  • The [workarounds]/disable_native_luksv1 configuration option has been introduced. This can be used by operators to workaround recently discovered performance issues found within the libgcrypt library used by QEMU when natively decrypting LUKSv1 encrypted disks. Enabling this option will result in the use of the legacy dm-crypt based os-brick provided encryptors.

    Operators should be aware that this workaround only applies when using the libvirt compute driver with attached encrypted Cinder volumes using the luks encryption provider. The luks2 encryption provider will continue to use the dm-crypt based os-brick encryptors regardless of what this configurable is set to.

    This workaround is temporary and will be removed during the W release once all impacted distributions have been able to update their versions of the libgcrypt library.

    Warning

    Operators must ensure no instances are running on the compute host before enabling this workaround. Any instances with encrypted LUKSv1 disks left running on the hosts will fail to migrate or stop after this workaround has been enabled.

  • The [workarounds]/rbd_volume_local_attach configuration option has been introduced. This can be used by operators to ensure RBD volumes are connected to compute hosts as block devices. This can be used with the [worarounds]/disable_native_luksv1 configuration option to workaround recently discovered performance issues found within the libgcrypt library used by QEMU when natively decrypting LUKSv1 encrypted disks.

    This workaround does not currently support extending attached volumes.

    This workaround is temporary and will be removed during the W release once all impacted distributions have been able to update their versions of the libgcrypt library.

    Warning

    Operators must ensure no instances are running on the compute host before enabling this workaround. Any instances with attached RBD volumes left running on the hosts will fail to migrate or stop after this workaround has been enabled.