Chapter 9. nova

allow_resize_to_same_host = False

boolean value

Allow destination machine to match source for resize. Useful when testing in single-host environments. By default it is not allowed to resize to the same host. Setting this option to true will add the same host to the destination options. Also set to true if you allow the ServerGroupAffinityFilter and need to resize. For changes to this option to take effect, the nova-api service needs to be restarted.

arq_binding_timeout = 300

integer value

Timeout for Accelerator Request (ARQ) bind event message arrival.

Number of seconds to wait for ARQ bind resolution event to arrive. The event indicates that every ARQ for an instance has either bound successfully or failed to bind. If it does not arrive, instance bringup is aborted with an exception.

backdoor_port = None

string value

Enable eventlet backdoor. Acceptable values are 0, <port>, and <start>:<end>, where 0 results in listening on a random tcp port number; <port> results in listening on the specified port number (and not enabling backdoor if that port is in use); and <start>:<end> results in listening on the smallest unused port number within the specified range of port numbers. The chosen port is displayed in the service’s log file.

backdoor_socket = None

string value

Enable eventlet backdoor, using the provided path as a unix socket that can receive connections. This option is mutually exclusive with backdoor_port in that only one should be provided. If both are provided then the existence of this option overrides the usage of that option. Inside the path {pid} will be replaced with the PID of the current process.

block_device_allocate_retries = 60

integer value

The number of times to check for a volume to be "available" before attaching it during server create.

When creating a server with block device mappings where source_type is one of blank, image or snapshot and the destination_type is volume, the nova-compute service will create a volume and then attach it to the server. Before the volume can be attached, it must be in status "available". This option controls how many times to check for the created volume to be "available" before it is attached.

If the operation times out, the volume will be deleted if the block device mapping delete_on_termination value is True.

It is recommended to configure the image cache in the block storage service to speed up this operation. See https://docs.openstack.org/cinder/latest/admin/blockstorage-image-volume-cache.html for details.

Possible values:

Related options:

block_device_allocate_retries_interval = 3

integer value

Interval (in seconds) between block device allocation retries on failures.

This option allows the user to specify the time interval between consecutive retries. The block_device_allocate_retries option specifies the maximum number of retries.

Possible values:

Related options:

cert = self.pem

string value

Path to SSL certificate file.

Related options:

compute_driver = None

string value

Defines which driver to use for controlling virtualization.

Possible values:

compute_monitors = []

list value

A comma-separated list of monitors that can be used for getting compute metrics. You can use the alias/name from the setuptools entry points for nova.compute.monitors.* namespaces. If no namespace is supplied, the "cpu." namespace is assumed for backwards-compatibility.

Possible values:

config_drive_format = iso9660

string value

Config drive format.

Config drive format that will contain metadata attached to the instance when it boots.

Related options:

Deprecated since: 19.0.0

Reason: This option was originally added as a workaround for bug in libvirt, #1246201, that was resolved in libvirt v1.2.17. As a result, this option is no longer necessary or useful.

conn_pool_min_size = 2

integer value

The pool size limit for connections expiration policy

conn_pool_ttl = 1200

integer value

The time-to-live in sec of idle connections in the pool

console_host = <based on operating system>

string value

Console proxy host to be used to connect to instances on this host. It is the publicly visible name for the console host.

Possible values:

control_exchange = nova

string value

The default exchange under which topics are scoped. May be overridden by an exchange name specified in the transport_url option.

cpu_allocation_ratio = None

floating point value

Virtual CPU to physical CPU allocation ratio.

This option is used to influence the hosts selected by the Placement API by configuring the allocation ratio for VCPU inventory.

Possible values:

Related options:

daemon = False

boolean value

Run as a background process.

debug = False

boolean value

If set to true, the logging level will be set to DEBUG instead of the default INFO level.

default_access_ip_network_name = None

string value

Name of the network to be used to set access IPs for instances. If there are multiple IPs to choose from, an arbitrary one will be chosen.

Possible values:

default_availability_zone = nova

string value

Default availability zone for compute services.

This option determines the default availability zone for nova-compute services, which will be used if the service(s) do not belong to aggregates with availability zone metadata.

Possible values:

default_ephemeral_format = None

string value

The default format an ephemeral_volume will be formatted with on creation.

Possible values:

default_log_levels = ['amqp=WARN', 'amqplib=WARN', 'boto=WARN', 'qpid=WARN', 'sqlalchemy=WARN', 'suds=INFO', 'oslo.messaging=INFO', 'oslo_messaging=INFO', 'iso8601=WARN', 'requests.packages.urllib3.connectionpool=WARN', 'urllib3.connectionpool=WARN', 'websocket=WARN', 'requests.packages.urllib3.util.retry=WARN', 'urllib3.util.retry=WARN', 'keystonemiddleware=WARN', 'routes.middleware=WARN', 'stevedore=WARN', 'taskflow=WARN', 'keystoneauth=WARN', 'oslo.cache=INFO', 'oslo_policy=INFO', 'dogpile.core.dogpile=INFO', 'glanceclient=WARN', 'oslo.privsep.daemon=INFO']

list value

List of package logging levels in logger=LEVEL pairs. This option is ignored if log_config_append is set.

default_schedule_zone = None

string value

Default availability zone for instances.

This option determines the default availability zone for instances, which will be used when a user does not specify one when creating an instance. The instance(s) will be bound to this availability zone for their lifetime.

Possible values:

Related options:

disk_allocation_ratio = None

floating point value

Virtual disk to physical disk allocation ratio.

This option is used to influence the hosts selected by the Placement API by configuring the allocation ratio for DISK_GB inventory.

When configured, a ratio greater than 1.0 will result in over-subscription of the available physical disk, which can be useful for more efficiently packing instances created with images that do not use the entire virtual disk, such as sparse or compressed images. It can be set to a value between 0.0 and 1.0 in order to preserve a percentage of the disk for uses other than instances.

Possible values:

Related options:

enable_new_services = True

boolean value

Enable new nova-compute services on this host automatically.

When a new nova-compute service starts up, it gets registered in the database as an enabled service. Sometimes it can be useful to register new compute services in disabled state and then enabled them at a later point in time. This option only sets this behavior for nova-compute services, it does not auto-disable other services like nova-conductor, nova-scheduler, or nova-osapi_compute.

Possible values:

enabled_apis = ['osapi_compute', 'metadata']

list value

List of APIs to be enabled by default.

enabled_ssl_apis = []

list value

List of APIs with enabled SSL.

Nova provides SSL support for the API servers. enabled_ssl_apis option allows configuring the SSL support.

executor_thread_pool_size = 64

integer value

Size of executor thread pool when executor is threading or eventlet.

fatal_deprecations = False

boolean value

Enables or disables fatal status of deprecations.

flat_injected = False

boolean value

This option determines whether the network setup information is injected into the VM before it is booted. While it was originally designed to be used only by nova-network, it is also used by the vmware virt driver to control whether network information is injected into a VM. The libvirt virt driver also uses it when we use config_drive to configure network to control whether network information is injected into a VM.

force_config_drive = False

boolean value

Force injection to take place on a config drive

When this option is set to true config drive functionality will be forced enabled by default, otherwise users can still enable config drives via the REST API or image metadata properties. Launched instances are not affected by this option.

Possible values:

Related options:

force_raw_images = True

boolean value

Force conversion of backing images to raw format.

Possible values:

Related options:

graceful_shutdown_timeout = 60

integer value

Specify a timeout after which a gracefully shutdown server will exit. Zero value means endless wait.

heal_instance_info_cache_interval = 60

integer value

Interval between instance network information cache updates.

Number of seconds after which each compute node runs the task of querying Neutron for all of its instances networking information, then updates the Nova db with that information. Nova will never update it’s cache if this option is set to 0. If we don’t update the cache, the metadata service and nova-api endpoints will be proxying incorrect network data about the instance. So, it is not recommended to set this option to 0.

Possible values:

host = <based on operating system>

host domain value

Hostname, FQDN or IP address of this host.

Used as:

Must be valid within AMQP key.

Possible values:

initial_cpu_allocation_ratio = 4.0

floating point value

Initial virtual CPU to physical CPU allocation ratio.

This is only used when initially creating the computes_nodes table record for a given nova-compute service.

See https://docs.openstack.org/nova/latest/admin/configuration/schedulers.html for more details and usage scenarios.

Related options:

initial_disk_allocation_ratio = 1.0

floating point value

Initial virtual disk to physical disk allocation ratio.

This is only used when initially creating the computes_nodes table record for a given nova-compute service.

See https://docs.openstack.org/nova/latest/admin/configuration/schedulers.html for more details and usage scenarios.

Related options:

initial_ram_allocation_ratio = 1.0

floating point value

Initial virtual RAM to physical RAM allocation ratio.

This is only used when initially creating the computes_nodes table record for a given nova-compute service.

See https://docs.openstack.org/nova/latest/admin/configuration/schedulers.html for more details and usage scenarios.

Related options:

injected_network_template = $pybasedir/nova/virt/interfaces.template

string value

Path to /etc/network/interfaces template.

The path to a template file for the /etc/network/interfaces-style file, which will be populated by nova and subsequently used by cloudinit. This provides a method to configure network connectivity in environments without a DHCP server.

The template will be rendered using Jinja2 template engine, and receive a top-level key called interfaces. This key will contain a list of dictionaries, one for each interface.

Refer to the cloudinit documentation for more information:

https://cloudinit.readthedocs.io/en/latest/topics/datasources.html

Possible values:

Related options:

instance_build_timeout = 0

integer value

Maximum time in seconds that an instance can take to build.

If this timer expires, instance status will be changed to ERROR. Enabling this option will make sure an instance will not be stuck in BUILD state for a longer period.

Possible values:

instance_delete_interval = 300

integer value

Interval for retrying failed instance file deletes.

This option depends on maximum_instance_delete_attempts. This option specifies how often to retry deletes whereas maximum_instance_delete_attempts specifies the maximum number of retry attempts that can be made.

Possible values:

Related options:

`instance_format = [instance: %(uuid)s] `

string value

The format for an instance that is passed with the log message.

instance_name_template = instance-%08x

string value

Template string to be used to generate instance names.

This template controls the creation of the database name of an instance. This is not the display name you enter when creating an instance (via Horizon or CLI). For a new deployment it is advisable to change the default value (which uses the database autoincrement) to another value which makes use of the attributes of an instance, like instance-%(uuid)s. If you already have instances in your deployment when you change this, your deployment will break.

Possible values:

instance_usage_audit = False

boolean value

This option enables periodic compute.instance.exists notifications. Each compute node must be configured to generate system usage data. These notifications are consumed by OpenStack Telemetry service.

instance_usage_audit_period = month

string value

Time period to generate instance usages for. It is possible to define optional offset to given period by appending @ character followed by a number defining offset.

Possible values:

`instance_uuid_format = [instance: %(uuid)s] `

string value

The format for an instance UUID that is passed with the log message.

instances_path = $state_path/instances

string value

Specifies where instances are stored on the hypervisor’s disk. It can point to locally attached storage or a directory on NFS.

Possible values:

Related options:

internal_service_availability_zone = internal

string value

Availability zone for internal services.

This option determines the availability zone for the various internal nova services, such as nova-scheduler, nova-conductor, etc.

Possible values:

key = None

string value

SSL key file (if separate from cert).

Related options:

live_migration_retry_count = 30

integer value

Maximum number of 1 second retries in live_migration. It specifies number of retries to iptables when it complains. It happens when an user continuously sends live-migration request to same host leading to concurrent request to iptables.

Possible values:

log-config-append = None

string value

The name of a logging configuration file. This file is appended to any existing logging configuration files. For details about logging configuration files, see the Python logging module documentation. Note that when logging configuration files are used then all logging configuration is set in the configuration file and other logging configuration options are ignored (for example, log-date-format).

log-date-format = %Y-%m-%d %H:%M:%S

string value

Defines the format string for %%(asctime)s in log records. Default: %(default)s . This option is ignored if log_config_append is set.

log-dir = None

string value

(Optional) The base directory used for relative log_file paths. This option is ignored if log_config_append is set.

log-file = None

string value

(Optional) Name of log file to send logging output to. If no default is set, logging will go to stderr as defined by use_stderr. This option is ignored if log_config_append is set.

log_options = True

boolean value

Enables or disables logging values of all registered options when starting a service (at DEBUG level).

log_rotate_interval = 1

integer value

The amount of time before the log files are rotated. This option is ignored unless log_rotation_type is set to "interval".

log_rotate_interval_type = days

string value

Rotation interval type. The time of the last file change (or the time when the service was started) is used when scheduling the next rotation.

log_rotation_type = none

string value

Log rotation type.

logging_context_format_string = %(asctime)s.%(msecs)03d %(process)d %(levelname)s %(name)s [%(global_request_id)s %(request_id)s %(user_identity)s] %(instance)s%(message)s

string value

Format string to use for log messages with context. Used by oslo_log.formatters.ContextFormatter

logging_debug_format_suffix = %(funcName)s %(pathname)s:%(lineno)d

string value

Additional data to append to log message when logging level for the message is DEBUG. Used by oslo_log.formatters.ContextFormatter

logging_default_format_string = %(asctime)s.%(msecs)03d %(process)d %(levelname)s %(name)s [-] %(instance)s%(message)s

string value

Format string to use for log messages when context is undefined. Used by oslo_log.formatters.ContextFormatter

logging_exception_prefix = %(asctime)s.%(msecs)03d %(process)d ERROR %(name)s %(instance)s

string value

Prefix each line of exception output with this format. Used by oslo_log.formatters.ContextFormatter

logging_user_identity_format = %(user)s %(project)s %(domain)s %(system_scope)s %(user_domain)s %(project_domain)s

string value

Defines the format string for %(user_identity)s that is used in logging_context_format_string. Used by oslo_log.formatters.ContextFormatter

long_rpc_timeout = 1800

integer value

This option allows setting an alternate timeout value for RPC calls that have the potential to take a long time. If set, RPC calls to other services will use this value for the timeout (in seconds) instead of the global rpc_response_timeout value.

Operations with RPC calls that utilize this value:

Related options:

max_concurrent_builds = 10

integer value

Limits the maximum number of instance builds to run concurrently by nova-compute. Compute service can attempt to build an infinite number of instances, if asked to do so. This limit is enforced to avoid building unlimited instance concurrently on a compute node. This value can be set per compute node.

Possible Values:

max_concurrent_live_migrations = 1

integer value

Maximum number of live migrations to run concurrently. This limit is enforced to avoid outbound live migrations overwhelming the host/network and causing failures. It is not recommended that you change this unless you are very sure that doing so is safe and stable in your environment.

Possible values:

max_concurrent_snapshots = 5

integer value

Maximum number of instance snapshot operations to run concurrently. This limit is enforced to prevent snapshots overwhelming the host/network/storage and causing failure. This value can be set per compute node.

Possible Values:

max_local_block_devices = 3

integer value

Maximum number of devices that will result in a local image being created on the hypervisor node.

A negative number means unlimited. Setting max_local_block_devices to 0 means that any request that attempts to create a local disk will fail. This option is meant to limit the number of local discs (so root local disc that is the result of imageRef being used when creating a server, and any other ephemeral and swap disks). 0 does not mean that images will be automatically converted to volumes and boot instances from volumes - it just means that all requests that attempt to create a local disk will fail.

Possible values:

max_logfile_count = 30

integer value

Maximum number of rotated log files.

max_logfile_size_mb = 200

integer value

Log file maximum size in MB. This option is ignored if "log_rotation_type" is not set to "size".

maximum_instance_delete_attempts = 5

integer value

The number of times to attempt to reap an instance’s files.

This option specifies the maximum number of retry attempts that can be made.

Possible values:

Related options:

metadata_listen = 0.0.0.0

string value

IP address on which the metadata API will listen.

The metadata API service listens on this IP address for incoming requests.

metadata_listen_port = 8775

port value

Port on which the metadata API will listen.

The metadata API service listens on this port number for incoming requests.

metadata_workers = <based on operating system>

integer value

Number of workers for metadata service. If not specified the number of available CPUs will be used.

The metadata service can be configured to run as multi-process (workers). This overcomes the problem of reduction in throughput when API request concurrency increases. The metadata service will run in the specified number of processes.

Possible Values:

migrate_max_retries = -1

integer value

Number of times to retry live-migration before failing.

Possible values:

mkisofs_cmd = genisoimage

string value

Name or path of the tool used for ISO image creation.

Use the mkisofs_cmd flag to set the path where you install the genisoimage program. If genisoimage is on the system path, you do not need to change the default value.

To use a config drive with Hyper-V, you must set the mkisofs_cmd value to the full path to an mkisofs.exe installation. Additionally, you must set the qemu_img_cmd value in the hyperv configuration section to the full path to an qemu-img command installation.

Possible values:

Related options:

my_block_storage_ip = $my_ip

string value

The IP address which is used to connect to the block storage network.

Possible values:

Related options:

my_ip = <based on operating system>

string value

The IP address which the host is using to connect to the management network.

Possible values:

Related options:

network_allocate_retries = 0

integer value

Number of times to retry network allocation. It is required to attempt network allocation retries if the virtual interface plug fails.

Possible values:

non_inheritable_image_properties = ['cache_in_nova', 'bittorrent']

list value

Image properties that should not be inherited from the instance when taking a snapshot.

This option gives an opportunity to select which image-properties should not be inherited by newly created snapshots.

Possible values:

osapi_compute_listen = 0.0.0.0

string value

IP address on which the OpenStack API will listen.

The OpenStack API service listens on this IP address for incoming requests.

osapi_compute_listen_port = 8774

port value

Port on which the OpenStack API will listen.

The OpenStack API service listens on this port number for incoming requests.

`osapi_compute_unique_server_name_scope = `

string value

Sets the scope of the check for unique instance names.

The default doesn’t check for unique names. If a scope for the name check is set, a launch of a new instance or an update of an existing instance with a duplicate name will result in an 'InstanceExists' error. The uniqueness is case-insensitive. Setting this option can increase the usability for end users as they don’t have to distinguish among instances with the same name by their IDs.

osapi_compute_workers = None

integer value

Number of workers for OpenStack API service. The default will be the number of CPUs available.

OpenStack API services can be configured to run as multi-process (workers). This overcomes the problem of reduction in throughput when API request concurrency increases. OpenStack API service will run in the specified number of processes.

Possible Values:

password_length = 12

integer value

Length of generated instance admin passwords.

periodic_enable = True

boolean value

Enable periodic tasks.

If set to true, this option allows services to periodically run tasks on the manager.

In case of running multiple schedulers or conductors you may want to run periodic tasks on only one host - in this case disable this option for all hosts but one.

periodic_fuzzy_delay = 60

integer value

Number of seconds to randomly delay when starting the periodic task scheduler to reduce stampeding.

When compute workers are restarted in unison across a cluster, they all end up running the periodic tasks at the same time causing problems for the external services. To mitigate this behavior, periodic_fuzzy_delay option allows you to introduce a random initial delay when starting the periodic task scheduler.

Possible Values:

pointer_model = usbtablet

string value

Generic property to specify the pointer type.

Input devices allow interaction with a graphical framebuffer. For example to provide a graphic tablet for absolute cursor movement.

If set, either the hw_input_bus or hw_pointer_model image metadata properties will take precedence over this configuration option.

Related options:

preallocate_images = none

string value

The image preallocation mode to use.

Image preallocation allows storage for instance images to be allocated up front when the instance is initially provisioned. This ensures immediate feedback is given if enough space isn’t available. In addition, it should significantly improve performance on writes to new blocks and may even improve I/O performance to prewritten blocks due to reduced fragmentation.

publish_errors = False

boolean value

Enables or disables publication of error events.

pybasedir = /usr/lib/python3.9/site-packages

string value

The directory where the Nova python modules are installed.

This directory is used to store template files for networking and remote console access. It is also the default path for other config options which need to persist Nova internal data. It is very unlikely that you need to change this option from its default value.

Possible values:

Related options:

ram_allocation_ratio = None

floating point value

Virtual RAM to physical RAM allocation ratio.

This option is used to influence the hosts selected by the Placement API by configuring the allocation ratio for MEMORY_MB inventory.

Possible values:

Related options:

rate_limit_burst = 0

integer value

Maximum number of logged messages per rate_limit_interval.

rate_limit_except_level = CRITICAL

string value

Log level name used by rate limiting: CRITICAL, ERROR, INFO, WARNING, DEBUG or empty string. Logs with level greater or equal to rate_limit_except_level are not filtered. An empty string means that all levels are filtered.

rate_limit_interval = 0

integer value

Interval, number of seconds, of log rate limiting.

reboot_timeout = 0

integer value

Time interval after which an instance is hard rebooted automatically.

When doing a soft reboot, it is possible that a guest kernel is completely hung in a way that causes the soft reboot task to not ever finish. Setting this option to a time period in seconds will automatically hard reboot an instance if it has been stuck in a rebooting state longer than N seconds.

Possible values:

reclaim_instance_interval = 0

integer value

Interval for reclaiming deleted instances.

A value greater than 0 will enable SOFT_DELETE of instances. This option decides whether the server to be deleted will be put into the SOFT_DELETED state. If this value is greater than 0, the deleted server will not be deleted immediately, instead it will be put into a queue until it’s too old (deleted time greater than the value of reclaim_instance_interval). The server can be recovered from the delete queue by using the restore action. If the deleted server remains longer than the value of reclaim_instance_interval, it will be deleted by a periodic task in the compute service automatically.

Note that this option is read from both the API and compute nodes, and must be set globally otherwise servers could be put into a soft deleted state in the API and never actually reclaimed (deleted) on the compute node.

Possible values:

Related options:

record = None

string value

Filename that will be used for storing websocket frames received and sent by a proxy service (like VNC, spice, serial) running on this host. If this is not set, no recording will be done.

reimage_timeout_per_gb = 20

integer value

Timeout for reimaging a volume.

Number of seconds to wait for volume-reimaged events to arrive before continuing or failing.

This is a per gigabyte time which has a default value of 20 seconds and will be multiplied by the GB size of image. Eg: an image of 6 GB will have a timeout of 20 * 6 = 120 seconds. Try increasing the timeout if the image copy per GB takes more time and you are hitting timeout failures.

report_interval = 10

integer value

Number of seconds indicating how frequently the state of services on a given hypervisor is reported. Nova needs to know this to determine the overall health of the deployment.

Related Options:

rescue_timeout = 0

integer value

Interval to wait before un-rescuing an instance stuck in RESCUE.

Possible values:

reserved_host_cpus = 0

integer value

Number of host CPUs to reserve for host processes.

The host resources usage is reported back to the scheduler continuously from nova-compute running on the compute node. This value is used to determine the reserved value reported to placement.

This option cannot be set if the [compute] cpu_shared_set or [compute] cpu_dedicated_set config options have been defined. When these options are defined, any host CPUs not included in these values are considered reserved for the host.

Possible values:

Related options:

reserved_host_disk_mb = 0

integer value

Amount of disk resources in MB to make them always available to host. The disk usage gets reported back to the scheduler from nova-compute running on the compute nodes. To prevent the disk resources from being considered as available, this option can be used to reserve disk space for that host.

Possible values:

reserved_host_memory_mb = 512

integer value

Amount of memory in MB to reserve for the host so that it is always available to host processes. The host resources usage is reported back to the scheduler continuously from nova-compute running on the compute node. To prevent the host memory from being considered as available, this option is used to reserve memory for the host.

Possible values:

reserved_huge_pages = None

dict value

Number of huge/large memory pages to reserved per NUMA host cell.

Possible values:

resize_confirm_window = 0

integer value

Automatically confirm resizes after N seconds.

Resize functionality will save the existing server before resizing. After the resize completes, user is requested to confirm the resize. The user has the opportunity to either confirm or revert all changes. Confirm resize removes the original server and changes server status from resized to active. Setting this option to a time period (in seconds) will automatically confirm the resize if the server is in resized state longer than that time.

Possible values:

resize_fs_using_block_device = False

boolean value

Enable resizing of filesystems via a block device.

If enabled, attempt to resize the filesystem by accessing the image over a block device. This is done by the host and may not be necessary if the image contains a recent version of cloud-init. Possible mechanisms require the nbd driver (for qcow and raw), or loop (for raw).

resume_guests_state_on_host_boot = False

boolean value

This option specifies whether to start guests that were running before the host rebooted. It ensures that all of the instances on a Nova compute node resume their state each time the compute node boots or restarts.

rootwrap_config = /etc/nova/rootwrap.conf

string value

Path to the rootwrap configuration file.

Goal of the root wrapper is to allow a service-specific unprivileged user to run a number of actions as the root user in the safest manner possible. The configuration file used here must match the one defined in the sudoers entry.

rpc_conn_pool_size = 30

integer value

Size of RPC connection pool.

rpc_ping_enabled = False

boolean value

Add an endpoint to answer to ping calls. Endpoint is named oslo_rpc_server_ping

rpc_response_timeout = 60

integer value

Seconds to wait for a response from a call.

run_external_periodic_tasks = True

boolean value

Some periodic tasks can be run in a separate process. Should we run them here?

running_deleted_instance_action = reap

string value

The compute service periodically checks for instances that have been deleted in the database but remain running on the compute node. The above option enables action to be taken when such instances are identified.

Related options:

running_deleted_instance_poll_interval = 1800

integer value

Time interval in seconds to wait between runs for the clean up action. If set to 0, above check will be disabled. If "running_deleted_instance _action" is set to "log" or "reap", a value greater than 0 must be set.

Possible values:

Related options:

running_deleted_instance_timeout = 0

integer value

Time interval in seconds to wait for the instances that have been marked as deleted in database to be eligible for cleanup.

Possible values:

Related options:

scheduler_instance_sync_interval = 120

integer value

Interval between sending the scheduler a list of current instance UUIDs to verify that its view of instances is in sync with nova.

If the CONF option scheduler_tracks_instance_changes is False, the sync calls will not be made. So, changing this option will have no effect.

If the out of sync situations are not very common, this interval can be increased to lower the number of RPC messages being sent. Likewise, if sync issues turn out to be a problem, the interval can be lowered to check more frequently.

Possible values:

Related options:

service_down_time = 60

integer value

Maximum time in seconds since last check-in for up service

Each compute node periodically updates their database status based on the specified report interval. If the compute node hasn’t updated the status for more than service_down_time, then the compute node is considered down.

Related Options:

servicegroup_driver = db

string value

This option specifies the driver to be used for the servicegroup service.

ServiceGroup API in nova enables checking status of a compute node. When a compute worker running the nova-compute daemon starts, it calls the join API to join the compute group. Services like nova scheduler can query the ServiceGroup API to check if a node is alive. Internally, the ServiceGroup client driver automatically updates the compute worker status. There are multiple backend implementations for this service: Database ServiceGroup driver and Memcache ServiceGroup driver.

Related Options:

shelved_offload_time = 0

integer value

Time before a shelved instance is eligible for removal from a host.

By default this option is set to 0 and the shelved instance will be removed from the hypervisor immediately after shelve operation. Otherwise, the instance will be kept for the value of shelved_offload_time(in seconds) so that during the time period the unshelve action will be faster, then the periodic task will remove the instance from hypervisor after shelved_offload_time passes.

Possible values:

shelved_poll_interval = 3600

integer value

Interval for polling shelved instances to offload.

The periodic task runs for every shelved_poll_interval number of seconds and checks if there are any shelved instances. If it finds a shelved instance, based on the shelved_offload_time config value it offloads the shelved instances. Check shelved_offload_time config option description for details.

Possible values:

Related options:

shutdown_timeout = 60

integer value

Total time to wait in seconds for an instance to perform a clean shutdown.

It determines the overall period (in seconds) a VM is allowed to perform a clean shutdown. While performing stop, rescue and shelve, rebuild operations, configuring this option gives the VM a chance to perform a controlled shutdown before the instance is powered off. The default timeout is 60 seconds. A value of 0 (zero) means the guest will be powered off immediately with no opportunity for guest OS clean-up.

The timeout value can be overridden on a per image basis by means of os_shutdown_timeout that is an image metadata setting allowing different types of operating systems to specify how much time they need to shut down cleanly.

Possible values:

source_is_ipv6 = False

boolean value

Set to True if source host is addressed with IPv6.

ssl_only = False

boolean value

Disallow non-encrypted connections.

Related options:

state_path = $pybasedir

string value

The top-level directory for maintaining Nova’s state.

This directory is used to store Nova’s internal state. It is used by a variety of other config options which derive from this. In some scenarios (for example migrations) it makes sense to use a storage location which is shared between multiple compute hosts (for example via NFS). Unless the option instances_path gets overwritten, this directory can grow very large.

Possible values:

sync_power_state_interval = 600

integer value

Interval to sync power states between the database and the hypervisor.

The interval that Nova checks the actual virtual machine power state and the power state that Nova has in its database. If a user powers down their VM, Nova updates the API to report the VM has been powered down. Should something turn on the VM unexpectedly, Nova will turn the VM back off to keep the system in the expected state.

Possible values:

Related options:

sync_power_state_pool_size = 1000

integer value

Number of greenthreads available for use to sync power states.

This option can be used to reduce the number of concurrent requests made to the hypervisor or system with real instance power states for performance reasons, for example, with Ironic.

Possible values:

syslog-log-facility = LOG_USER

string value

Syslog facility to receive log lines. This option is ignored if log_config_append is set.

tempdir = None

string value

Explicitly specify the temporary working directory.

timeout_nbd = 10

integer value

Amount of time, in seconds, to wait for NBD device start up.

transport_url = rabbit://

string value

The network address and optional user credentials for connecting to the messaging backend, in URL format. The expected format is:

driver://[user:pass@]host:port[,[userN:passN@]hostN:portN]/virtual_host?query

Example: rabbit://rabbitmq:password@127.0.0.1:5672//

For full details on the fields in the URL see the documentation of oslo_messaging.TransportURL at https://docs.openstack.org/oslo.messaging/latest/reference/transport.html

update_resources_interval = 0

integer value

Interval for updating compute resources.

This option specifies how often the update_available_resource periodic task should run. A number less than 0 means to disable the task completely. Leaving this at the default of 0 will cause this to run at the default periodic interval. Setting it to any positive value will cause it to run at approximately that number of seconds.

Possible values:

use-journal = False

boolean value

Enable journald for logging. If running in a systemd environment you may wish to enable journal support. Doing so will use the journal native protocol which includes structured metadata in addition to log messages.This option is ignored if log_config_append is set.

use-json = False

boolean value

Use JSON formatting for logging. This option is ignored if log_config_append is set.

use-syslog = False

boolean value

Use syslog for logging. Existing syslog format is DEPRECATED and will be changed later to honor RFC5424. This option is ignored if log_config_append is set.

use_cow_images = True

boolean value

Enable use of copy-on-write (cow) images.

QEMU/KVM allow the use of qcow2 as backing files. By disabling this, backing files will not be used.

use_eventlog = False

boolean value

Log output to Windows Event Log.

use_rootwrap_daemon = False

boolean value

Start and use a daemon that can run the commands that need to be run with root privileges. This option is usually enabled on nodes that run nova compute processes.

use_stderr = False

boolean value

Log output to standard error. This option is ignored if log_config_append is set.

vcpu_pin_set = None

string value

Mask of host CPUs that can be used for VCPU resources.

The behavior of this option depends on the definition of the [compute] cpu_dedicated_set option and affects the behavior of the [compute] cpu_shared_set option.

Possible values:

Related options:

Deprecated since: 20.0.0

Reason: This option has been superseded by the ``[compute] cpu_dedicated_set`` and ``[compute] cpu_shared_set`` options, which allow things like the co-existence of pinned and unpinned instances on the same host (for the libvirt driver).

vif_plugging_is_fatal = True

boolean value

Determine if instance should boot or fail on VIF plugging timeout.

Nova sends a port update to Neutron after an instance has been scheduled, providing Neutron with the necessary information to finish setup of the port. Once completed, Neutron notifies Nova that it has finished setting up the port, at which point Nova resumes the boot of the instance since network connectivity is now supposed to be present. A timeout will occur if the reply is not received after a given interval.

This option determines what Nova does when the VIF plugging timeout event happens. When enabled, the instance will error out. When disabled, the instance will continue to boot on the assumption that the port is ready.

Possible values:

vif_plugging_timeout = 300

integer value

Timeout for Neutron VIF plugging event message arrival.

Number of seconds to wait for Neutron vif plugging events to arrive before continuing or failing (see vif_plugging_is_fatal).

If you are hitting timeout failures at scale, consider running rootwrap in "daemon mode" in the neutron agent via the [agent]/root_helper_daemon neutron configuration option.

Related options:

virt_mkfs = []

multi valued

Name of the mkfs commands for ephemeral device.

The format is <os_type>=<mkfs command>

volume_usage_poll_interval = 0

integer value

Interval for gathering volume usages.

This option updates the volume usage cache for every volume_usage_poll_interval number of seconds.

Possible values:

watch-log-file = False

boolean value

Uses logging handler designed to watch file system. When log file is moved or removed this handler will open a new log file with specified path instantaneously. It makes sense only if log_file option is specified and Linux platform is used. This option is ignored if log_config_append is set.

web = /usr/share/spice-html5

string value

Path to directory with content which will be served by a web server.

auth_strategy = keystone

string value

Determine the strategy to use for authentication.

Deprecated since: 21.0.0

Reason: The only non-default choice, ``noauth2``, is for internal development and testing purposes only and should not be used in deployments. This option and its middleware, NoAuthMiddleware[V2_18], will be removed in a future release.

compute_link_prefix = None

string value

This string is prepended to the normal URL that is returned in links to the OpenStack Compute API. If it is empty (the default), the URLs are returned unchanged.

Possible values:

config_drive_skip_versions = 1.0 2007-01-19 2007-03-01 2007-08-29 2007-10-10 2007-12-15 2008-02-01 2008-09-01

string value

When gathering the existing metadata for a config drive, the EC2-style metadata is returned for all versions that don’t appear in this option. As of the Liberty release, the available versions are:

The option is in the format of a single string, with each version separated by a space.

Possible values:

dhcp_domain = novalocal

string value

Domain name used to configure FQDN for instances.

Configure a fully-qualified domain name for instance hostnames. The value is suffixed to the instance hostname from the database to construct the hostname that appears in the metadata API. To disable this behavior (for example in order to correctly support microversion’s 2.94 FQDN hostnames), set this to the empty string.

Possible values:

enable_instance_password = True

boolean value

Enables returning of the instance password by the relevant server API calls such as create, rebuild, evacuate, or rescue. If the hypervisor does not support password injection, then the password returned will not be correct, so if your hypervisor does not support password injection, set this to False.

glance_link_prefix = None

string value

This string is prepended to the normal URL that is returned in links to Glance resources. If it is empty (the default), the URLs are returned unchanged.

Possible values:

instance_list_cells_batch_fixed_size = 100

integer value

This controls the batch size of instances requested from each cell database if instance_list_cells_batch_strategy` is set to fixed. This integral value will define the limit issued to each cell every time a batch of instances is requested, regardless of the number of cells in the system or any other factors. Per the general logic called out in the documentation for instance_list_cells_batch_strategy, the minimum value for this is 100 records per batch.

Related options:

instance_list_cells_batch_strategy = distributed

string value

This controls the method by which the API queries cell databases in smaller batches during large instance list operations. If batching is performed, a large instance list operation will request some fraction of the overall API limit from each cell database initially, and will re-request that same batch size as records are consumed (returned) from each cell as necessary. Larger batches mean less chattiness between the API and the database, but potentially more wasted effort processing the results from the database which will not be returned to the user. Any strategy will yield a batch size of at least 100 records, to avoid a user causing many tiny database queries in their request.

Related options:

instance_list_per_project_cells = False

boolean value

When enabled, this will cause the API to only query cell databases in which the tenant has mapped instances. This requires an additional (fast) query in the API database before each list, but also (potentially) limits the number of cell databases that must be queried to provide the result. If you have a small number of cells, or tenants are likely to have instances in all cells, then this should be False. If you have many cells, especially if you confine tenants to a small subset of those cells, this should be True.

list_records_by_skipping_down_cells = True

boolean value

When set to False, this will cause the API to return a 500 error if there is an infrastructure failure like non-responsive cells. If you want the API to skip the down cells and return the results from the up cells set this option to True.

Note that from API microversion 2.69 there could be transient conditions in the deployment where certain records are not available and the results could be partial for certain requests containing those records. In those cases this option will be ignored. See "Handling Down Cells" section of the Compute API guide (https://docs.openstack.org/api-guide/compute/down_cells.html) for more information.

local_metadata_per_cell = False

boolean value

Indicates that the nova-metadata API service has been deployed per-cell, so that we can have better performance and data isolation in a multi-cell deployment. Users should consider the use of this configuration depending on how neutron is setup. If you have networks that span cells, you might need to run nova-metadata API service globally. If your networks are segmented along cell boundaries, then you can run nova-metadata API service per cell. When running nova-metadata API service per cell, you should also configure each Neutron metadata-agent to point to the corresponding nova-metadata API service.

max_limit = 1000

integer value

As a query can potentially return many thousands of items, you can limit the maximum number of items in a single response by setting this option.

metadata_cache_expiration = 15

integer value

This option is the time (in seconds) to cache metadata. When set to 0, metadata caching is disabled entirely; this is generally not recommended for performance reasons. Increasing this setting should improve response times of the metadata API when under heavy load. Higher values may increase memory usage, and result in longer times for host metadata changes to take effect.

neutron_default_tenant_id = default

string value

Tenant ID for getting the default network from Neutron API (also referred in some places as the project ID) to use.

Related options:

use_forwarded_for = False

boolean value

When True, the X-Forwarded-For header is treated as the canonical remote address. When False (the default), the remote_address header is used.

You should only enable this if you have an HTML sanitizing proxy.

Deprecated since: 26.0.0

*Reason:*This feature is duplicate of the HTTPProxyToWSGI middleware in oslo.middleware

use_neutron_default_nets = False

boolean value

When True, the TenantNetworkController will query the Neutron API to get the default networks to use.

Related options:

vendordata_dynamic_connect_timeout = 5

integer value

Maximum wait time for an external REST service to connect.

Possible values:

Related options:

vendordata_dynamic_failure_fatal = False

boolean value

Should failures to fetch dynamic vendordata be fatal to instance boot?

Related options:

vendordata_dynamic_read_timeout = 5

integer value

Maximum wait time for an external REST service to return data once connected.

Possible values:

Related options:

`vendordata_dynamic_ssl_certfile = `

string value

Path to an optional certificate file or CA bundle to verify dynamic vendordata REST services ssl certificates against.

Possible values:

Related options:

vendordata_dynamic_targets = []

list value

A list of targets for the dynamic vendordata provider. These targets are of the form <name>@<url>.

The dynamic vendordata provider collects metadata by contacting external REST services and querying them for information about the instance. This behaviour is documented in the vendordata.rst file in the nova developer reference.

vendordata_jsonfile_path = None

string value

Cloud providers may store custom data in vendor data file that will then be available to the instances via the metadata service, and to the rendering of config-drive. The default class for this, JsonFileVendorData, loads this information from a JSON file, whose path is configured by this option. If there is no path set by this option, the class returns an empty dictionary.

Note that when using this to provide static vendor data to a configuration drive, the nova-compute service must be configured with this option and the file must be accessible from the nova-compute host.

Possible values:

vendordata_providers = ['StaticJSON']

list value

A list of vendordata providers.

vendordata providers are how deployers can provide metadata via configdrive and metadata that is specific to their deployment.

For more information on the requirements for implementing a vendordata dynamic endpoint, please see the vendordata.rst file in the nova developer reference.

Related options:

backend = sqlalchemy

string value

The back end to use for the database.

connection = None

string value

The SQLAlchemy connection string to use to connect to the database.

connection_debug = 0

integer value

Verbosity of SQL debugging information: 0=None, 100=Everything.

`connection_parameters = `

string value

Optional URL parameters to append onto the connection URL at connect time; specify as param1=value1&param2=value2&…​

connection_recycle_time = 3600

integer value

Connections which have been present in the connection pool longer than this number of seconds will be replaced with a new one the next time they are checked out from the pool.

connection_trace = False

boolean value

Add Python stack traces to SQL as comment strings.

db_inc_retry_interval = True

boolean value

If True, increases the interval between retries of a database operation up to db_max_retry_interval.

db_max_retries = 20

integer value

Maximum retries in case of connection error or deadlock error before error is raised. Set to -1 to specify an infinite retry count.

db_max_retry_interval = 10

integer value

If db_inc_retry_interval is set, the maximum seconds between retries of a database operation.

db_retry_interval = 1

integer value

Seconds between retries of a database transaction.

max_overflow = 50

integer value

If set, use this value for max_overflow with SQLAlchemy.

max_pool_size = 5

integer value

Maximum number of SQL connections to keep open in a pool. Setting a value of 0 indicates no limit.

max_retries = 10

integer value

Maximum number of database connection retries during startup. Set to -1 to specify an infinite retry count.

mysql_enable_ndb = False

boolean value

If True, transparently enables support for handling MySQL Cluster (NDB). Deprecated since: 12.1.0

*Reason:*Support for the MySQL NDB Cluster storage engine has been deprecated and will be removed in a future release.

mysql_sql_mode = TRADITIONAL

string value

The SQL mode to be used for MySQL sessions. This option, including the default, overrides any server-set SQL mode. To use whatever SQL mode is set by the server configuration, set this to no value. Example: mysql_sql_mode=

mysql_wsrep_sync_wait = None

integer value

For Galera only, configure wsrep_sync_wait causality checks on new connections. Default is None, meaning don’t configure any setting.

pool_timeout = None

integer value

If set, use this value for pool_timeout with SQLAlchemy.

retry_interval = 10

integer value

Interval between retries of opening a SQL connection.

slave_connection = None

string value

The SQLAlchemy connection string to use to connect to the slave database.

sqlite_synchronous = True

boolean value

If True, SQLite uses synchronous mode.

auth_endpoint = http://localhost/identity/v3

string value

Use this endpoint to connect to Keystone

barbican_api_version = None

string value

Version of the Barbican API, for example: "v1"

barbican_endpoint = None

string value

Use this endpoint to connect to Barbican, for example: "http://localhost:9311/"

barbican_endpoint_type = public

string value

Specifies the type of endpoint. Allowed values are: public, private, and admin

barbican_region_name = None

string value

Specifies the region of the chosen endpoint.

number_of_retries = 60

integer value

Number of times to retry poll for key creation completion

retry_delay = 1

integer value

Number of seconds to wait before retrying poll for key creation completion

send_service_user_token = False

boolean value

When True, if sending a user token to a REST API, also send a service token.

Nova often reuses the user token provided to the nova-api to talk to other REST APIs, such as Cinder, Glance and Neutron. It is possible that while the user token was valid when the request was made to Nova, the token may expire before it reaches the other service. To avoid any failures, and to make it clear it is Nova calling the service on the user’s behalf, we include a service token along with the user token. Should the user’s token have expired, a valid service token ensures the REST API request will still be accepted by the keystone middleware.

verify_ssl = True

boolean value

Specifies if insecure TLS (https) requests. If False, the server’s certificate will not be validated, if True, we can set the verify_ssl_path config meanwhile.

verify_ssl_path = None

string value

A path to a bundle or CA certs to check against, or None for requests to attempt to locate and use certificates which verify_ssh is True. If verify_ssl is False, this is ignored.

auth_section = None

string value

Config Section from which to load plugin specific options

auth_type = None

string value

Authentication type to load

cafile = None

string value

PEM encoded Certificate Authority to use when verifying HTTPs connections.

certfile = None

string value

PEM encoded client certificate cert file

collect-timing = False

boolean value

Collect per-API call timing information.

insecure = False

boolean value

Verify HTTPS connections.

keyfile = None

string value

PEM encoded client certificate key file

split-loggers = False

boolean value

Log requests to multiple loggers.

timeout = None

integer value

Timeout value for http requests

backend = dogpile.cache.null

string value

Cache backend module. For eventlet-based or environments with hundreds of threaded servers, Memcache with pooling (oslo_cache.memcache_pool) is recommended. For environments with less than 100 threaded servers, Memcached (dogpile.cache.memcached) or Redis (dogpile.cache.redis) is recommended. Test environments with a single instance of the server can use the dogpile.cache.memory backend.

backend_argument = []

multi valued

Arguments supplied to the backend module. Specify this option once per argument to be passed to the dogpile.cache backend. Example format: "<argname>:<value>".

config_prefix = cache.oslo

string value

Prefix for building the configuration dictionary for the cache region. This should not need to be changed unless there is another dogpile.cache region with the same configuration name.

dead_timeout = 60

floating point value

Time in seconds before attempting to add a node back in the pool in the HashClient’s internal mechanisms.

debug_cache_backend = False

boolean value

Extra debugging from the cache backend (cache keys, get/set/delete/etc calls). This is only really useful if you need to see the specific cache-backend get/set/delete calls with the keys/values. Typically this should be left set to false.

enable_retry_client = False

boolean value

Enable retry client mechanisms to handle failure. Those mechanisms can be used to wrap all kind of pymemcache clients. The wrapper allows you to define how many attempts to make and how long to wait between attemots.

enable_socket_keepalive = False

boolean value

Global toggle for the socket keepalive of dogpile’s pymemcache backend

enabled = False

boolean value

Global toggle for caching.

expiration_time = 600

integer value

Default TTL, in seconds, for any cached item in the dogpile.cache region. This applies to any cached method that doesn’t have an explicit cache expiration time defined for it.

hashclient_retry_attempts = 2

integer value

Amount of times a client should be tried before it is marked dead and removed from the pool in the HashClient’s internal mechanisms.

hashclient_retry_delay = 1

floating point value

Time in seconds that should pass between retry attempts in the HashClient’s internal mechanisms.

memcache_dead_retry = 300

integer value

Number of seconds memcached server is considered dead before it is tried again. (dogpile.cache.memcache and oslo_cache.memcache_pool backends only).

`memcache_password = `

string value

the password for the memcached which SASL enabled

memcache_pool_connection_get_timeout = 10

integer value

Number of seconds that an operation will wait to get a memcache client connection.

memcache_pool_flush_on_reconnect = False

boolean value

Global toggle if memcache will be flushed on reconnect. (oslo_cache.memcache_pool backend only).

memcache_pool_maxsize = 10

integer value

Max total number of open connections to every memcached server. (oslo_cache.memcache_pool backend only).

memcache_pool_unused_timeout = 60

integer value

Number of seconds a connection to memcached is held unused in the pool before it is closed. (oslo_cache.memcache_pool backend only).

memcache_sasl_enabled = False

boolean value

Enable the SASL(Simple Authentication and SecurityLayer) if the SASL_enable is true, else disable.

memcache_servers = ['localhost:11211']

list value

Memcache servers in the format of "host:port". This is used by backends dependent on Memcached.If dogpile.cache.memcached or oslo_cache.memcache_pool is used and a given host refer to an IPv6 or a given domain refer to IPv6 then you should prefix the given address withthe address family (inet6) (e.g inet6[::1]:11211, inet6:[fd12:3456:789a:1::1]:11211, inet6:[controller-0.internalapi]:11211). If the address family is not given then these backends will use the default inet address family which corresponds to IPv4

memcache_socket_timeout = 1.0

floating point value

Timeout in seconds for every call to a server. (dogpile.cache.memcache and oslo_cache.memcache_pool backends only).

`memcache_username = `

string value

the user name for the memcached which SASL enabled

proxies = []

list value

Proxy classes to import that will affect the way the dogpile.cache backend functions. See the dogpile.cache documentation on changing-backend-behavior.

retry_attempts = 2

integer value

Number of times to attempt an action before failing.

retry_delay = 0

floating point value

Number of seconds to sleep between each attempt.

socket_keepalive_count = 1

integer value

The maximum number of keepalive probes TCP should send before dropping the connection. Should be a positive integer greater than zero.

socket_keepalive_idle = 1

integer value

The time (in seconds) the connection needs to remain idle before TCP starts sending keepalive probes. Should be a positive integer most greater than zero.

socket_keepalive_interval = 1

integer value

The time (in seconds) between individual keepalive probes. Should be a positive integer greater than zero.

tls_allowed_ciphers = None

string value

Set the available ciphers for sockets created with the TLS context. It should be a string in the OpenSSL cipher list format. If not specified, all OpenSSL enabled ciphers will be available.

tls_cafile = None

string value

Path to a file of concatenated CA certificates in PEM format necessary to establish the caching servers' authenticity. If tls_enabled is False, this option is ignored.

tls_certfile = None

string value

Path to a single file in PEM format containing the client’s certificate as well as any number of CA certificates needed to establish the certificate’s authenticity. This file is only required when client side authentication is necessary. If tls_enabled is False, this option is ignored.

tls_enabled = False

boolean value

Global toggle for TLS usage when comunicating with the caching servers.

tls_keyfile = None

string value

Path to a single file containing the client’s private key in. Otherwise the private key will be taken from the file specified in tls_certfile. If tls_enabled is False, this option is ignored.

auth-url = None

string value

Authentication URL

auth_section = None

string value

Config Section from which to load plugin specific options

auth_type = None

string value

Authentication type to load

cafile = None

string value

PEM encoded Certificate Authority to use when verifying HTTPs connections.

catalog_info = volumev3::publicURL

string value

Info to match when looking for cinder in the service catalog.

The <service_name> is optional and omitted by default since it should not be necessary in most deployments.

Possible values:

Note: Nova does not support the Cinder v2 API since the Nova 17.0.0 Queens release.

Related options:

certfile = None

string value

PEM encoded client certificate cert file

collect-timing = False

boolean value

Collect per-API call timing information.

cross_az_attach = True

boolean value

Allow attach between instance and volume in different availability zones.

If False, volumes attached to an instance must be in the same availability zone in Cinder as the instance availability zone in Nova.

This also means care should be taken when booting an instance from a volume where source is not "volume" because Nova will attempt to create a volume using the same availability zone as what is assigned to the instance.

If that AZ is not in Cinder (or allow_availability_zone_fallback=False in cinder.conf), the volume create request will fail and the instance will fail the build request.

By default there is no availability zone restriction on volume attach.

Related options:

debug = False

boolean value

Enable DEBUG logging with cinderclient and os_brick independently of the rest of Nova.

default-domain-id = None

string value

Optional domain ID to use with v3 and v2 parameters. It will be used for both the user and project domain in v3 and ignored in v2 authentication.

default-domain-name = None

string value

Optional domain name to use with v3 API and v2 parameters. It will be used for both the user and project domain in v3 and ignored in v2 authentication.

domain-id = None

string value

Domain ID to scope to

domain-name = None

string value

Domain name to scope to

endpoint_template = None

string value

If this option is set then it will override service catalog lookup with this template for cinder endpoint

Possible values:

Note: Nova does not support the Cinder v2 API since the Nova 17.0.0 Queens release.

Related options:

http_retries = 3

integer value

Number of times cinderclient should retry on any failed http call. 0 means connection is attempted only once. Setting it to any positive integer means that on failure connection is retried that many times e.g. setting it to 3 means total attempts to connect will be 4.

Possible values:

insecure = False

boolean value

Verify HTTPS connections.

keyfile = None

string value

PEM encoded client certificate key file

os_region_name = None

string value

Region name of this node. This is used when picking the URL in the service catalog.

Possible values:

password = None

string value

User’s password

project-domain-id = None

string value

Domain ID containing project

project-domain-name = None

string value

Domain name containing project

project-id = None

string value

Project ID to scope to

project-name = None

string value

Project name to scope to

split-loggers = False

boolean value

Log requests to multiple loggers.

system-scope = None

string value

Scope for system operations

tenant-id = None

string value

Tenant ID

tenant-name = None

string value

Tenant Name

timeout = None

integer value

Timeout value for http requests

trust-id = None

string value

ID of the trust to use as a trustee use

user-domain-id = None

string value

User’s domain id

user-domain-name = None

string value

User’s domain name

user-id = None

string value

User ID

username = None

string value

Username

consecutive_build_service_disable_threshold = 10

integer value

Enables reporting of build failures to the scheduler.

Any nonzero value will enable sending build failure statistics to the scheduler for use by the BuildFailureWeigher.

Possible values:

Related options:

cpu_dedicated_set = None

string value

Mask of host CPUs that can be used for PCPU resources.

The behavior of this option affects the behavior of the deprecated vcpu_pin_set option.

This behavior will be simplified in a future release when vcpu_pin_set is removed.

Possible values:

Related options:

cpu_shared_set = None

string value

Mask of host CPUs that can be used for VCPU resources and offloaded emulator threads.

The behavior of this option depends on the definition of the deprecated vcpu_pin_set option.

This behavior will be simplified in a future release when vcpu_pin_set is removed.

Possible values:

Related options:

image_type_exclude_list = []

list value

A list of image formats that should not be advertised as supported by this compute node.

In some situations, it may be desirable to have a compute node refuse to support an expensive or complex image format. This factors into the decisions made by the scheduler about which compute node to select when booted with a given image.

Possible values:

Related options:

live_migration_wait_for_vif_plug = True

boolean value

Determine if the source compute host should wait for a network-vif-plugged event from the (neutron) networking service before starting the actual transfer of the guest to the destination compute host.

Note that this option is read on the destination host of a live migration. If you set this option the same on all of your compute hosts, which you should do if you use the same networking backend universally, you do not have to worry about this.

Before starting the transfer of the guest, some setup occurs on the destination compute host, including plugging virtual interfaces. Depending on the networking backend on the destination host, a network-vif-plugged event may be triggered and then received on the source compute host and the source compute can wait for that event to ensure networking is set up on the destination host before starting the guest transfer in the hypervisor.

Possible values:

Related options:

max_concurrent_disk_ops = 0

integer value

Number of concurrent disk-IO-intensive operations (glance image downloads, image format conversions, etc.) that we will do in parallel. If this is set too high then response time suffers. The default value of 0 means no limit.

max_disk_devices_to_attach = -1

integer value

Maximum number of disk devices allowed to attach to a single server. Note that the number of disks supported by an server depends on the bus used. For example, the ide disk bus is limited to 4 attached devices. The configured maximum is enforced during server create, rebuild, evacuate, unshelve, live migrate, and attach volume.

Usually, disk bus is determined automatically from the device type or disk device, and the virtualization type. However, disk bus can also be specified via a block device mapping or an image property. See the disk_bus field in :doc:/user/block-device-mapping for more information about specifying disk bus in a block device mapping, and see https://docs.openstack.org/glance/latest/admin/useful-image-properties.html for more information about the hw_disk_bus image property.

Operators changing the [compute]/max_disk_devices_to_attach on a compute service that is hosting servers should be aware that it could cause rebuilds to fail, if the maximum is decreased lower than the number of devices already attached to servers. For example, if server A has 26 devices attached and an operators changes [compute]/max_disk_devices_to_attach to 20, a request to rebuild server A will fail and go into ERROR state because 26 devices are already attached and exceed the new configured maximum of 20.

Operators setting [compute]/max_disk_devices_to_attach should also be aware that during a cold migration, the configured maximum is only enforced in-place and the destination is not checked before the move. This means if an operator has set a maximum of 26 on compute host A and a maximum of 20 on compute host B, a cold migration of a server with 26 attached devices from compute host A to compute host B will succeed. Then, once the server is on compute host B, a subsequent request to rebuild the server will fail and go into ERROR state because 26 devices are already attached and exceed the configured maximum of 20 on compute host B.

The configured maximum is not enforced on shelved offloaded servers, as they have no compute host.

Possible values:

packing_host_numa_cells_allocation_strategy = False

boolean value

This option controls allocation strategy used to choose NUMA cells on host for placing VM’s NUMA cells (for VMs with defined numa topology). By default host’s NUMA cell with more resources consumed will be chosen last for placing attempt. When the packing_host_numa_cells_allocation_strategy variable is set to False, host’s NUMA cell with more resources available will be used. When set to True cells with some usage will be packed with VM’s cell until it will be completely exhausted, before a new free host’s cell will be used.

Possible values:

provider_config_location = /etc/nova/provider_config/

string value

Location of YAML files containing resource provider configuration data.

These files allow the operator to specify additional custom inventory and traits to assign to one or more resource providers.

Additional documentation is available here:

https://docs.openstack.org/nova/latest/admin/managing-resource-providers.html

resource_provider_association_refresh = 300

integer value

Interval for updating nova-compute-side cache of the compute node resource provider’s inventories, aggregates, and traits.

This option specifies the number of seconds between attempts to update a provider’s inventories, aggregates and traits in the local cache of the compute node.

A value of zero disables cache refresh completely.

The cache can be cleared manually at any time by sending SIGHUP to the compute process, causing it to be repopulated the next time the data is accessed.

Possible values:

shutdown_retry_interval = 10

integer value

Time to wait in seconds before resending an ACPI shutdown signal to instances.

The overall time to wait is set by shutdown_timeout.

Possible values:

Related options:

vmdk_allowed_types = ['streamOptimized', 'monolithicSparse']

list value

A list of strings describing allowed VMDK "create-type" subformats that will be allowed. This is recommended to only include single-file-with-sparse-header variants to avoid potential host file exposure due to processing named extents. If this list is empty, then no form of VMDK image will be allowed.

workers = None

integer value

Number of workers for OpenStack Conductor service. The default will be the number of CPUs available.

allowed_origins = []

list value

Adds list of allowed origins to the console websocket proxy to allow connections from other origin hostnames. Websocket proxy matches the host header with the origin header to prevent cross-site requests. This list specifies if any there are values other than host are allowed in the origin header.

Possible values:

ssl_ciphers = None

string value

OpenSSL cipher preference string that specifies what ciphers to allow for TLS connections from clients. For example::

ssl_ciphers = "kEECDH+aECDSA+AES:kEECDH+AES+aRSA:kEDH+aRSA+AES"

See the man page for the OpenSSL ciphers command for details of the cipher preference string format and allowed values::

https://www.openssl.org/docs/man1.1.0/man1/ciphers.html

Related options:

ssl_minimum_version = default

string value

Minimum allowed SSL/TLS protocol version.

Related options:

token_ttl = 600

integer value

The lifetime of a console auth token (in seconds).

A console auth token is used in authorizing console access for a user. Once the auth token time to live count has elapsed, the token is considered expired. Expired tokens are then deleted.

allow_credentials = True

boolean value

Indicate that the actual request can include user credentials

allow_headers = ['X-Auth-Token', 'X-Openstack-Request-Id', 'X-Identity-Status', 'X-Roles', 'X-Service-Catalog', 'X-User-Id', 'X-Tenant-Id', 'X-OpenStack-Nova-API-Version', 'OpenStack-API-Version']

list value

Indicate which header field names may be used during the actual request.

allow_methods = ['GET', 'PUT', 'POST', 'DELETE', 'PATCH']

list value

Indicate which methods can be used during the actual request.

allowed_origin = None

list value

Indicate whether this resource may be shared with the domain received in the requests "origin" header. Format: "<protocol>://<host>[:<port>]", no trailing slash. Example: https://horizon.example.com

expose_headers = ['X-Auth-Token', 'X-Openstack-Request-Id', 'X-Subject-Token', 'X-Service-Token', 'X-OpenStack-Nova-API-Version', 'OpenStack-API-Version']

list value

Indicate which headers are safe to expose to the API. Defaults to HTTP Simple Headers.

max_age = 3600

integer value

Maximum cache age of CORS preflight requests.

cafile = None

string value

PEM encoded Certificate Authority to use when verifying HTTPs connections.

certfile = None

string value

PEM encoded client certificate cert file

collect-timing = False

boolean value

Collect per-API call timing information.

connect-retries = None

integer value

The maximum number of retries that should be attempted for connection errors.

connect-retry-delay = None

floating point value

Delay (in seconds) between two retries for connection errors. If not set, exponential retry starting with 0.5 seconds up to a maximum of 60 seconds is used.

endpoint-override = None

string value

Always use this endpoint URL for requests for this client. NOTE: The unversioned endpoint should be specified here; to request a particular API version, use the version, min-version, and/or max-version options.

insecure = False

boolean value

Verify HTTPS connections.

keyfile = None

string value

PEM encoded client certificate key file

region-name = None

string value

The default region_name for endpoint URL discovery.

service-name = None

string value

The default service_name for endpoint URL discovery.

service-type = accelerator

string value

The default service_type for endpoint URL discovery.

split-loggers = False

boolean value

Log requests to multiple loggers.

status-code-retries = None

integer value

The maximum number of retries that should be attempted for retriable HTTP status codes.

status-code-retry-delay = None

floating point value

Delay (in seconds) between two retries for retriable status codes. If not set, exponential retry starting with 0.5 seconds up to a maximum of 60 seconds is used.

timeout = None

integer value

Timeout value for http requests

valid-interfaces = ['internal', 'public']

list value

List of interfaces, in order of preference, for endpoint URL.

backend = sqlalchemy

string value

The back end to use for the database.

connection = None

string value

The SQLAlchemy connection string to use to connect to the database.

connection_debug = 0

integer value

Verbosity of SQL debugging information: 0=None, 100=Everything.

`connection_parameters = `

string value

Optional URL parameters to append onto the connection URL at connect time; specify as param1=value1&param2=value2&…​

connection_recycle_time = 3600

integer value

Connections which have been present in the connection pool longer than this number of seconds will be replaced with a new one the next time they are checked out from the pool.

connection_trace = False

boolean value

Add Python stack traces to SQL as comment strings.

db_inc_retry_interval = True

boolean value

If True, increases the interval between retries of a database operation up to db_max_retry_interval.

db_max_retries = 20

integer value

Maximum retries in case of connection error or deadlock error before error is raised. Set to -1 to specify an infinite retry count.

db_max_retry_interval = 10

integer value

If db_inc_retry_interval is set, the maximum seconds between retries of a database operation.

db_retry_interval = 1

integer value

Seconds between retries of a database transaction.

max_overflow = 50

integer value

If set, use this value for max_overflow with SQLAlchemy.

max_pool_size = 5

integer value

Maximum number of SQL connections to keep open in a pool. Setting a value of 0 indicates no limit.

max_retries = 10

integer value

Maximum number of database connection retries during startup. Set to -1 to specify an infinite retry count.

mysql_enable_ndb = False

boolean value

If True, transparently enables support for handling MySQL Cluster (NDB). Deprecated since: 12.1.0

*Reason:*Support for the MySQL NDB Cluster storage engine has been deprecated and will be removed in a future release.

mysql_sql_mode = TRADITIONAL

string value

The SQL mode to be used for MySQL sessions. This option, including the default, overrides any server-set SQL mode. To use whatever SQL mode is set by the server configuration, set this to no value. Example: mysql_sql_mode=

mysql_wsrep_sync_wait = None

integer value

For Galera only, configure wsrep_sync_wait causality checks on new connections. Default is None, meaning don’t configure any setting.

pool_timeout = None

integer value

If set, use this value for pool_timeout with SQLAlchemy.

retry_interval = 10

integer value

Interval between retries of opening a SQL connection.

slave_connection = None

string value

The SQLAlchemy connection string to use to connect to the slave database.

sqlite_synchronous = True

boolean value

If True, SQLite uses synchronous mode.

use_db_reconnect = False

boolean value

Enable the experimental use of database reconnect on connection lost.

use_tpool = False

boolean value

Enable the experimental use of thread pooling for all DB API calls Deprecated since: 10.0.0

*Reason:*This feature has never graduated from experimental status and is now being removed due to lack of maintenance and test coverage

enabled_mdev_types = []

list value

The mdev types enabled in the compute node.

Some hardware (e.g. NVIDIA GRID K1) support different mdev types. User can use this option to specify a list of enabled mdev types that may be assigned to a guest instance.

If more than one single mdev type is provided, then for each mdev type an additional section, [mdev_$(MDEV_TYPE)], must be added to the configuration file. Each section then must be configured with a single configuration option, device_addresses, which should be a list of PCI addresses corresponding to the physical GPU(s) or mdev-capable hardware to assign to this type.

If one or more sections are missing (meaning that a specific type is not wanted to use for at least one physical device) or if no device addresses are provided , then Nova will only use the first type that was provided by [devices]/enabled_mdev_types.

If the same PCI address is provided for two different types, nova-compute will return an InvalidLibvirtMdevConfig exception at restart.

As an interim period, old configuration groups named [vgpu_$(MDEV_TYPE)] will be accepted. A valid configuration could then be::

[devices]
enabled_mdev_types = nvidia-35, nvidia-36

[mdev_nvidia-35]
device_addresses = 0000:84:00.0,0000:85:00.0

[vgpu_nvidia-36]
device_addresses = 0000:86:00.0

cipher = aes-xts-plain64

string value

Cipher-mode string to be used.

The cipher and mode to be used to encrypt ephemeral storage. The set of cipher-mode combinations available depends on kernel support. According to the dm-crypt documentation, the cipher is expected to be in the format: "<cipher>-<chainmode>-<ivmode>".

Possible values:

enabled = False

boolean value

Enables/disables LVM ephemeral storage encryption.

key_size = 512

integer value

Encryption key length in bits.

The bit length of the encryption key to be used to encrypt ephemeral storage. In XTS mode only half of the bits are used for encryption key.

aggregate_image_properties_isolation_namespace = None

string value

Image property namespace for use in the host aggregate.

Images and hosts can be configured so that certain images can only be scheduled to hosts in a particular aggregate. This is done with metadata values set on the host aggregate that are identified by beginning with the value of this option. If the host is part of an aggregate with such a metadata key, the image in the request spec must have the value of that metadata in its properties in order for the scheduler to consider the host as acceptable.

Note that this setting only affects scheduling if the AggregateImagePropertiesIsolation filter is enabled.

Possible values:

Related options:

aggregate_image_properties_isolation_separator = .

string value

Separator character(s) for image property namespace and name.

When using the aggregate_image_properties_isolation filter, the relevant metadata keys are prefixed with the namespace defined in the aggregate_image_properties_isolation_namespace configuration option plus a separator. This option defines the separator to be used.

Note that this setting only affects scheduling if the AggregateImagePropertiesIsolation filter is enabled.

Possible values:

Related options:

available_filters = ['nova.scheduler.filters.all_filters']

multi valued

Filters that the scheduler can use.

An unordered list of the filter classes the nova scheduler may apply. Only the filters specified in the [filter_scheduler] enabled_filters option will be used, but any filter appearing in that option must also be included in this list.

By default, this is set to all filters that are included with nova.

Possible values:

Related options:

build_failure_weight_multiplier = 1000000.0

floating point value

Multiplier used for weighing hosts that have had recent build failures.

This option determines how much weight is placed on a compute node with recent build failures. Build failures may indicate a failing, misconfigured, or otherwise ailing compute node, and avoiding it during scheduling may be beneficial. The weight is inversely proportional to the number of recent build failures the compute node has experienced. This value should be set to some high value to offset weight given by other enabled weighers due to available resources. To disable weighing compute hosts by the number of recent failures, set this to zero.

Note that this setting only affects scheduling if the BuildFailureWeigher weigher is enabled.

Possible values:

Related options:

cpu_weight_multiplier = 1.0

floating point value

CPU weight multiplier ratio.

Multiplier used for weighting free vCPUs. Negative numbers indicate stacking rather than spreading.

Note that this setting only affects scheduling if the CPUWeigher weigher is enabled.

Possible values:

Related options:

cross_cell_move_weight_multiplier = 1000000.0

floating point value

Multiplier used for weighing hosts during a cross-cell move.

This option determines how much weight is placed on a host which is within the same source cell when moving a server, for example during cross-cell resize. By default, when moving an instance, the scheduler will prefer hosts within the same cell since cross-cell move operations can be slower and riskier due to the complicated nature of cross-cell migrations.

Note that this setting only affects scheduling if the CrossCellWeigher weigher is enabled. If your cloud is not configured to support cross-cell migrations, then this option has no effect.

The value of this configuration option can be overridden per host aggregate by setting the aggregate metadata key with the same name (cross_cell_move_weight_multiplier).

Possible values:

Related options:

disk_weight_multiplier = 1.0

floating point value

Disk weight multipler ratio.

Multiplier used for weighing free disk space. Negative numbers mean to stack vs spread.

Note that this setting only affects scheduling if the DiskWeigher weigher is enabled.

Possible values:

enabled_filters = ['ComputeFilter', 'ComputeCapabilitiesFilter', 'ImagePropertiesFilter', 'ServerGroupAntiAffinityFilter', 'ServerGroupAffinityFilter']

list value

Filters that the scheduler will use.

An ordered list of filter class names that will be used for filtering hosts. These filters will be applied in the order they are listed so place your most restrictive filters first to make the filtering process more efficient.

All of the filters in this option must be present in the [scheduler_filter] available_filter option, or a SchedulerHostFilterNotFound exception will be raised.

Possible values:

Related options:

host_subset_size = 1

integer value

Size of subset of best hosts selected by scheduler.

New instances will be scheduled on a host chosen randomly from a subset of the N best hosts, where N is the value set by this option.

Setting this to a value greater than 1 will reduce the chance that multiple scheduler processes handling similar requests will select the same host, creating a potential race condition. By selecting a host randomly from the N hosts that best fit the request, the chance of a conflict is reduced. However, the higher you set this value, the less optimal the chosen host may be for a given request.

Possible values:

hypervisor_version_weight_multiplier = 1.0

floating point value

Hypervisor Version weight multiplier ratio.

The multiplier is used for weighting hosts based on the reported hypervisor version. Negative numbers indicate preferring older hosts, the default is to prefer newer hosts to aid with upgrades.

Possible values:

Example:

Related options:

image_properties_default_architecture = None

string value

The default architecture to be used when using the image properties filter.

When using the ImagePropertiesFilter, it is possible that you want to define a default architecture to make the user experience easier and avoid having something like x86_64 images landing on AARCH64 compute nodes because the user did not specify the hw_architecture property in Glance.

Possible values:

io_ops_weight_multiplier = -1.0

floating point value

IO operations weight multipler ratio.

This option determines how hosts with differing workloads are weighed. Negative values, such as the default, will result in the scheduler preferring hosts with lighter workloads whereas positive values will prefer hosts with heavier workloads. Another way to look at it is that positive values for this option will tend to schedule instances onto hosts that are already busy, while negative values will tend to distribute the workload across more hosts. The absolute value, whether positive or negative, controls how strong the io_ops weigher is relative to other weighers.

Note that this setting only affects scheduling if the IoOpsWeigher weigher is enabled.

Possible values:

Related options:

isolated_hosts = []

list value

List of hosts that can only run certain images.

If there is a need to restrict some images to only run on certain designated hosts, list those host names here.

Note that this setting only affects scheduling if the IsolatedHostsFilter filter is enabled.

Possible values:

Related options:

isolated_images = []

list value

List of UUIDs for images that can only be run on certain hosts.

If there is a need to restrict some images to only run on certain designated hosts, list those image UUIDs here.

Note that this setting only affects scheduling if the IsolatedHostsFilter filter is enabled.

Possible values:

Related options:

max_instances_per_host = 50

integer value

Maximum number of instances that can exist on a host.

If you need to limit the number of instances on any given host, set this option to the maximum number of instances you want to allow. The NumInstancesFilter and AggregateNumInstancesFilter will reject any host that has at least as many instances as this option’s value.

Note that this setting only affects scheduling if the NumInstancesFilter or AggregateNumInstancesFilter filter is enabled.

Possible values:

Related options:

max_io_ops_per_host = 8

integer value

The number of instances that can be actively performing IO on a host.

Instances performing IO includes those in the following states: build, resize, snapshot, migrate, rescue, unshelve.

Note that this setting only affects scheduling if the IoOpsFilter filter is enabled.

Possible values:

Related options:

pci_in_placement = False

boolean value

Enable scheduling and claiming PCI devices in Placement.

This can be enabled after [pci]report_in_placement is enabled on all compute hosts.

When enabled the scheduler queries Placement about the PCI device availability to select destination for a server with PCI request. The scheduler also allocates the selected PCI devices in Placement. Note that this logic does not replace the PCIPassthroughFilter but extends it.

pci_weight_multiplier = 1.0

floating point value

PCI device affinity weight multiplier.

The PCI device affinity weighter computes a weighting based on the number of PCI devices on the host and the number of PCI devices requested by the instance.

Note that this setting only affects scheduling if the PCIWeigher weigher and NUMATopologyFilter filter are enabled.

Possible values:

Related options:

ram_weight_multiplier = 1.0

floating point value

RAM weight multipler ratio.

This option determines how hosts with more or less available RAM are weighed. A positive value will result in the scheduler preferring hosts with more available RAM, and a negative number will result in the scheduler preferring hosts with less available RAM. Another way to look at it is that positive values for this option will tend to spread instances across many hosts, while negative values will tend to fill up (stack) hosts as much as possible before scheduling to a less-used host. The absolute value, whether positive or negative, controls how strong the RAM weigher is relative to other weighers.

Note that this setting only affects scheduling if the RAMWeigher weigher is enabled.

Possible values:

Related options:

restrict_isolated_hosts_to_isolated_images = True

boolean value

Prevent non-isolated images from being built on isolated hosts.

Note that this setting only affects scheduling if the IsolatedHostsFilter filter is enabled. Even then, this option doesn’t affect the behavior of requests for isolated images, which will always be restricted to isolated hosts.

Related options:

shuffle_best_same_weighed_hosts = False

boolean value

Enable spreading the instances between hosts with the same best weight.

Enabling it is beneficial for cases when [filter_scheduler] host_subset_size is 1 (default), but there is a large number of hosts with same maximal weight. This scenario is common in Ironic deployments where there are typically many baremetal nodes with identical weights returned to the scheduler. In such case enabling this option will reduce contention and chances for rescheduling events. At the same time it will make the instance packing (even in unweighed case) less dense.

soft_affinity_weight_multiplier = 1.0

floating point value

Multiplier used for weighing hosts for group soft-affinity.

Note that this setting only affects scheduling if the ServerGroupSoftAffinityWeigher weigher is enabled.

Possible values:

Related options:

soft_anti_affinity_weight_multiplier = 1.0

floating point value

Multiplier used for weighing hosts for group soft-anti-affinity.

Note that this setting only affects scheduling if the ServerGroupSoftAntiAffinityWeigher weigher is enabled.

Possible values:

Related options:

track_instance_changes = True

boolean value

Enable querying of individual hosts for instance information.

The scheduler may need information about the instances on a host in order to evaluate its filters and weighers. The most common need for this information is for the (anti-)affinity filters, which need to choose a host based on the instances already running on a host.

If the configured filters and weighers do not need this information, disabling this option will improve performance. It may also be disabled when the tracking overhead proves too heavy, although this will cause classes requiring host usage data to query the database on each request instead.

Related options:

weight_classes = ['nova.scheduler.weights.all_weighers']

list value

Weighers that the scheduler will use.

Only hosts which pass the filters are weighed. The weight for any host starts at 0, and the weighers order these hosts by adding to or subtracting from the weight assigned by the previous weigher. Weights may become negative. An instance will be scheduled to one of the N most-weighted hosts, where N is [filter_scheduler] host_subset_size.

By default, this is set to all weighers that are included with Nova.

Possible values:

api_servers = None

list value

List of glance api servers endpoints available to nova.

https is used for ssl-based glance api servers.

Possible values:

Deprecated since: 21.0.0

Reason: Support for image service configuration via standard keystoneauth1 Adapter options was added in the 17.0.0 Queens release. The api_servers option was retained temporarily to allow consumers time to cut over to a real load balancing solution.

cafile = None

string value

PEM encoded Certificate Authority to use when verifying HTTPs connections.

certfile = None

string value

PEM encoded client certificate cert file

collect-timing = False

boolean value

Collect per-API call timing information.

connect-retries = None

integer value

The maximum number of retries that should be attempted for connection errors.

connect-retry-delay = None

floating point value

Delay (in seconds) between two retries for connection errors. If not set, exponential retry starting with 0.5 seconds up to a maximum of 60 seconds is used.

debug = False

boolean value

Enable or disable debug logging with glanceclient.

default_trusted_certificate_ids = []

list value

List of certificate IDs for certificates that should be trusted.

May be used as a default list of trusted certificate IDs for certificate validation. The value of this option will be ignored if the user provides a list of trusted certificate IDs with an instance API request. The value of this option will be persisted with the instance data if signature verification and certificate validation are enabled and if the user did not provide an alternative list. If left empty when certificate validation is enabled the user must provide a list of trusted certificate IDs otherwise certificate validation will fail.

Related options:

enable_certificate_validation = False

boolean value

Enable certificate validation for image signature verification.

During image signature verification nova will first verify the validity of the image’s signing certificate using the set of trusted certificates associated with the instance. If certificate validation fails, signature verification will not be performed and the instance will be placed into an error state. This provides end users with stronger assurances that the image data is unmodified and trustworthy. If left disabled, image signature verification can still occur but the end user will not have any assurance that the signing certificate used to generate the image signature is still trustworthy.

Related options:

Deprecated since: 16.0.0

Reason: This option is intended to ease the transition for deployments leveraging image signature verification. The intended state long-term is for signature verification and certificate validation to always happen together.

enable_rbd_download = False

boolean value

Enable Glance image downloads directly via RBD.

Allow non-rbd computes using local storage to download and cache images from Ceph via rbd rather than the Glance API via http.

Related options:

endpoint-override = None

string value

Always use this endpoint URL for requests for this client. NOTE: The unversioned endpoint should be specified here; to request a particular API version, use the version, min-version, and/or max-version options.

insecure = False

boolean value

Verify HTTPS connections.

keyfile = None

string value

PEM encoded client certificate key file

num_retries = 3

integer value

Enable glance operation retries.

Specifies the number of retries when uploading / downloading an image to / from glance. 0 means no retries.

`rbd_ceph_conf = `

string value

Path to the ceph configuration file to use.

Related options:

rbd_connect_timeout = 5

integer value

The RADOS client timeout in seconds when initially connecting to the cluster.

Related options:

`rbd_pool = `

string value

The RADOS pool in which the Glance images are stored as rbd volumes.

Related options:

`rbd_user = `

string value

The RADOS client name for accessing Glance images stored as rbd volumes.

Related options:

region-name = None

string value

The default region_name for endpoint URL discovery.

service-name = None

string value

The default service_name for endpoint URL discovery.

service-type = image

string value

The default service_type for endpoint URL discovery.

split-loggers = False

boolean value

Log requests to multiple loggers.

status-code-retries = None

integer value

The maximum number of retries that should be attempted for retriable HTTP status codes.

status-code-retry-delay = None

floating point value

Delay (in seconds) between two retries for retriable status codes. If not set, exponential retry starting with 0.5 seconds up to a maximum of 60 seconds is used.

timeout = None

integer value

Timeout value for http requests

valid-interfaces = ['internal', 'public']

list value

List of interfaces, in order of preference, for endpoint URL.

verify_glance_signatures = False

boolean value

Enable image signature verification.

nova uses the image signature metadata from glance and verifies the signature of a signed image while downloading that image. If the image signature cannot be verified or if the image signature metadata is either incomplete or unavailable, then nova will not boot the image and instead will place the instance into an error state. This provides end users with stronger assurances of the integrity of the image data they are using to create servers.

Related options:

debug = False

boolean value

Enable/disables guestfs logging.

This configures guestfs to debug messages and push them to OpenStack logging system. When set to True, it traces libguestfs API calls and enable verbose debug messages. In order to use the above feature, "libguestfs" package must be installed.

Related options:

Since libguestfs access and modifies VM’s managed by libvirt, below options should be set to give access to those VM’s.

backends = []

list value

Additional backends that can perform health checks and report that information back as part of a request.

detailed = False

boolean value

Show more detailed information as part of the response. Security note: Enabling this option may expose sensitive details about the service being monitored. Be sure to verify that it will not violate your security policies.

disable_by_file_path = None

string value

Check the presence of a file to determine if an application is running on a port. Used by DisableByFileHealthcheck plugin.

disable_by_file_paths = []

list value

Check the presence of a file based on a port to determine if an application is running on a port. Expects a "port:path" list of strings. Used by DisableByFilesPortsHealthcheck plugin.

path = /healthcheck

string value

The path to respond to healtcheck requests on.

config_drive_cdrom = False

boolean value

Mount config drive as a CD drive.

OpenStack can be configured to write instance metadata to a config drive, which is then attached to the instance before it boots. The config drive can be attached as a disk drive (default) or as a CD drive.

Related options:

config_drive_inject_password = False

boolean value

Inject password to config drive.

When enabled, the admin password will be available from the config drive image.

Related options:

dynamic_memory_ratio = 1.0

floating point value

Dynamic memory ratio

Enables dynamic memory allocation (ballooning) when set to a value greater than 1. The value expresses the ratio between the total RAM assigned to an instance and its startup RAM amount. For example a ratio of 2.0 for an instance with 1024MB of RAM implies 512MB of RAM allocated at startup.

Possible values:

enable_instance_metrics_collection = False

boolean value

Enable instance metrics collection

Enables metrics collections for an instance by using Hyper-V’s metric APIs. Collected data can be retrieved by other apps and services, e.g.: Ceilometer.

enable_remotefx = False

boolean value

Enable RemoteFX feature

This requires at least one DirectX 11 capable graphics adapter for Windows / Hyper-V Server 2012 R2 or newer and RDS-Virtualization feature has to be enabled.

Instances with RemoteFX can be requested with the following flavor extra specs:

3840x2160 is only available on Windows / Hyper-V Server 2016.

os:vram. Guest VM VRAM amount. Only available on Windows / Hyper-V Server 2016. Acceptable values::

64, 128, 256, 512, 1024

`instances_path_share = `

string value

Instances path share

The name of a Windows share mapped to the "instances_path" dir and used by the resize feature to copy files to the target host. If left blank, an administrative share (hidden network share) will be used, looking for the same "instances_path" used locally.

Possible values:

Related options:

iscsi_initiator_list = []

list value

List of iSCSI initiators that will be used for establishing iSCSI sessions.

If none are specified, the Microsoft iSCSI initiator service will choose the initiator.

limit_cpu_features = False

boolean value

Limit CPU features

This flag is needed to support live migration to hosts with different CPU features and checked during instance creation in order to limit the CPU features used by the instance.

mounted_disk_query_retry_count = 10

integer value

Mounted disk query retry count

The number of times to retry checking for a mounted disk. The query runs until the device can be found or the retry count is reached.

Possible values:

Related options:

mounted_disk_query_retry_interval = 5

integer value

Mounted disk query retry interval

Interval between checks for a mounted disk, in seconds.

Possible values:

Related options:

power_state_check_timeframe = 60

integer value

Power state check timeframe

The timeframe to be checked for instance power state changes. This option is used to fetch the state of the instance from Hyper-V through the WMI interface, within the specified timeframe.

Possible values:

power_state_event_polling_interval = 2

integer value

Power state event polling interval

Instance power state change event polling frequency. Sets the listener interval for power state events to the given value. This option enhances the internal lifecycle notifications of instances that reboot themselves. It is unlikely that an operator has to change this value.

Possible values:

qemu_img_cmd = qemu-img.exe

string value

qemu-img command

qemu-img is required for some of the image related operations like converting between different image types. You can get it from here: (http://qemu.weilnetz.de/) or you can install the Cloudbase OpenStack Hyper-V Compute Driver (https://cloudbase.it/openstack-hyperv-driver/) which automatically sets the proper path for this config option. You can either give the full path of qemu-img.exe or set its path in the PATH environment variable and leave this option to the default value.

Possible values:

Related options:

use_multipath_io = False

boolean value

Use multipath connections when attaching iSCSI or FC disks.

This requires the Multipath IO Windows feature to be enabled. MPIO must be configured to claim such devices.

volume_attach_retry_count = 10

integer value

Volume attach retry count

The number of times to retry attaching a volume. Volume attachment is retried until success or the given retry count is reached.

Possible values:

Related options:

volume_attach_retry_interval = 5

integer value

Volume attach retry interval

Interval between volume attachment attempts, in seconds.

Possible values:

Related options:

vswitch_name = None

string value

External virtual switch name

The Hyper-V Virtual Switch is a software-based layer-2 Ethernet network switch that is available with the installation of the Hyper-V server role. The switch includes programmatically managed and extensible capabilities to connect virtual machines to both virtual networks and the physical network. In addition, Hyper-V Virtual Switch provides policy enforcement for security, isolation, and service levels. The vSwitch represented by this config option must be an external one (not internal or private).

Possible values:

wait_soft_reboot_seconds = 60

integer value

Wait soft reboot seconds

Number of seconds to wait for instance to shut down after soft reboot request is made. We fall back to hard reboot if instance does not shutdown within this window.

Possible values:

manager_interval = 2400

integer value

Number of seconds to wait between runs of the image cache manager.

Note that when using shared storage for the [DEFAULT]/instances_path configuration option across multiple nova-compute services, this periodic could process a large number of instances. Similarly, using a compute driver that manages a cluster (like vmwareapi.VMwareVCDriver) could result in processing a large number of instances. Therefore you may need to adjust the time interval for the anticipated load, or only run on one nova-compute service within a shared storage aggregate. Additional note, every time the image_cache_manager runs the timestamps of images in [DEFAULT]/instances_path are updated.

Possible values:

Related options:

precache_concurrency = 1

integer value

Maximum number of compute hosts to trigger image precaching in parallel.

When an image precache request is made, compute nodes will be contacted to initiate the download. This number constrains the number of those that will happen in parallel. Higher numbers will cause more computes to work in parallel and may result in reduced time to complete the operation, but may also DDoS the image service. Lower numbers will result in more sequential operation, lower image service load, but likely longer runtime to completion.

remove_unused_base_images = True

boolean value

Should unused base images be removed?

When there are no remaining instances on the hypervisor created from this base image or linked to it, the base image is considered unused.

remove_unused_original_minimum_age_seconds = 86400

integer value

Unused unresized base images younger than this will not be removed.

remove_unused_resized_minimum_age_seconds = 3600

integer value

Unused resized base images younger than this will not be removed.

subdirectory_name = _base

string value

Location of cached images.

This is NOT the full path - just a folder name relative to $instances_path. For per-compute-host cached images, set to base$my_ip

api_max_retries = 60

integer value

The number of times to retry when a request conflicts. If set to 0, only try once, no retries.

Related options:

api_retry_interval = 2

integer value

The number of seconds to wait before retrying the request.

Related options:

auth-url = None

string value

Authentication URL

auth_section = None

string value

Config Section from which to load plugin specific options

auth_type = None

string value

Authentication type to load

cafile = None

string value

PEM encoded Certificate Authority to use when verifying HTTPs connections.

certfile = None

string value

PEM encoded client certificate cert file

collect-timing = False

boolean value

Collect per-API call timing information.

connect-retries = None

integer value

The maximum number of retries that should be attempted for connection errors.

connect-retry-delay = None

floating point value

Delay (in seconds) between two retries for connection errors. If not set, exponential retry starting with 0.5 seconds up to a maximum of 60 seconds is used.

domain-id = None

string value

Domain ID to scope to

domain-name = None

string value

Domain name to scope to

endpoint-override = None

string value

Always use this endpoint URL for requests for this client. NOTE: The unversioned endpoint should be specified here; to request a particular API version, use the version, min-version, and/or max-version options.

insecure = False

boolean value

Verify HTTPS connections.

keyfile = None

string value

PEM encoded client certificate key file

partition_key = None

string value

Case-insensitive key to limit the set of nodes that may be managed by this service to the set of nodes in Ironic which have a matching conductor_group property. If unset, all available nodes will be eligible to be managed by this service. Note that setting this to the empty string ("") will match the default conductor group, and is different than leaving the option unset.

password = None

string value

User’s password

peer_list = []

list value

List of hostnames for all nova-compute services (including this host) with this partition_key config value. Nodes matching the partition_key value will be distributed between all services specified here. If partition_key is unset, this option is ignored.

project-domain-id = None

string value

Domain ID containing project

project-domain-name = None

string value

Domain name containing project

project-id = None

string value

Project ID to scope to

project-name = None

string value

Project name to scope to

region-name = None

string value

The default region_name for endpoint URL discovery.

serial_console_state_timeout = 10

integer value

Timeout (seconds) to wait for node serial console state changed. Set to 0 to disable timeout.

service-name = None

string value

The default service_name for endpoint URL discovery.

service-type = baremetal

string value

The default service_type for endpoint URL discovery.

split-loggers = False

boolean value

Log requests to multiple loggers.

status-code-retries = None

integer value

The maximum number of retries that should be attempted for retriable HTTP status codes.

status-code-retry-delay = None

floating point value

Delay (in seconds) between two retries for retriable status codes. If not set, exponential retry starting with 0.5 seconds up to a maximum of 60 seconds is used.

system-scope = None

string value

Scope for system operations

timeout = None

integer value

Timeout value for http requests

trust-id = None

string value

ID of the trust to use as a trustee use

user-domain-id = None

string value

User’s domain id

user-domain-name = None

string value

User’s domain name

user-id = None

string value

User ID

username = None

string value

Username

valid-interfaces = ['internal', 'public']

list value

List of interfaces, in order of preference, for endpoint URL.

auth_type = None

string value

The type of authentication credential to create. Possible values are token, password, keystone_token, and keystone_password. Required if no context is passed to the credential factory.

auth_url = None

string value

Use this endpoint to connect to Keystone.

backend = barbican

string value

Specify the key manager implementation. Options are "barbican" and "vault". Default is "barbican". Will support the values earlier set using [key_manager]/api_class for some time.

domain_id = None

string value

Domain ID for domain scoping. Optional for keystone_token and keystone_password auth_type.

domain_name = None

string value

Domain name for domain scoping. Optional for keystone_token and keystone_password auth_type.

fixed_key = None

string value

Fixed key returned by key manager, specified in hex.

Possible values:

password = None

string value

Password for authentication. Required for password and keystone_password auth_type.

project_domain_id = None

string value

Project’s domain ID for project. Optional for keystone_token and keystone_password auth_type.

project_domain_name = None

string value

Project’s domain name for project. Optional for keystone_token and keystone_password auth_type.

project_id = None

string value

Project ID for project scoping. Optional for keystone_token and keystone_password auth_type.

project_name = None

string value

Project name for project scoping. Optional for keystone_token and keystone_password auth_type.

reauthenticate = True

boolean value

Allow fetching a new token if the current one is going to expire. Optional for keystone_token and keystone_password auth_type.

token = None

string value

Token for authentication. Required for token and keystone_token auth_type if no context is passed to the credential factory.

trust_id = None

string value

Trust ID for trust scoping. Optional for keystone_token and keystone_password auth_type.

user_domain_id = None

string value

User’s domain ID for authentication. Optional for keystone_token and keystone_password auth_type.

user_domain_name = None

string value

User’s domain name for authentication. Optional for keystone_token and keystone_password auth_type.

user_id = None

string value

User ID for authentication. Optional for keystone_token and keystone_password auth_type.

username = None

string value

Username for authentication. Required for password auth_type. Optional for the keystone_password auth_type.

auth-url = None

string value

Authentication URL

auth_section = None

string value

Config Section from which to load plugin specific options

auth_type = None

string value

Authentication type to load

cafile = None

string value

PEM encoded Certificate Authority to use when verifying HTTPs connections.

certfile = None

string value

PEM encoded client certificate cert file

collect-timing = False

boolean value

Collect per-API call timing information.

connect-retries = None

integer value

The maximum number of retries that should be attempted for connection errors.

connect-retry-delay = None

floating point value

Delay (in seconds) between two retries for connection errors. If not set, exponential retry starting with 0.5 seconds up to a maximum of 60 seconds is used.

default-domain-id = None

string value

Optional domain ID to use with v3 and v2 parameters. It will be used for both the user and project domain in v3 and ignored in v2 authentication.

default-domain-name = None

string value

Optional domain name to use with v3 API and v2 parameters. It will be used for both the user and project domain in v3 and ignored in v2 authentication.

domain-id = None

string value

Domain ID to scope to

domain-name = None

string value

Domain name to scope to

endpoint-override = None

string value

Always use this endpoint URL for requests for this client. NOTE: The unversioned endpoint should be specified here; to request a particular API version, use the version, min-version, and/or max-version options.

insecure = False

boolean value

Verify HTTPS connections.

keyfile = None

string value

PEM encoded client certificate key file

password = None

string value

User’s password

project-domain-id = None

string value

Domain ID containing project

project-domain-name = None

string value

Domain name containing project

project-id = None

string value

Project ID to scope to

project-name = None

string value

Project name to scope to

region-name = None

string value

The default region_name for endpoint URL discovery.

service-name = None

string value

The default service_name for endpoint URL discovery.

service-type = identity

string value

The default service_type for endpoint URL discovery.

split-loggers = False

boolean value

Log requests to multiple loggers.

status-code-retries = None

integer value

The maximum number of retries that should be attempted for retriable HTTP status codes.

status-code-retry-delay = None

floating point value

Delay (in seconds) between two retries for retriable status codes. If not set, exponential retry starting with 0.5 seconds up to a maximum of 60 seconds is used.

system-scope = None

string value

Scope for system operations

tenant-id = None

string value

Tenant ID

tenant-name = None

string value

Tenant Name

timeout = None

integer value

Timeout value for http requests

trust-id = None

string value

ID of the trust to use as a trustee use

user-domain-id = None

string value

User’s domain id

user-domain-name = None

string value

User’s domain name

user-id = None

string value

User ID

username = None

string value

Username

valid-interfaces = ['internal', 'public']

list value

List of interfaces, in order of preference, for endpoint URL.

auth_section = None

string value

Config Section from which to load plugin specific options

auth_type = None

string value

Authentication type to load

auth_uri = None

string value

Complete "public" Identity API endpoint. This endpoint should not be an "admin" endpoint, as it should be accessible by all end users. Unauthenticated clients are redirected to this endpoint to authenticate. Although this endpoint should ideally be unversioned, client support in the wild varies. If you’re using a versioned v2 endpoint here, then this should not be the same endpoint the service user utilizes for validating tokens, because normal end users may not be able to reach that endpoint. This option is deprecated in favor of www_authenticate_uri and will be removed in the S release. Deprecated since: Queens

*Reason:*The auth_uri option is deprecated in favor of www_authenticate_uri and will be removed in the S release.

auth_version = None

string value

API version of the Identity API endpoint.

cache = None

string value

Request environment key where the Swift cache object is stored. When auth_token middleware is deployed with a Swift cache, use this option to have the middleware share a caching backend with swift. Otherwise, use the memcached_servers option instead.

cafile = None

string value

A PEM encoded Certificate Authority to use when verifying HTTPs connections. Defaults to system CAs.

certfile = None

string value

Required if identity server requires client certificate

delay_auth_decision = False

boolean value

Do not handle authorization requests within the middleware, but delegate the authorization decision to downstream WSGI components.

enforce_token_bind = permissive

string value

Used to control the use and type of token binding. Can be set to: "disabled" to not check token binding. "permissive" (default) to validate binding information if the bind type is of a form known to the server and ignore it if not. "strict" like "permissive" but if the bind type is unknown the token will be rejected. "required" any form of token binding is needed to be allowed. Finally the name of a binding method that must be present in tokens.

http_connect_timeout = None

integer value

Request timeout value for communicating with Identity API server.

http_request_max_retries = 3

integer value

How many times are we trying to reconnect when communicating with Identity API Server.

include_service_catalog = True

boolean value

(Optional) Indicate whether to set the X-Service-Catalog header. If False, middleware will not ask for service catalog on token validation and will not set the X-Service-Catalog header.

insecure = False

boolean value

Verify HTTPS connections.

interface = internal

string value

Interface to use for the Identity API endpoint. Valid values are "public", "internal" (default) or "admin".

keyfile = None

string value

Required if identity server requires client certificate

memcache_pool_conn_get_timeout = 10

integer value

(Optional) Number of seconds that an operation will wait to get a memcached client connection from the pool.

memcache_pool_dead_retry = 300

integer value

(Optional) Number of seconds memcached server is considered dead before it is tried again.

memcache_pool_maxsize = 10

integer value

(Optional) Maximum total number of open connections to every memcached server.

memcache_pool_socket_timeout = 3

integer value

(Optional) Socket timeout in seconds for communicating with a memcached server.

memcache_pool_unused_timeout = 60

integer value

(Optional) Number of seconds a connection to memcached is held unused in the pool before it is closed.

memcache_secret_key = None

string value

(Optional, mandatory if memcache_security_strategy is defined) This string is used for key derivation.

memcache_security_strategy = None

string value

(Optional) If defined, indicate whether token data should be authenticated or authenticated and encrypted. If MAC, token data is authenticated (with HMAC) in the cache. If ENCRYPT, token data is encrypted and authenticated in the cache. If the value is not one of these options or empty, auth_token will raise an exception on initialization.

memcache_use_advanced_pool = True

boolean value

(Optional) Use the advanced (eventlet safe) memcached client pool.

memcached_servers = None

list value

Optionally specify a list of memcached server(s) to use for caching. If left undefined, tokens will instead be cached in-process.

region_name = None

string value

The region in which the identity server can be found.

service_token_roles = ['service']

list value

A choice of roles that must be present in a service token. Service tokens are allowed to request that an expired token can be used and so this check should tightly control that only actual services should be sending this token. Roles here are applied as an ANY check so any role in this list must be present. For backwards compatibility reasons this currently only affects the allow_expired check.

service_token_roles_required = False

boolean value

For backwards compatibility reasons we must let valid service tokens pass that don’t pass the service_token_roles check as valid. Setting this true will become the default in a future release and should be enabled if possible.

service_type = None

string value

The name or type of the service as it appears in the service catalog. This is used to validate tokens that have restricted access rules.

token_cache_time = 300

integer value

In order to prevent excessive effort spent validating tokens, the middleware caches previously-seen tokens for a configurable duration (in seconds). Set to -1 to disable caching completely.

www_authenticate_uri = None

string value

Complete "public" Identity API endpoint. This endpoint should not be an "admin" endpoint, as it should be accessible by all end users. Unauthenticated clients are redirected to this endpoint to authenticate. Although this endpoint should ideally be unversioned, client support in the wild varies. If you’re using a versioned v2 endpoint here, then this should not be the same endpoint the service user utilizes for validating tokens, because normal end users may not be able to reach that endpoint.

`connection_uri = `

string value

Overrides the default libvirt URI of the chosen virtualization type.

If set, Nova will use this URI to connect to libvirt.

Possible values:

Related options:

cpu_mode = None

string value

Is used to set the CPU mode an instance should have.

If virt_type="kvm|qemu", it will default to host-model, otherwise it will default to none.

Related options:

cpu_model_extra_flags = []

list value

Enable or disable guest CPU flags.

To explicitly enable or disable CPU flags, use the +flag or -flag notation — the + sign will enable the CPU flag for the guest, while a - sign will disable it. If neither + nor - is specified, the flag will be enabled, which is the default behaviour. For example, if you specify the following (assuming the said CPU model and features are supported by the host hardware and software)::

[libvirt]
cpu_mode = custom
cpu_models = Cascadelake-Server
cpu_model_extra_flags = -hle, -rtm, +ssbd, mtrr

Nova will disable the hle and rtm flags for the guest; and it will enable ssbd and mttr (because it was specified with neither + nor - prefix).

The CPU flags are case-insensitive. In the following example, the pdpe1gb flag will be disabled for the guest; vmx and pcid flags will be enabled::

[libvirt]
cpu_mode = custom
cpu_models = Haswell-noTSX-IBRS
cpu_model_extra_flags = -PDPE1GB, +VMX, pcid

Specifying extra CPU flags is valid in combination with all the three possible values of cpu_mode config attribute: custom (this also requires an explicit CPU model to be specified via the cpu_models config attribute), host-model, or host-passthrough.

There can be scenarios where you may need to configure extra CPU flags even for host-passthrough CPU mode, because sometimes QEMU may disable certain CPU features. An example of this is Intel’s "invtsc" (Invariable Time Stamp Counter) CPU flag — if you need to expose this flag to a Nova instance, you need to explicitly enable it.

The possible values for cpu_model_extra_flags depends on the CPU model in use. Refer to /usr/share/libvirt/cpu_map/*.xml for possible CPU feature flags for a given CPU model.

A special note on a particular CPU flag: pcid (an Intel processor feature that alleviates guest performance degradation as a result of applying the Meltdown CVE fixes). When configuring this flag with the custom CPU mode, not all CPU models (as defined by QEMU and libvirt) need it:

The libvirt driver’s default CPU mode, host-model, will do the right thing with respect to handling PCID CPU flag for the guest — assuming you are running updated processor microcode, host and guest kernel, libvirt, and QEMU. The other mode, host-passthrough, checks if PCID is available in the hardware, and if so directly passes it through to the Nova guests. Thus, in context of PCID, with either of these CPU modes (host-model or host-passthrough), there is no need to use the cpu_model_extra_flags.

Related options:

cpu_models = []

list value

An ordered list of CPU models the host supports.

It is expected that the list is ordered so that the more common and less advanced CPU models are listed earlier. Here is an example: SandyBridge,IvyBridge,Haswell,Broadwell, the latter CPU model’s features is richer that the previous CPU model.

Possible values:

Related options:

cpu_power_governor_high = performance

string value

Governor to use in order to have best CPU performance

cpu_power_governor_low = powersave

string value

Governor to use in order to reduce CPU power consumption

cpu_power_management = False

boolean value

Use libvirt to manage CPU cores performance.

cpu_power_management_strategy = cpu_state

string value

Tuning strategy to reduce CPU power consumption when unused

device_detach_attempts = 8

integer value

Maximum number of attempts the driver tries to detach a device in libvirt.

Related options:

device_detach_timeout = 20

integer value

Maximum number of seconds the driver waits for the success or the failure event from libvirt for a given device detach attempt before it re-trigger the detach.

Related options:

disk_cachemodes = []

list value

Specific cache modes to use for different disk types.

For example: file=directsync,block=none,network=writeback

For local or direct-attached storage, it is recommended that you use writethrough (default) mode, as it ensures data integrity and has acceptable I/O performance for applications running in the guest, especially for read operations. However, caching mode none is recommended for remote NFS storage, because direct I/O operations (O_DIRECT) perform better than synchronous I/O operations (with O_SYNC). Caching mode none effectively turns all guest I/O operations into direct I/O operations on the host, which is the NFS client in this environment.

Possible cache modes:

disk_prefix = None

string value

Override the default disk prefix for the devices attached to an instance.

If set, this is used to identify a free disk device name for a bus.

Possible values:

Related options:

enabled_perf_events = []

list value

Performance events to monitor and collect statistics for.

This will allow you to specify a list of events to monitor low-level performance of guests, and collect related statistics via the libvirt driver, which in turn uses the Linux kernel’s perf infrastructure. With this config attribute set, Nova will generate libvirt guest XML to monitor the specified events.

For example, to monitor the count of CPU cycles (total/elapsed) and the count of cache misses, enable them as follows::

[libvirt]
enabled_perf_events = cpu_clock, cache_misses

Possible values: A string list. The list of supported events can be found here`__. Note that Intel CMT events - `cmt, mbmbt and mbml - are unsupported by recent Linux kernel versions (4.14+) and will be ignored by nova.

__ https://libvirt.org/formatdomain.html#elementsPerf.

file_backed_memory = 0

integer value

Available capacity in MiB for file-backed memory.

Set to 0 to disable file-backed memory.

When enabled, instances will create memory files in the directory specified in /etc/libvirt/qemu.conf's memory_backing_dir option. The default location is /var/lib/libvirt/qemu/ram.

When enabled, the value defined for this option is reported as the node memory capacity. Compute node system memory will be used as a cache for file-backed memory, via the kernel’s pagecache mechanism.

Related options:

gid_maps = []

list value

List of guid targets and ranges.Syntax is guest-gid:host-gid:count. Maximum of 5 allowed.

hw_disk_discard = None

string value

Discard option for nova managed disks.

Requires:

hw_machine_type = None

list value

For qemu or KVM guests, set this option to specify a default machine type per host architecture. You can find a list of supported machine types in your environment by checking the output of the :command:virsh capabilities command. The format of the value for this config option is host-arch=machine-type. For example: x86_64=machinetype1,armv7l=machinetype2.

`images_rbd_ceph_conf = `

string value

Path to the ceph configuration file to use

images_rbd_glance_copy_poll_interval = 15

integer value

The interval in seconds with which to poll Glance after asking for it to copy an image to the local rbd store. This affects how often we ask Glance to report on copy completion, and thus should be short enough that we notice quickly, but not too aggressive that we generate undue load on the Glance server.

Related options:

images_rbd_glance_copy_timeout = 600

integer value

The overall maximum time we will wait for Glance to complete an image copy to our local rbd store. This should be long enough to allow large images to be copied over the network link between our local store and the one where images typically reside. The downside of setting this too long is just to catch the case where the image copy is stalled or proceeding too slowly to be useful. Actual errors will be reported by Glance and noticed according to the poll interval.

Related options:

`images_rbd_glance_store_name = `

string value

The name of the Glance store that represents the rbd cluster in use by this node. If set, this will allow Nova to request that Glance copy an image from an existing non-local store into the one named by this option before booting so that proper Copy-on-Write behavior is maintained.

Related options:

images_rbd_pool = rbd

string value

The RADOS pool in which rbd volumes are stored

images_type = default

string value

VM Images format.

If default is specified, then use_cow_images flag is used instead of this one.

Related options:

images_volume_group = None

string value

LVM Volume Group that is used for VM images, when you specify images_type=lvm

Related options:

inject_key = False

boolean value

Allow the injection of an SSH key at boot time.

There is no agent needed within the image to do this. If libguestfs is available on the host, it will be used. Otherwise nbd is used. The file system of the image will be mounted and the SSH key, which is provided in the REST API call will be injected as SSH key for the root user and appended to the authorized_keys of that user. The SELinux context will be set if necessary. Be aware that the injection is not possible when the instance gets launched from a volume.

This config option will enable directly modifying the instance disk and does not affect what cloud-init may do using data from config_drive option or the metadata service.

Linux distribution guest only.

Related options:

inject_partition = -2

integer value

Determines how the file system is chosen to inject data into it.

libguestfs is used to inject data. If libguestfs is not able to determine the root partition (because there are more or less than one root partition) or cannot mount the file system it will result in an error and the instance won’t boot.

Possible values:

Linux distribution guest only.

Related options:

inject_password = False

boolean value

Allow the injection of an admin password for instance only at create and rebuild process.

There is no agent needed within the image to do this. If libguestfs is available on the host, it will be used. Otherwise nbd is used. The file system of the image will be mounted and the admin password, which is provided in the REST API call will be injected as password for the root user. If no root user is available, the instance won’t be launched and an error is thrown. Be aware that the injection is not possible when the instance gets launched from a volume.

Linux distribution guest only.

Possible values:

Related options:

iscsi_iface = None

string value

The iSCSI transport iface to use to connect to target in case offload support is desired.

Default format is of the form <transport_name>.<hwaddress>, where <transport_name> is one of (be2iscsi, bnx2i, cxgb3i, cxgb4i, qla4xxx, ocs, tcp) and <hwaddress> is the MAC address of the interface and can be generated via the iscsiadm -m iface command. Do not confuse the iscsi_iface parameter to be provided here with the actual transport name.

iser_use_multipath = False

boolean value

Use multipath connection of the iSER volume.

iSER volumes can be connected as multipath devices. This will provide high availability and fault tolerance.

live_migration_bandwidth = 0

integer value

Maximum bandwidth(in MiB/s) to be used during migration.

If set to 0, the hypervisor will choose a suitable default. Some hypervisors do not support this feature and will return an error if bandwidth is not 0. Please refer to the libvirt documentation for further details.

live_migration_completion_timeout = 800

integer value

Time to wait, in seconds, for migration to successfully complete transferring data before aborting the operation.

Value is per GiB of guest RAM + disk to be transferred, with lower bound of a minimum of 2 GiB. Should usually be larger than downtime delay * downtime steps. Set to 0 to disable timeouts.

Related options:

live_migration_downtime = 500

integer value

Target maximum period of time Nova will try to keep the instance paused during the last part of the memory copy, in milliseconds.

Will be rounded up to a minimum of 100ms. You can increase this value if you want to allow live-migrations to complete faster, or avoid live-migration timeout errors by allowing the guest to be paused for longer during the live-migration switch over. This value may be exceeded if there is any reduction on the transfer rate after the VM is paused.

Related options:

live_migration_downtime_delay = 75

integer value

Time to wait, in seconds, between each step increase of the migration downtime.

Minimum delay is 3 seconds. Value is per GiB of guest RAM + disk to be transferred, with lower bound of a minimum of 2 GiB per device.

live_migration_downtime_steps = 10

integer value

Number of incremental steps to reach max downtime value.

Will be rounded up to a minimum of 3 steps.

live_migration_inbound_addr = None

host domain value

IP address used as the live migration address for this host.

This option indicates the IP address which should be used as the target for live migration traffic when migrating to this hypervisor. This metadata is then used by the source of the live migration traffic to construct a migration URI.

If this option is set to None, the hostname of the migration target compute node will be used.

This option is useful in environments where the live-migration traffic can impact the network plane significantly. A separate network for live-migration traffic can then use this config option and avoids the impact on the management network.

live_migration_permit_auto_converge = False

boolean value

This option allows nova to start live migration with auto converge on.

Auto converge throttles down CPU if a progress of on-going live migration is slow. Auto converge will only be used if this flag is set to True and post copy is not permitted or post copy is unavailable due to the version of libvirt and QEMU in use.

Related options:

live_migration_permit_post_copy = False

boolean value

This option allows nova to switch an on-going live migration to post-copy mode, i.e., switch the active VM to the one on the destination node before the migration is complete, therefore ensuring an upper bound on the memory that needs to be transferred. Post-copy requires libvirt>=1.3.3 and QEMU>=2.5.0.

When permitted, post-copy mode will be automatically activated if we reach the timeout defined by live_migration_completion_timeout and live_migration_timeout_action is set to force_complete. Note if you change to no timeout or choose to use abort, i.e. live_migration_completion_timeout = 0, then there will be no automatic switch to post-copy.

The live-migration force complete API also uses post-copy when permitted. If post-copy mode is not available, force complete falls back to pausing the VM to ensure the live-migration operation will complete.

When using post-copy mode, if the source and destination hosts lose network connectivity, the VM being live-migrated will need to be rebooted. For more details, please see the Administration guide.

Related options:

live_migration_scheme = None

string value

URI scheme for live migration used by the source of live migration traffic.

Override the default libvirt live migration scheme (which is dependent on virt_type). If this option is set to None, nova will automatically choose a sensible default based on the hypervisor. It is not recommended that you change this unless you are very sure that hypervisor supports a particular scheme.

Related options:

live_migration_timeout_action = abort

string value

This option will be used to determine what action will be taken against a VM after live_migration_completion_timeout expires. By default, the live migrate operation will be aborted after completion timeout. If it is set to force_complete, the compute service will either pause the VM or trigger post-copy depending on if post copy is enabled and available (live_migration_permit_post_copy is set to True).

Related options:

live_migration_tunnelled = False

boolean value

Enable tunnelled migration.

This option enables the tunnelled migration feature, where migration data is transported over the libvirtd connection. If enabled, we use the VIR_MIGRATE_TUNNELLED migration flag, avoiding the need to configure the network to allow direct hypervisor to hypervisor communication. If False, use the native transport. If not set, Nova will choose a sensible default based on, for example the availability of native encryption support in the hypervisor. Enabling this option will definitely impact performance massively.

Note that this option is NOT compatible with use of block migration.

Deprecated since: 23.0.0

Reason: The "tunnelled live migration" has two inherent limitations: it cannot handle live migration of disks in a non-shared storage setup; and it has a huge performance cost. Both these problems are solved by ``live_migration_with_native_tls`` (requires a pre-configured TLS environment), which is the recommended approach for securing all live migration streams.

live_migration_uri = None

string value

Live migration target URI used by the source of live migration traffic.

Override the default libvirt live migration target URI (which is dependent on virt_type). Any included "%s" is replaced with the migration target hostname, or live_migration_inbound_addr if set.

If this option is set to None (which is the default), Nova will automatically generate the live_migration_uri value based on only 4 supported virt_type in following list:

Related options:

Deprecated since: 15.0.0

Reason: live_migration_uri is deprecated for removal in favor of two other options that allow to change live migration scheme and target URI: ``live_migration_scheme`` and ``live_migration_inbound_addr`` respectively.

live_migration_with_native_tls = False

boolean value

Use QEMU-native TLS encryption when live migrating.

This option will allow both migration stream (guest RAM plus device state) and disk stream to be transported over native TLS, i.e. TLS support built into QEMU.

Prerequisite: TLS environment is configured correctly on all relevant Compute nodes. This means, Certificate Authority (CA), server, client certificates, their corresponding keys, and their file permissions are in place, and are validated.

Notes:

Related options:

live_migration_tunnelled: This transports migration stream (but not disk stream) over libvirtd.

max_queues = None

integer value

The maximum number of virtio queue pairs that can be enabled when creating a multiqueue guest. The number of virtio queues allocated will be the lesser of the CPUs requested by the guest and the max value defined. By default, this value is set to none meaning the legacy limits based on the reported kernel major version will be used.

mem_stats_period_seconds = 10

integer value

A number of seconds to memory usage statistics period. Zero or negative value mean to disable memory usage statistics.

nfs_mount_options = None

string value

Mount options passed to the NFS client. See section of the nfs man page for details.

Mount options controls the way the filesystem is mounted and how the NFS client behaves when accessing files on this mount point.

Possible values:

nfs_mount_point_base = $state_path/mnt

string value

Directory where the NFS volume is mounted on the compute node. The default is mnt directory of the location where nova’s Python module is installed.

NFS provides shared storage for the OpenStack Block Storage service.

Possible values:

num_aoe_discover_tries = 3

integer value

Number of times to rediscover AoE target to find volume.

Nova provides support for block storage attaching to hosts via AOE (ATA over Ethernet). This option allows the user to specify the maximum number of retry attempts that can be made to discover the AoE device.

num_iser_scan_tries = 5

integer value

Number of times to scan iSER target to find volume.

iSER is a server network protocol that extends iSCSI protocol to use Remote Direct Memory Access (RDMA). This option allows the user to specify the maximum number of scan attempts that can be made to find iSER volume.

num_memory_encrypted_guests = None

integer value

Maximum number of guests with encrypted memory which can run concurrently on this compute host.

For now this is only relevant for AMD machines which support SEV (Secure Encrypted Virtualization). Such machines have a limited number of slots in their memory controller for storing encryption keys. Each running guest with encrypted memory will consume one of these slots.

The option may be reused for other equivalent technologies in the future. If the machine does not support memory encryption, the option will be ignored and inventory will be set to 0.

If the machine does support memory encryption, for now a value of None means an effectively unlimited inventory, i.e. no limit will be imposed by Nova on the number of SEV guests which can be launched, even though the underlying hardware will enforce its own limit. However it is expected that in the future, auto-detection of the inventory from the hardware will become possible, at which point None will cause auto-detection to automatically impose the correct limit.

Related options:

num_nvme_discover_tries = 5

integer value

Number of times to rediscover NVMe target to find volume

Nova provides support for block storage attaching to hosts via NVMe (Non-Volatile Memory Express). This option allows the user to specify the maximum number of retry attempts that can be made to discover the NVMe device.

num_pcie_ports = 0

integer value

The number of PCIe ports an instance will get.

Libvirt allows a custom number of PCIe ports (pcie-root-port controllers) a target instance will get. Some will be used by default, rest will be available for hotplug use.

By default we have just 1-2 free ports which limits hotplug.

More info: https://github.com/qemu/qemu/blob/master/docs/pcie.txt

Due to QEMU limitations for aarch64/virt maximum value is set to 28.

Default value 0 moves calculating amount of ports to libvirt.

num_volume_scan_tries = 5

integer value

Number of times to scan given storage protocol to find volume.

pmem_namespaces = []

list value

Configure persistent memory(pmem) namespaces.

These namespaces must have been already created on the host. This config option is in the following format::

"$LABEL:$NSNAME[|$NSNAME][,$LABEL:$NSNAME[|$NSNAME]]"

quobyte_client_cfg = None

string value

Path to a Quobyte Client configuration file.

quobyte_mount_point_base = $state_path/mnt

string value

Directory where the Quobyte volume is mounted on the compute node.

Nova supports Quobyte volume driver that enables storing Block Storage service volumes on a Quobyte storage back end. This Option specifies the path of the directory where Quobyte volume is mounted.

Possible values:

rbd_connect_timeout = 5

integer value

The RADOS client timeout in seconds when initially connecting to the cluster.

rbd_destroy_volume_retries = 12

integer value

Number of retries to destroy a RBD volume.

Related options:

rbd_destroy_volume_retry_interval = 5

integer value

Number of seconds to wait between each consecutive retry to destroy a RBD volume.

Related options:

rbd_secret_uuid = None

string value

The libvirt UUID of the secret for the rbd_user volumes.

rbd_user = None

string value

The RADOS client name for accessing rbd(RADOS Block Devices) volumes.

Libvirt will refer to this user when connecting and authenticating with the Ceph RBD server.

realtime_scheduler_priority = 1

integer value

In a realtime host context vCPUs for guest will run in that scheduling priority. Priority depends on the host kernel (usually 1-99)

remote_filesystem_transport = ssh

string value

libvirt’s transport method for remote file operations.

Because libvirt cannot use RPC to copy files over network to/from other compute nodes, other method must be used for:

rescue_image_id = None

string value

The ID of the image to boot from to rescue data from a corrupted instance.

If the rescue REST API operation doesn’t provide an ID of an image to use, the image which is referenced by this ID is used. If this option is not set, the image from the instance is used.

Possible values:

Related options:

rescue_kernel_id = None

string value

The ID of the kernel (AKI) image to use with the rescue image.

If the chosen rescue image allows the separate definition of its kernel disk, the value of this option is used, if specified. This is the case when Amazon's AMI/AKI/ARI image format is used for the rescue image.

Possible values:

Related options:

rescue_ramdisk_id = None

string value

The ID of the RAM disk (ARI) image to use with the rescue image.

If the chosen rescue image allows the separate definition of its RAM disk, the value of this option is used, if specified. This is the case when Amazon's AMI/AKI/ARI image format is used for the rescue image.

Possible values:

Related options:

rng_dev_path = /dev/urandom

string value

The path to an RNG (Random Number Generator) device that will be used as the source of entropy on the host. Since libvirt 1.3.4, any path (that returns random numbers when read) is accepted. The recommended source of entropy is /dev/urandom — it is non-blocking, therefore relatively fast; and avoids the limitations of /dev/random, which is a legacy interface. For more details (and comparison between different RNG sources), refer to the "Usage" section in the Linux kernel API documentation for [u]random: http://man7.org/linux/man-pages/man4/urandom.4.html and http://man7.org/linux/man-pages/man7/random.7.html.

rx_queue_size = None

integer value

Configure virtio rx queue size.

This option is only usable for virtio-net device with vhost and vhost-user backend. Available only with QEMU/KVM. Requires libvirt v2.3 QEMU v2.7.

`smbfs_mount_options = `

string value

Mount options passed to the SMBFS client.

Provide SMBFS options as a single string containing all parameters. See mount.cifs man page for details. Note that the libvirt-qemu uid and gid must be specified.

smbfs_mount_point_base = $state_path/mnt

string value

Directory where the SMBFS shares are mounted on the compute node.

snapshot_compression = False

boolean value

Enable snapshot compression for qcow2 images.

Note: you can set snapshot_image_format to qcow2 to force all snapshots to be in qcow2 format, independently from their original image type.

Related options:

snapshot_image_format = None

string value

Determine the snapshot image format when sending to the image service.

If set, this decides what format is used when sending the snapshot to the image service. If not set, defaults to same type as source image.

snapshots_directory = $instances_path/snapshots

string value

Location where libvirt driver will store snapshots before uploading them to image service

sparse_logical_volumes = False

boolean value

Create sparse logical volumes (with virtualsize) if this flag is set to True.

Deprecated since: 18.0.0

Reason: Sparse logical volumes is a feature that is not tested hence not supported. LVM logical volumes are preallocated by default. If you want thin provisioning, use Cinder thin-provisioned volumes.

swtpm_enabled = False

boolean value

Enable emulated TPM (Trusted Platform Module) in guests.

swtpm_group = tss

string value

Group that swtpm binary runs as.

When using emulated TPM, the swtpm binary will run to emulate a TPM device. The user this binary runs as depends on libvirt configuration, with tss being the default.

In order to support cold migration and resize, nova needs to know what group the swtpm binary is running as in order to ensure that files get the proper ownership after being moved between nodes.

Related options:

swtpm_user = tss

string value

User that swtpm binary runs as.

When using emulated TPM, the swtpm binary will run to emulate a TPM device. The user this binary runs as depends on libvirt configuration, with tss being the default.

In order to support cold migration and resize, nova needs to know what user the swtpm binary is running as in order to ensure that files get the proper ownership after being moved between nodes.

Related options:

sysinfo_serial = unique

string value

The data source used to the populate the host "serial" UUID exposed to guest in the virtual BIOS. All choices except unique will change the serial when migrating the instance to another host. Changing the choice of this option will also affect existing instances on this host once they are stopped and started again. It is recommended to use the default choice (unique) since that will not change when an instance is migrated. However, if you have a need for per-host serials in addition to per-instance serial numbers, then consider restricting flavors via host aggregates.

tx_queue_size = None

integer value

Configure virtio tx queue size.

This option is only usable for virtio-net device with vhost-user backend. Available only with QEMU/KVM. Requires libvirt v3.7 QEMU v2.10.

uid_maps = []

list value

List of uid targets and ranges.Syntax is guest-uid:host-uid:count. Maximum of 5 allowed.

use_virtio_for_bridges = True

boolean value

Use virtio for bridge interfaces with KVM/QEMU