nova compute #20

Supports: xenial bionic eoan focal trusty groovy
Add to new model


OpenStack Compute, codenamed Nova, is a cloud computing fabric controller. In addition to its "native" API (the OpenStack API), it also supports the Amazon EC2 API. . This charm provides the Nova Compute hypervisor service and should be deployed directly to physical servers.


This charm provides Nova Compute, the OpenStack compute service. Its target platform is Ubuntu (preferably LTS) + OpenStack.


The following interfaces are provided:

  • cloud-compute - Used to relate (at least) with one or more of nova-cloud-controller, glance, ceph, cinder, mysql, ceilometer-agent, rabbitmq-server, neutron

  • nrpe-external-master - Used to generate Nagios checks.


Nova compute only requires database access if using nova-network. If using Neutron, no direct database access is required and the shared-db relation need not be added. The nova-network feature is not available in Ussuri and later, and so this interface produces a warning if added.


This charm support nova-network (legacy) and Neutron networking.


This charm supports a number of different storage backends depending on your hypervisor type and storage relations.

In order to have cinder ceph rbd support for OpenStack Ocata and newer releases, ceph-access relation must be added to cinder-ceph to allow nova-compute units to communicate with multiple ceph backends using different cephx keys and user names.

$ juju add-relation nova-compute cinder-ceph

See LP Bug #1671422 for more information.

Availability Zones

There are two options to provide default_availability_zone config for nova nodes:

  • default-availability-zone
  • customize-failure-domain

The order of precedence is as follows:

  1. Information from a Juju provider (JUJU_AVAILABILITY_ZONE) if customize-failure-domain is set to True and Juju has set the JUJU_AVAILABILITY_ZONE to a non-empty value;
  2. The value of default-availability-zone will be used if customize-failure-domain is set to True but no JUJU_AVAILABILITY_ZONE is provided via hook context by the Juju provider;
  3. Otherwise, the value of default-availability-zone charm option will be used.

The default_availability_zone in Nova affects scheduling if a given Nova node was not placed into an aggregate with an availability zone present as a property by an operator. Using customize-failure-domain is recommended as it provides AZ-aware scheduling out of the box if an operator specifies an AZ during instance creation.

These options also affect the AZ propagated down to networking subordinates which is useful for AZ-aware Neutron agent scheduling.

NFV support

This charm (in conjunction with the nova-cloud-controller and neutron-api charms) supports use of nova-compute nodes configured for use in Telco NFV deployments; specifically the following configuration options (yaml excerpt):

  hugepages: 60%
  vcpu-pin-set: "^0,^2"
  reserved-host-memory: 1024
  pci-passthrough-whitelist: {"vendor_id":"1137","product_id":"0071","address":"*:0a:00.*","physical_network":"physnet1"}

In this example, compute nodes will be configured with 60% of available RAM for hugepage use (decreasing memory fragmentation in virtual machines, improving performance), and Nova will be configured to reserve CPU cores 0 and 2 and 1024M of RAM for host usage and use the supplied PCI device whitelist as PCI devices that as consumable by virtual machines, including any mapping to underlying provider network names (used for SR-IOV VF/PF port scheduling with Nova and Neutron's SR-IOV support).

The vcpu-pin-set configuration option is a comma-separated list of physical CPU numbers that virtual CPUs can be allocated to by default. Each element should be either a single CPU number, a range of CPU numbers, or a caret followed by a CPU number to be excluded from a previous range. For example:

vcpu-pin-set: "4-12,^8,15"

The pci-passthrough-whitelist configuration must be specified as follows:

A JSON dictionary which describe a whitelisted PCI device. It should take the following format:

["device_id": "<id>",] ["product_id": "<id>",]
["address": "[[[[<domain>]:]<bus>]:][<slot>][.[<function>]]" |
"devname": "PCI Device Name",]
{"tag": "<tag_value>",}

where '[' indicates zero or one occurrences, '{' indicates zero or multiple occurrences, and '|' mutually exclusive options. Note that any missing fields are automatically wildcarded. Valid examples are:

pci-passthrough-whitelist: {"devname":"eth0", "physical_network":"physnet"}
pci-passthrough-whitelist: {"address":"*:0a:00.*"}
pci-passthrough-whitelist: {"address":":0a:00.", "physical_network":"physnet1"}
pci-passthrough-whitelist: {"vendor_id":"1137", "product_id":"0071"}
pci-passthrough-whitelist: {"vendor_id":"1137", "product_id":"0071", "address": "0000:0a:00.1", "physical_network":"physnet1"}

The following is invalid, as it specifies mutually exclusive options:

pci-passthrough-whitelist: {"devname":"eth0", "physical_network":"physnet", "address":"*:0a:00.*"}

A JSON list of JSON dictionaries corresponding to the above format. For example:

pci-passthrough-whitelist: [{"product_id":"0001", "vendor_id":"8086"}, {"product_id":"0002", "vendor_id":"8086"}]

The OpenStack advanced networking documentation provides further details on whitelist configuration and how to create instances with Neutron ports wired to SR-IOV devices.

Network Spaces

This charm supports use of network spaces by binding relation endpoints.

In addition this charm declares two extra-bindings:

  • internal: this binding is used to determine the network space to use for console access to instances.

  • migration: this binding is used to determine which network space should be used for live and cold migrations between hypervisors.

Note that the nova-cloud-controller application must have bindings to the same network spaces used for both 'internal' and 'migration' extra bindings.

See the Juju documentation for more information on using network spaces.


(string) Experimental enable apparmor profile. Valid settings: 'complain', 'enforce' or 'disable'. . AA disabled by default.
(boolean) If True enables openstack upgrades for this charm via juju actions. You will still need to set openstack-origin to the new repository but instead of an upgrade running automatically across all units, it will wait for you to execute the openstack-upgrade action for this charm on each unit. If False it will revert to existing behavior of upgrading all units on config change.
(string) Only used when migration-auth-type is set to ssh. . Full path to authorized_keys file, can be useful for systems with non-default AuthorizedKeysFile location. It will be formatted using the following variables: . homedir - user's home directory username - username .
(string) Bridge interface to be configured.
(string) IP to be assigned to bridge interface.
(string) Netmask to be assigned to bridge interface.
(int) This value dictates the number of replicas ceph must make of any object it stores within the nova rbd pool. Of course, this only applies if using Ceph as a backend store. Note that once the nova rbd pool has been created, changing this value will not have any effect (although it can be changed in ceph by manually configuring your ceph cluster).
(int) Defines a relative weighting of the pool as a percentage of the total amount of data in the Ceph cluster. This effectively weights the number of placement groups for the pool created to be appropriately portioned to the amount of data expected. For example, if the ephemeral volumes for the OpenStack compute instances are expected to take up 20% of the overall configuration then this value would be specified as 20. Note - it is important to choose an appropriate value for the pool weight as this directly affects the number of placement groups which will be created for the pool. The number of placement groups for a pool can only be increased, never decreased - so it is important to identify the percent of data that will likely reside in the pool.
(string) Comma-separated list of key=value config flags. These values will be placed in the nova.conf [DEFAULT] section.
(string) Sets compute/cpu_dedicated_set option in nova.conf defines which physical CPUs will be used for dedicated guest vCPU resources. This option can't be used with vcpu-pin-set.
(string) Set to 'host-model' to clone the host CPU feature flags; to 'host-passthrough' to use the host CPU model exactly; to 'custom' to use a named CPU model; to 'none' to not set any CPU model. If virt_type='kvm|qemu', it will default to 'host-model', otherwise it will default to 'none'. Defaults to 'host-passthrough' for ppc64el, ppc64le if no value is set.
(string) Set to a named libvirt CPU model (see names listed in /usr/share/libvirt/cpu_map.xml). Only has effect if cpu_mode='custom' and virt_type='kvm|qemu'.
(string) Space delimited list of specific CPU flags for libvirt.
(string) Sets compute/cpu_shared_set option in nova.conf defines which physical CPUs will be used for best-effort guest vCPU resources. Currently only used by libvirt driver to place guest emulator threads when hw:emulator_threads_policy:share is set.
(boolean) Juju propagates availability zone information to charms from the underlying machine provider such as MAAS and this option allows the charm to use JUJU_AVAILABILITY_ZONE to set default_availability_zone for Nova nodes. This option overrides the default-availability-zone charm config setting only when the Juju provider sets JUJU_AVAILABILITY_ZONE.
(string) Nova database name.
(string) Username for database access.
(boolean) Enable debug logging.
(string) Default compute node availability zone. . This option determines the availability zone to be used when it is not specified in the VM creation request. If this option is not set, the default availability zone 'nova' is used. If customize-failure-domain is set to True, it will override this option only if an AZ is set by the Juju provider. If JUJU_AVAILABILITY_ZONE is not set, the value specified by this option will be used regardless of customize-failure-domain's setting. . NOTE: Availability zones must be created manually using the 'openstack aggregate create' command. .
(string) The default format an ephemeral volume will be formatted with on creation. Possible values: ext2 ext3 ext4 xfs ntfs (only for Windows guests)
(string) Specific cachemodes to use for different disk types e.g: file=directsync,block=none
(string) (lrc plugin) The type of the crush bucket in which each set of chunks defined by l will be stored. For instance, if it is set to rack, each group of l chunks will be placed in a different rack. It is used to create a CRUSH rule step such as step choose rack. If it is not set, no such grouping is done.
(string) Device class from CRUSH map to use for placement groups for erasure profile - valid values: ssd, hdd or nvme (or leave unset to not use a device class).
(int) (shec plugin - c) The number of parity chunks each of which includes each data chunk in its calculation range. The number is used as a durability estimator. For instance, if c=2, 2 OSDs can be down without losing data.
(int) (clay plugin - d) Number of OSDs requested to send data during recovery of a single chunk. d needs to be chosen such that k+1 <= d <= k+m-1. Larger the d, the better the savings.
(int) Number of data chunks that will be used for EC data pool. K+M factors should never be greater than the number of available zones (or hosts) for balancing.
(int) (lrc plugin - l) Group the coding and data chunks into sets of size l. For instance, for k=4 and m=2, when l=3 two groups of three are created. Each set can be recovered without reading chunks from another set. Note that using the lrc plugin does incur more raw storage usage than isa or jerasure in order to reduce the cost of recovery operations.
(int) Number of coding chunks that will be used for EC data pool. K+M factors should never be greater than the number of available zones (or hosts) for balancing.
(string) Name for the EC profile to be created for the EC pools. If not defined a profile name will be generated based on the name of the pool used by the application.
(string) EC plugin to use for this applications pool. The following list of plugins acceptable - jerasure, lrc, isa, shec, clay.
(string) (clay plugin) specifies the plugin that is used as a building block in the layered construction. It can be one of jerasure, isa, shec (defaults to jerasure).
(string) EC profile technique used for this applications pool - will be validated based on the plugin configured via ec-profile-plugin. Supported techniques are ‘reed_sol_van’, ‘reed_sol_r6_op’, ‘cauchy_orig’, ‘cauchy_good’, ‘liber8tion’ for jerasure, ‘reed_sol_van’, ‘cauchy’ for isa and ‘single’, ‘multiple’ for shec.
(string) Name of the metadata pool to be created (for RBD use-cases). If not defined a metadata pool name will be generated based on the name of the data pool used by the application. The metadata pool is always replicated, not erasure coded.
(boolean) Configure libvirt or lxd for live migration. . Live migration support for lxd is still considered experimental. . NOTE: This also enables passwordless SSH access for user 'root' between compute hosts.
(boolean) Enable instance resizing. . NOTE: This also enables passwordless SSH access for user 'nova' between compute hosts.
(boolean) Encrypt block devices used for nova instances using dm-crypt, making use of vault for encryption key management; requires a relation to vault.
(string) Block devices to use for storage of ephermeral disks to support nova instances; generally used in-conjunction with 'encrypt' to support data-at-rest encryption of instance direct attached storage volumes.
(string) Cloud instances provide ephermeral storage which is normally mounted on /mnt. . Setting this option to the path of the ephemeral mountpoint will force an unmount of the corresponding device so that it can be used for as the backing store for local instances. This is useful for testing purposes (cloud deployment is not a typical use case).
(string) Network interface on which to build bridge.
(boolean) Force conversion of backing images to raw format. Note that the conversion process in Pike uses O_DIRECT calls - certain file systems do not support this, for example ZFS; e.g. if using the LXD provider with ZFS backend, this option should be set to False.
(string) Apply system hardening. Supports a space-delimited list of modules to run. Supported modules currently include os, ssh, apache and mysql.
(string) The percentage of system memory to use for hugepages eg '10%' or the total number of 2M hugepages - eg "1024". For a systemd system (wily and later) the prefered approach is to enable hugepages via kernel parameters set in MAAS and systemd will mount them automatically. . NOTE: For hugepages to work it must be enabled on the machine deployed to. This can be accomplished by setting kernel parameters on capable machines in MAAS, tagging them and using these tags as constraints in the model.
(string) Path used for storing Nova instances data - empty means default of /var/lib/nova/instances.
(string) Set to 1 to enable KSM, 0 to disable KSM, and AUTO to use default settings. . Please note that the AUTO value works for qemu 2.2+ (> Kilo), older releases will be set to 1 as default.
(string) Tell Nova which libvirt image backend to use. Supported backends are raw, qcow2, rbd and flat. If no backend is specified, the Nova default (qcow2) is used. NOTE: "rbd" imagebackend is only supported with >= Juno. NOTE: "flat" imagebackend is only supported with >= Newton and replaces "raw".
(string) Specify a network in cidr notation (, which directs libvirt to use a specific network address as the live_migration_inbound_addr to make use of a dedicated migration network if possible. . Please note that if the migration binding has been declared and set, the primary address for that space has precedence over this configuration option. . This option doesn't have any effect on clouds running a release < Ocata.
(int) 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.
(int) Maximum permitted downtime, in milliseconds, for live migration switchover. Will be rounded up to a minimum of 100ms. Use a large value if guest liveness is unimportant.
(int) Time to wait, in seconds, between each step increase of the migration downtime. Minimum delay is 10 seconds. Value is per GiB of guest RAM + disk to be transferred, with lower bound of a minimum of 2 GiB per device.
(int) Number of incremental steps to reach max downtime value. Will be rounded up to a minimum of 3 steps.
(boolean) If live-migration is enabled, this option allows Nova to throttle down CPU when an on-going live migration is slow.
(boolean) If live-migration is enabled, this option allows Nova to switch an on- going live migration to post-copy mode.
(string) TCP authentication scheme for libvirt live migration. Available options include ssh.
(string) Whether to run nova-api and nova-network on the compute nodes. Note that nova-network is not available on Ussuri and later.
(string) Used by the nrpe-external-master subordinate charm. A string that will be prepended to instance name to set the host name in nagios. So for instance the hostname would be something like: . juju-myservice-0 . If you're running multiple environments with the same services in them this allows you to differentiate between them.
(string) A comma-separated list of nagios servicegroups. If left empty, the nagios_context will be used as the servicegroup.
(string) There are two types of notifications in Nova: legacy notifications which have an unversioned payload and newer notifications which have a versioned payload. . Setting this option to `versioned` will use the versioned notification concept, `unversioned`, the unversioned notification concept and finally `both` will use the two concepts. . Starting in the Pike release, the notification_format includes both the versioned and unversioned message notifications. Ceilometer does not yet consume the versioned message notifications, so intentionally make the default notification format unversioned until this is implemented. . Possible Values are both, versioned, unversioned.
(string) Full path to Nova configuration file.
(string) Repository from which to install. May be one of the following: distro (default), ppa:somecustom/ppa, a deb url sources entry or a supported Ubuntu Cloud Archive (UCA) release pocket. . Supported UCA sources include: . cloud:<series>-<openstack-release> cloud:<series>-<openstack-release>/updates cloud:<series>-<openstack-release>/staging cloud:<series>-<openstack-release>/proposed . For series=Precise we support UCA for openstack-release= * icehouse . For series=Trusty we support UCA for openstack-release= * juno * kilo * ... . NOTE: updating this setting to a source that is known to provide a later version of OpenStack will trigger a software upgrade. .
(string) The IP address and netmask of the OpenStack Internal network (e.g. . This network will be used to bind vncproxy client.
(string) The pci-passthrough-whitelist option of nova-compute charm is used for specifying which PCI devices are allowed passthrough. pci-alias is more a convenience that can be used in conjunction with Nova flavor properties to automatically assign required PCI devices to new instances. You could, for example, have a GPU flavor or a SR-IOV flavor: . pci-alias='{"vendor_id":"8086","product_id":"10ca","name":"a1"}' . This configures a new PCI alias 'a1' which will request a PCI device with a vendor id of 0x8086 and a product id of 10ca. . For more information about the syntax of pci_alias, refer to
(string) Sets the pci_passthrough_whitelist option in nova.conf which allows PCI passthrough of specific devices to VMs. . Example applications: GPU processing, SR-IOV networking, etc. . NOTE: For PCI passthrough to work IOMMU must be enabled on the machine deployed to. This can be accomplished by setting kernel parameters on capable machines in MAAS, tagging them and using these tags as constraints in the model.
(string) Ceph pool type to use for storage - valid values include ‘replicated’ and ‘erasure-coded’.
(boolean) If True enables IPv6 support. The charm will expect network interfaces to be configured with an IPv6 address. If set to False (default) IPv4 is expected. . NOTE: these charms do not currently support IPv6 privacy extension. In order for this charm to function correctly, the privacy extension must be disabled and a non-temporary address must be configured/available on your network interface.
(string) Username used to access rabbitmq queue.
(string) Rabbitmq vhost.
(string) Enable/disable rbd client cache. Leaving this value unset will result in default Ceph rbd client settings being used (rbd cache is enabled by default for Ceph >= Giant). Supported values here are "enabled" or "disabled".
(string) RBD pool to use with Nova libvirt RBDImageBackend. Only required when you have libvirt-image-backend set to 'rbd'.
(int) Amount of memory in MB to reserve for the host. Defaults to 512MB.
(string) Sets a reserved amont of huge pages per NUMA nodes which are used by third-party components. Semicolons are used as separator. . reserved_huge_pages = node:0,size:2048,count:64;node:1,size:1GB,count:1 . The above will consider 64 pages of 2MiB on NUMA node 0 and 1 page of 1GiB on NUMA node 1 reserved. They will not be used by Nova to map guests memory.
(boolean) Optionally restrict Ceph key permissions to access pools as required.
(boolean) This option determines whether to start guests that were running before the host rebooted.
(boolean) Ensure notifications are included in the log files. It will set an additional log driver for oslo messaging notifications.
(string) YAML formatted associative array of sysctl values, e.g.: '{ kernel.pid_max : 4194303 }'
{ net.ipv4.neigh.default.gc_thresh1 : 128, net.ipv4.neigh.default.gc_thresh2 : 28672, net.ipv4.neigh.default.gc_thresh3 : 32768, net.ipv6.neigh.default.gc_thresh1 : 128, net.ipv6.neigh.default.gc_thresh2 : 28672, net.ipv6.neigh.default.gc_thresh3 : 32768, net.nf_conntrack_max : 1000000, net.netfilter.nf_conntrack_buckets : 204800, net.netfilter.nf_conntrack_max : 1000000 }
(boolean) Openstack mostly defaults to using public endpoints for internal communication between services. If set to True this option will configure services to use internal endpoints where possible.
(boolean) Use a multipath connection for iSCSI or FC volumes. Enabling this feature causes libvirt to login, discover and scan available targets before presenting the disk via device mapper (/dev/mapper/XX) to the VM instead of a single path (/dev/disk/by-path/XX). If changed after deployment, each VM will require a full stop/start for changes to take affect.
(boolean) Setting this to True will allow supporting services to log to syslog.
(string) Sets vcpu_pin_set option in nova.conf which defines which pcpus that instance vcpus can or cannot use. For example '^0,^2' to reserve two cpus for the host.
(boolean) Enable verbose logging.
(string) Virtualisation flavor. Supported and tested flavors are: kvm, lxd. If using the lxd flavor, the lxd subordinate charm must also be related. Other native libvirt flavors available to exercise (dev/test only): uml, lxc, qemu. NOTE: Changing virtualisation flavor after deployment is not supported.
(int) Sets libvirt/rx_queue_size option in nova.conf. Larger queues sizes for virtio-net devices increases networking performance by amortizing vCPU preemption and avoiding packet drops. Only works with Rocky and later, since QEMU 2.7.0 and libvirt 2.3.0. Default value 256. Authorized values [256, 512, 1024].
(int) Sets libvirt/tx_queue_size option in nova.conf. Larger queues sizes for virtio-net devices increases networking performance by amortizing vCPU preemption and avoiding packet drops. Only works with Rocky and later, since QEMU 2.10.0 and libvirt 3.7.0. Default value 256. Authorized values [256, 512, 1024].
(float) The CPU core multiplier to use when configuring worker processes for this services e.g. metadata-api. By default, the number of workers for each daemon is set to twice the number of CPU cores a service unit has. When deployed in a LXD container, this default value will be capped to 4 workers unless this configuration option is set.