Skip to content
Docs
Reference
Instance options

Instance options

Instance options are configuration options that are directly related to the instance.

See Configure instance options for instructions on how to set the instance options.

The key/value configuration is namespaced. The following options are available:

Note that while a type is defined for each option, all values are stored as strings and should be exported over the REST API as strings (which makes it possible to support any extra values without breaking backward compatibility).

Miscellaneous options

In addition to the configuration options listed in the following sections, these instance options are supported:

agent.nic_config

Whether to use the name and MTU of the default network interfaces

  • Key: agent.nic_config
  • Type: bool
  • Default: false
  • Live update: no
  • Condition: virtual machine

When set to true, the name and MTU of the default network interfaces inside the virtual machine will match those of the instance devices.

cluster.evacuate

What to do when evacuating the instance

  • Key: cluster.evacuate
  • Type: string
  • Default: auto
  • Live update: no

The cluster.evacuate provides control over how instances are handled when a cluster member is being evacuated.

Available Modes:

  • auto (default): The system will automatically decide the best evacuation method based on the instance’s type and configured devices:
    • If any device is not suitable for migration, the instance will not be migrated (only stopped).
    • Live migration will be used only for virtual machines with the migration.stateful setting enabled and for which all its devices can be migrated as well.
  • live-migrate: Instances are live-migrated to another node. This means the instance remains running and operational during the migration process, ensuring minimal disruption.
  • migrate: In this mode, instances are migrated to another node in the cluster. The migration process will not be live, meaning there will be a brief downtime for the instance during the migration.
  • stop: Instances are not migrated. Instead, they are stopped on the current node.

See Evacuate a cluster member for more information.

linux.kernel_modules

Kernel modules to load or allow loading

  • Key: linux.kernel_modules
  • Type: string
  • Live update: yes
  • Condition: container

Specify the kernel modules as a comma-separated list.

The modules are loaded before the instance starts, or they can be loaded by a privileged user if linux.kernel_modules.load is set to ondemand.

linux.kernel_modules.load

How to load kernel modules

  • Key: linux.kernel_modules.load
  • Type: string
  • Default: boot
  • Live update: no
  • Condition: container

This option specifies how to load the kernel modules that are specified in linux.kernel_modules. Possible values are boot (load the modules when booting the container) and ondemand (intercept the finit_modules() syscall and allow a privileged user in the container’s user namespace to load the modules).

linux.sysctl.*

Override for the corresponding sysctl setting in the container

  • Key: linux.sysctl.*
  • Type: string
  • Live update: no
  • Condition: container

ubuntu_pro.guest_attach

Whether to auto-attach Ubuntu Pro.

  • Key: ubuntu_pro.guest_attach
  • Type: string
  • Live update: no

Indicate whether the guest should auto-attach Ubuntu Pro at start up. The allowed values are off, on, and available. If set to off, it will not be possible for the Ubuntu Pro client in the guest to obtain guest token via devlxd. If set to available, attachment via guest token is possible but will not be performed automatically by the Ubuntu Pro client in the guest at startup. If set to on, attachment will be performed automatically by the Ubuntu Pro client in the guest at startup. To allow guest attachment, the host must be an Ubuntu machine that is Pro attached, and guest attachment must be enabled via the Pro client. To do this, run pro config set lxd_guest_attach=on.

user.*

Free-form user key/value storage

  • Key: user.*
  • Type: string
  • Live update: no

User keys can be used in search.

environment.*

Environment variables for the instance

  • Key: environment.*
  • Type: string
  • Live update: yes (exec)

You can export key/value environment variables to the instance. These are then set for lxc exec.

Boot-related options

The following instance options control the boot-related behavior of the instance:

boot.autostart

Whether to always start the instance when LXD starts

  • Key: boot.autostart
  • Type: bool
  • Live update: no

If set to false, restore the last state.

boot.autostart.delay

Delay after starting the instance

  • Key: boot.autostart.delay
  • Type: integer
  • Default: 0
  • Live update: no

The number of seconds to wait after the instance started before starting the next one.

boot.autostart.priority

What order to start the instances in

  • Key: boot.autostart.priority
  • Type: integer
  • Default: 0
  • Live update: no

The instance with the highest value is started first.

boot.debug_edk2

Enable debug version of the edk2

  • Key: boot.debug_edk2
  • Type: bool

The instance should use a debug version of the edk2. A log file can be found in $LXD_DIR/logs/<instance_name>/edk2.log.

boot.host_shutdown_timeout

How long to wait for the instance to shut down

  • Key: boot.host_shutdown_timeout
  • Type: integer
  • Default: 30
  • Live update: yes

Number of seconds to wait for the instance to shut down before it is force-stopped.

boot.stop.priority

What order to shut down the instances in

  • Key: boot.stop.priority
  • Type: integer
  • Default: 0
  • Live update: no

The instance with the highest value is shut down first.

cloud-init configuration

The following instance options control the cloud-init configuration of the instance:

cloud-init.network-config

Network configuration for cloud-init

  • Key: cloud-init.network-config
  • Type: string
  • Default: DHCP on eth0
  • Live update: no
  • Condition: If supported by image

The content is used as seed value for cloud-init.

cloud-init.ssh-keys.KEYNAME

Additional SSH key to be injected on the instance by cloud-init

  • Key: cloud-init.ssh-keys.KEYNAME
  • Type: string
  • Live update: no
  • Condition: If supported by image

Represents an additional SSH public key to be merged into existing cloud-init seed data and injected into an instance. Has the format {user}:{key}, where {user} is a Linux username and {key} can be either a pure SSH public key or an import ID for a key hosted elsewhere. // For example: root:gh:githubUser, myUser:ssh-keyAlg publicKeyHash

cloud-init.user-data

User data for cloud-init

  • Key: cloud-init.user-data
  • Type: string
  • Default: #cloud-config
  • Live update: no
  • Condition: If supported by image

The content is used as seed value for cloud-init.

cloud-init.vendor-data

Vendor data for cloud-init

  • Key: cloud-init.vendor-data
  • Type: string
  • Default: #cloud-config
  • Live update: no
  • Condition: If supported by image

The content is used as seed value for cloud-init.

user.network-config

Legacy version of cloud-init.network-config

  • Key: user.network-config
  • Type: string
  • Default: DHCP on eth0
  • Live update: no
  • Condition: If supported by image

user.user-data

Legacy version of cloud-init.user-data

  • Key: user.user-data
  • Type: string
  • Default: #cloud-config
  • Live update: no
  • Condition: If supported by image

user.vendor-data

Legacy version of cloud-init.vendor-data

  • Key: user.vendor-data
  • Type: string
  • Default: #cloud-config
  • Live update: no
  • Condition: If supported by image

Support for these options depends on the image that is used and is not guaranteed.

If you specify both cloud-init.user-data and cloud-init.vendor-data, the content of both options is merged. Therefore, make sure that the cloud-init configuration you specify in those options does not contain the same keys.

Resource limits

The following instance options specify resource limits for the instance:

limits.cpu

Which CPUs to expose to the instance

  • Key: limits.cpu
  • Type: string
  • Default: 1 (VMs)
  • Live update: yes

A number or a specific range of CPUs to expose to the instance.

See CPU pinning for more information.

limits.cpu.allowance

How much of the CPU can be used

  • Key: limits.cpu.allowance
  • Type: string
  • Default: 100%
  • Live update: yes
  • Condition: container

To control how much of the CPU can be used, specify either a percentage (50%) for a soft limit or a chunk of time (25ms/100ms) for a hard limit.

See Allowance and priority (container only) for more information.

limits.cpu.nodes

Which NUMA nodes to place the instance CPUs on

  • Key: limits.cpu.nodes
  • Type: string
  • Live update: yes

A comma-separated list of NUMA node IDs or ranges to place the instance CPUs on.

See Allowance and priority (container only) for more information.

limits.cpu.pin_strategy

VM CPU auto pinning strategy

  • Key: limits.cpu.pin_strategy
  • Type: string
  • Default: none
  • Live update: no
  • Condition: virtual machine

Specify the strategy for VM CPU auto pinning. Possible values: none (disables CPU auto pinning) and auto (enables CPU auto pinning).

See CPU limits for virtual machines for more information.

limits.cpu.priority

CPU scheduling priority compared to other instances

  • Key: limits.cpu.priority
  • Type: integer
  • Default: 10 (maximum)
  • Live update: yes
  • Condition: container

When overcommitting resources, specify the CPU scheduling priority compared to other instances that share the same CPUs. Specify an integer between 0 and 10.

See Allowance and priority (container only) for more information.

limits.disk.priority

Priority of the instance’s I/O requests

  • Key: limits.disk.priority
  • Type: integer
  • Default: 5 (medium)
  • Live update: yes

Controls how much priority to give to the instance’s I/O requests when under load.

Specify an integer between 0 and 10.

limits.hugepages.1GB

Limit for the number of 1 GB huge pages

  • Key: limits.hugepages.1GB
  • Type: string
  • Live update: yes
  • Condition: container

Fixed value (in bytes) to limit the number of 1 GB huge pages. Various suffixes are supported (see Units for storage and network limits).

See Huge page limits for more information.

limits.hugepages.1MB

Limit for the number of 1 MB huge pages

  • Key: limits.hugepages.1MB
  • Type: string
  • Live update: yes
  • Condition: container

Fixed value (in bytes) to limit the number of 1 MB huge pages. Various suffixes are supported (see Units for storage and network limits).

See Huge page limits for more information.

limits.hugepages.2MB

Limit for the number of 2 MB huge pages

  • Key: limits.hugepages.2MB
  • Type: string
  • Live update: yes
  • Condition: container

Fixed value (in bytes) to limit the number of 2 MB huge pages. Various suffixes are supported (see Units for storage and network limits).

See Huge page limits for more information.

limits.hugepages.64KB

Limit for the number of 64 KB huge pages

  • Key: limits.hugepages.64KB
  • Type: string
  • Live update: yes
  • Condition: container

Fixed value (in bytes) to limit the number of 64 KB huge pages. Various suffixes are supported (see Units for storage and network limits).

See Huge page limits for more information.

limits.memory

Usage limit for the host’s memory

  • Key: limits.memory
  • Type: string
  • Default: 1GiB (VMs)
  • Live update: yes

Percentage of the host’s memory or a fixed value in bytes. Various suffixes are supported.

See Units for storage and network limits for details.

limits.memory.enforce

Whether the memory limit is hard or soft

  • Key: limits.memory.enforce
  • Type: string
  • Default: hard
  • Live update: yes
  • Condition: container

If the instance’s memory limit is hard, the instance cannot exceed its limit. If it is soft, the instance can exceed its memory limit when extra host memory is available.

limits.memory.hugepages

Whether to back the instance using huge pages

  • Key: limits.memory.hugepages
  • Type: bool
  • Default: false
  • Live update: no
  • Condition: virtual machine

If this option is set to false, regular system memory is used.

limits.memory.swap

Whether to encourage/discourage swapping less used pages for this instance

  • Key: limits.memory.swap
  • Type: bool
  • Default: true
  • Live update: yes
  • Condition: container

limits.memory.swap.priority

Prevents the instance from being swapped to disk

  • Key: limits.memory.swap.priority
  • Type: integer
  • Default: 10 (maximum)
  • Live update: yes
  • Condition: container

Specify an integer between 0 and 10. The higher the value, the less likely the instance is to be swapped to disk.

limits.processes

Maximum number of processes that can run in the instance

  • Key: limits.processes
  • Type: integer
  • Default: empty
  • Live update: yes
  • Condition: container

If left empty, no limit is set.

limits.kernel.*

Kernel resources per instance

  • Key: limits.kernel.*
  • Type: string
  • Live update: no
  • Condition: container

You can set kernel limits on an instance, for example, you can limit the number of open files. See Kernel resource limits for more information.

CPU limits

You have different options to limit CPU usage:

It is possible to set both options at the same time to restrict both which CPUs are visible to the instance and the allowed usage of those instances. However, if you use limits.cpu.allowance with a time limit, you should avoid using limits.cpu in addition, because that puts a lot of constraints on the scheduler and might lead to less efficient allocations.

The CPU limits are implemented through a mix of the cpuset and cpu cgroup controllers.

CPU pinning

limits.cpu results in CPU pinning through the cpuset controller. You can specify either which CPUs or how many CPUs are visible and available to the instance:

  • To specify which CPUs to use, set limits.cpu to either a set of CPUs (for example, 1,2,3) or a CPU range (for example, 0-3).

    To pin to a single CPU, use the range syntax (for example, 1-1) to differentiate it from a number of CPUs.

  • If you specify a number (for example, 4) of CPUs, LXD will do dynamic load-balancing of all instances that aren’t pinned to specific CPUs, trying to spread the load on the machine. Instances are re-balanced every time an instance starts or stops, as well as whenever a CPU is added to the system.

CPU limits for virtual machines

Note: LXD supports live-updating the limits.cpu option. However, for virtual machines, this only means that the respective CPUs are hotplugged. Depending on the guest operating system, you might need to either restart the instance or complete some manual actions to bring the new CPUs online.

LXD virtual machines default to having just one vCPU allocated, which shows up as matching the host CPU vendor and type, but has a single core and no threads.

When limits.cpu is set to a single integer, LXD allocates multiple vCPUs and exposes them to the guest as full cores. Unless limits.cpu.pin_strategy is set to auto, those vCPUs are not pinned to specific cores on the host. The number of vCPUs can be updated while the VM is running.

When limits.cpu is set to a range or comma-separated list of CPU IDs (as provided by lxc info --resources), the vCPUs are pinned to those cores. In this scenario, LXD checks whether the CPU configuration lines up with a realistic hardware topology and if it does, it replicates that topology in the guest. When doing CPU pinning, it is not possible to change the configuration while the VM is running.

For example, if the pinning configuration includes eight threads, with each pair of thread coming from the same core and an even number of cores spread across two CPUs, the guest will show two CPUs, each with two cores and each core with two threads. The NUMA layout is similarly replicated and in this scenario, the guest would most likely end up with two NUMA nodes, one for each CPU socket.

In such an environment with multiple NUMA nodes, the memory is similarly divided across NUMA nodes and be pinned accordingly on the host and then exposed to the guest.

All this allows for very high performance operations in the guest as the guest scheduler can properly reason about sockets, cores and threads as well as consider NUMA topology when sharing memory or moving processes across NUMA nodes.

Allowance and priority (container only)

limits.cpu.allowance drives either the CFS scheduler quotas when passed a time constraint, or the generic CPU shares mechanism when passed a percentage value:

  • The time constraint (for example, 20ms/50ms) is a hard limit. For example, if you want to allow the container to use a maximum of one CPU, set limits.cpu.allowance to a value like 100ms/100ms. The value is relative to one CPU worth of time, so to restrict to two CPUs worth of time, use something like 100ms/50ms or 200ms/100ms.
  • When using a percentage value, the limit is a soft limit that is applied only when under load. It is used to calculate the scheduler priority for the instance, relative to any other instance that is using the same CPU or CPUs. For example, to limit the CPU usage of the container to one CPU when under load, set limits.cpu.allowance to 100%.

limits.cpu.nodes can be used to restrict the CPUs that the instance can use to a specific set of NUMA nodes. To specify which NUMA nodes to use, set limits.cpu.nodes to either a set of NUMA node IDs (for example, 0,1) or a set of NUMA node ranges (for example, 0-1,2-4).

limits.cpu.priority is another factor that is used to compute the scheduler priority score when a number of instances sharing a set of CPUs have the same percentage of CPU assigned to them.

Huge page limits

LXD allows to limit the number of huge pages available to a container through the limits.hugepage.[size] key (for example, limits.hugepages.1MB).

Architectures often expose multiple huge-page sizes. The available huge-page sizes depend on the architecture.

Setting limits for huge pages is especially useful when LXD is configured to intercept the mount syscall for the hugetlbfs file system in unprivileged containers. When LXD intercepts a hugetlbfs mount syscall, it mounts the hugetlbfs file system for a container with correct uid and gid values as mount options. This makes it possible to use huge pages from unprivileged containers. However, it is recommended to limit the number of huge pages available to the container through limits.hugepages.[size] to stop the container from being able to exhaust the huge pages available to the host.

Limiting huge pages is done through the hugetlb cgroup controller, which means that the host system must expose the hugetlb controller in the legacy or unified cgroup hierarchy for these limits to apply.

Kernel resource limits

For container instances, LXD exposes a generic namespaced key limits.kernel.* that can be used to set resource limits.

It is generic in the sense that LXD does not perform any validation on the resource that is specified following the limits.kernel.* prefix. LXD cannot know about all the possible resources that a given kernel supports. Instead, LXD simply passes down the corresponding resource key after the limits.kernel.* prefix and its value to the kernel. The kernel does the appropriate validation. This allows users to specify any supported limit on their system.

Some common limits are:

KeyResourceDescription
limits.kernel.asRLIMIT_ASMaximum size of the process’s virtual memory
limits.kernel.coreRLIMIT_COREMaximum size of the process’s core dump file
limits.kernel.cpuRLIMIT_CPULimit in seconds on the amount of CPU time the process can consume
limits.kernel.dataRLIMIT_DATAMaximum size of the process’s data segment
limits.kernel.fsizeRLIMIT_FSIZEMaximum size of files the process may create
limits.kernel.locksRLIMIT_LOCKSLimit on the number of file locks that this process may establish
limits.kernel.memlockRLIMIT_MEMLOCKLimit on the number of bytes of memory that the process may lock in RAM
limits.kernel.niceRLIMIT_NICEMaximum value to which the process’s nice value can be raised
limits.kernel.nofileRLIMIT_NOFILEMaximum number of open files for the process
limits.kernel.nprocRLIMIT_NPROCMaximum number of processes that can be created for the user of the calling process
limits.kernel.rtprioRLIMIT_RTPRIOMaximum value on the real-time-priority that may be set for this process
limits.kernel.sigpendingRLIMIT_SIGPENDINGMaximum number of signals that may be queued for the user of the calling process

A full list of all available limits can be found in the manpages for the getrlimit(2)/setrlimit(2) system calls.

To specify a limit within the limits.kernel.* namespace, use the resource name in lowercase without the RLIMIT_ prefix. For example, RLIMIT_NOFILE should be specified as nofile.

A limit is specified as two colon-separated values that are either numeric or the word unlimited (for example, limits.kernel.nofile=1000:2000). A single value can be used as a shortcut to set both soft and hard limit to the same value (for example, limits.kernel.nofile=3000).

A resource with no explicitly configured limit will inherit its limit from the process that starts up the container. Note that this inheritance is not enforced by LXD but by the kernel.

Migration options

The following instance options control the behavior if the instance is moved from one LXD server to another:

migration.incremental.memory

Whether to use incremental memory transfer

  • Key: migration.incremental.memory
  • Type: bool
  • Default: false
  • Live update: yes
  • Condition: container

Using incremental memory transfer of the instance’s memory can reduce downtime.

migration.incremental.memory.goal

Percentage of memory to have in sync before stopping the instance

  • Key: migration.incremental.memory.goal
  • Type: integer
  • Default: 70
  • Live update: yes
  • Condition: container

migration.incremental.memory.iterations

Maximum number of transfer operations to go through before stopping the instance

  • Key: migration.incremental.memory.iterations
  • Type: integer
  • Default: 10
  • Live update: yes
  • Condition: container

migration.stateful

Whether to allow for stateful stop/start and snapshots

  • Key: migration.stateful
  • Type: bool
  • Default: false or value from profiles or instances.migration.stateful (if set)
  • Live update: no
  • Condition: virtual machine

Enabling this option prevents the use of some features that are incompatible with it.

NVIDIA and CUDA configuration

The following instance options specify the NVIDIA and CUDA configuration of the instance:

nvidia.driver.capabilities

What driver capabilities the instance needs

  • Key: nvidia.driver.capabilities
  • Type: string
  • Default: compute,utility
  • Live update: no
  • Condition: container

The specified driver capabilities are used to set libnvidia-container NVIDIA_DRIVER_CAPABILITIES.

nvidia.require.cuda

Required CUDA version

  • Key: nvidia.require.cuda
  • Type: string
  • Live update: no
  • Condition: container

The specified version expression is used to set libnvidia-container NVIDIA_REQUIRE_CUDA.

nvidia.require.driver

Required driver version

  • Key: nvidia.require.driver
  • Type: string
  • Live update: no
  • Condition: container

The specified version expression is used to set libnvidia-container NVIDIA_REQUIRE_DRIVER.

nvidia.runtime

Whether to pass the host NVIDIA and CUDA runtime libraries into the instance

  • Key: nvidia.runtime
  • Type: bool
  • Default: false
  • Live update: no
  • Condition: container

Raw instance configuration overrides

The following instance options allow direct interaction with the backend features that LXD itself uses:

raw.apparmor

AppArmor profile entries Key: raw.apparmor Type: blob Live update: yes The specified entries are appended to the generated profile.

raw.idmap

Raw idmap configuration Key: raw.idmap Type: blob Live update: no Condition: unprivileged container For example: both 1000 1000

raw.lxc

Raw LXC configuration to be appended to the generated one Key: raw.lxc Type: blob Live update: no Condition: container

raw.qemu

Raw QEMU configuration to be appended to the generated command line Key: raw.qemu Type: blob Live update: no Condition: virtual machine

raw.qemu.conf

Addition/override to the generated qemu.conf file Key: raw.qemu.conf Type: blob Live update: no Condition: virtual machine See Override QEMU configuration for more information.

raw.seccomp

Raw Seccomp configuration Key: raw.seccomp Type: blob Live update: no Condition: container

Important Setting these raw.* keys might break LXD in non-obvious ways. Therefore, you should avoid setting any of these keys.

Override QEMU configuration

For VM instances, LXD configures QEMU through a configuration file that is passed to QEMU with the -readconfig command-line option. This configuration file is generated for each instance before boot. It can be found at /var/log/lxd/<instance_name>/qemu.conf. The default configuration works fine for LXD’s most common use case: modern UEFI guests with VirtIO devices. In some situations, however, you might need to override the generated configuration. For example:

  • To run an old guest OS that doesn’t support UEFI.
  • To specify custom virtual devices when VirtIO is not supported by the guest OS.
  • To add devices that are not supported by LXD before the machines boots.
  • To remove devices that conflict with the guest OS.

To override the configuration, set the raw.qemu.conf option. It supports a format similar to qemu.conf, with some additions. Since it is a multi-line configuration option, you can use it to modify multiple sections or keys.

To replace a section or key in the generated configuration file, add a section with a different value. For example, use the following section to override the default virtio-gpu-pci GPU driver:

raw.qemu.conf: |-
[device "qemu_gpu"]
driver = "qxl-vga"

To remove a section, specify a section without any keys. For example:

raw.qemu.conf: |-
[device "qemu_gpu"]

To remove a key, specify an empty string as the value. For example:

raw.qemu.conf: |-
[device "qemu_gpu"]
driver = ""

To add a new section, specify a section name that is not present in the configuration file. The configuration file format used by QEMU allows multiple sections with the same name. Here’s a piece of the configuration generated by LXD:

[global]
driver = "ICH9-LPC"
property = "disable_s3"
value = "1"
[global]
driver = "ICH9-LPC"
property = "disable_s4"
value = "1"

To specify which section to override, specify an index. For example:

raw.qemu.conf: |-
[global][1]
value = "0"

Section indexes start at 0 (which is the default value when not specified), so the above example would generate the following configuration:

[global]
driver = "ICH9-LPC"
property = "disable_s3"
value = "1"
[global]
driver = "ICH9-LPC"
property = "disable_s4"
value = "0"

Security policies

The following instance options control the Security policies of the instance:

security.agent.metrics

Whether the lxd-agent is queried for state information and metrics Key: security.agent.metrics Type: bool Default: true Live update: no Condition: virtual machine

security.csm

Whether to use a firmware that supports UEFI-incompatible operating systems Key: security.csm Type: bool Default: false Live update: no Condition: virtual machine When enabling this option, set security.secureboot to false.

security.delegate_bpf

Whether to enable eBPF delegation using BPF Token mechanism Key: security.delegate_bpf Type: bool Default: false Live update: no Condition: unprivileged container This option enables BPF functionality delegation mechanism (using BPF Token). Note: security.delegate_bpf.cmd_types, security.delegate_bpf.map_types, security.delegate_bpf.prog_types, security.delegate_bpf.attach_types need to be configured depending on BPF workload in the container. See Privilege delegation using BPF Token for more information.

security.delegate_bpf.attach_types

Which eBPF attach types to allow with delegation mechanism Key: security.delegate_bpf.attach_types Type: bool Default: false Live update: no Condition: unprivileged container Which eBPF program attachment types to allow with delegation mechanism. Syntax follows a kernel one for delegate_attachs bpffs mount option. A number (bitmask) or :-separated list of attachment types to allow can be specified. For example, cgroup_inet_ingress allows BPF_CGROUP_INET_INGRESS attachment type.

security.delegate_bpf.cmd_types

Which eBPF commands to allow with delegation mechanism Key: security.delegate_bpf.cmd_types Type: bool Default: false Live update: no Condition: unprivileged container Which eBPF commands to allow with delegation mechanism. Syntax follows a kernel one for delegate_cmds bpffs mount option. A number (bitmask) or :-separated list of commands to allow can be specified. For example, prog_load:map_create allows eBPF programs loading and eBPF maps creation. Notice: security.delegate_bpf.prog_types and security.delegate_bpf.map_types still need to be configured accordingly.

security.delegate_bpf.map_types

Which eBPF maps to allow with delegation mechanism Key: security.delegate_bpf.map_types Type: bool Default: false Live update: no Condition: unprivileged container Which eBPF maps to allow with delegation mechanism. Syntax follows a kernel one for delegate_maps bpffs mount option. A number (bitmask) or :-separated list of map types to allow can be specified. For example, ringbuf allows BPF_MAP_TYPE_RINGBUF map.

security.delegate_bpf.prog_types

Which eBPF program types to allow with delegation mechanism Key: security.delegate_bpf.prog_types Type: bool Default: false Live update: no Condition: unprivileged container Which eBPF program types to allow with delegation mechanism. Syntax follows a kernel one for delegate_progs bpffs mount option. A number (bitmask) or :-separated list of program types to allow can be specified. For example, socket_filter allows BPF_PROG_TYPE_SOCKET_FILTER program type.

security.devlxd

Whether /dev/lxd is present in the instance Key: security.devlxd Type: bool Default: true Live update: no See Communication between instance and host for more information.

security.devlxd.images

Controls the availability of the /1.0/images API over devlxd Key: security.devlxd.images Type: bool Default: false Live update: yes

security.idmap.base

The base host ID to use for the allocation Key: security.idmap.base Type: integer Live update: no Condition: unprivileged container Setting this option overrides auto-detection.

security.idmap.isolated

Whether to use a unique idmap for this instance Key: security.idmap.isolated Type: bool Default: false Live update: no Condition: unprivileged container If specified, the idmap used for this instance is unique among instances that have this option set.

security.idmap.size

The size of the idmap to use Key: security.idmap.size Type: integer Live update: no Condition: unprivileged container

security.nesting

Whether to support running LXD (nested) inside the instance Key: security.nesting Type: bool Default: false Live update: yes Condition: container

security.privileged

Whether to run the instance in privileged mode Key: security.privileged Type: bool Default: false Live update: no Condition: container See Container security for more information.

security.protection.delete

Whether to prevent the instance from being deleted Key: security.protection.delete Type: bool Default: false Live update: container

security.protection.shift

Whether to protect the file system from being UID/GID shifted Key: security.protection.shift Type: bool Default: false Live update: yes Condition: container Set this option to true to prevent the instance’s file system from being UID/GID shifted on startup.

security.protection.start

Whether to prevent the instance from being started Key: security.protection.start Type: bool Default: false Live update: container

security.secureboot

Whether UEFI secure boot is enabled with the default Microsoft keys Key: security.secureboot Type: bool Default: true Live update: no Condition: virtual machine When disabling this option, consider enabling security.csm.

security.sev

Whether AMD SEV (Secure Encrypted Virtualization) is enabled for this VM Key: security.sev Type: bool Default: false Live update: no Condition: virtual machine

security.sev.policy.es

Whether AMD SEV-ES (SEV Encrypted State) is enabled for this VM Key: security.sev.policy.es Type: bool Default: false Live update: no Condition: virtual machine

security.sev.session.data

The guest owner’s base64-encoded session blob Key: security.sev.session.data Type: string Default: true Live update: no Condition: virtual machine

security.sev.session.dh

The guest owner’s base64-encoded Diffie-Hellman key Key: security.sev.session.dh Type: string Default: true Live update: no Condition: virtual machine

security.syscalls.allow

List of syscalls to allow Key: security.syscalls.allow Type: string Live update: no Condition: container A \n-separated list of syscalls to allow. This list must be mutually exclusive with security.syscalls.deny*.

security.syscalls.deny

List of syscalls to deny Key: security.syscalls.deny Type: string Live update: no Condition: container A \n-separated list of syscalls to deny. This list must be mutually exclusive with security.syscalls.allow.

security.syscalls.deny_compat

Whether to block compat_* syscalls (x86_64 only) Key: security.syscalls.deny_compat Type: bool Default: false Live update: no Condition: container On x86_64, this option controls whether to block compat_* syscalls. On other architectures, the option is ignored.

security.syscalls.deny_default

Whether to enable the default syscall deny Key: security.syscalls.deny_default Type: bool Default: true Live update: no Condition: container

security.syscalls.intercept.bpf

Whether to handle the bpf() system call Key: security.syscalls.intercept.bpf Type: bool Default: false Live update: no Condition: container

security.syscalls.intercept.bpf.devices

Whether to allow BPF programs Key: security.syscalls.intercept.bpf.devices Type: bool Default: false Live update: no Condition: container This option controls whether to allow BPF programs for the devices cgroup in the unified hierarchy to be loaded.

security.syscalls.intercept.mknod

Whether to handle the mknod and mknodat system calls Key: security.syscalls.intercept.mknod Type: bool Default: false Live update: no Condition: container These system calls allow creation of a limited subset of char/block devices.

security.syscalls.intercept.mount

Whether to handle the mount system call Key: security.syscalls.intercept.mount Type: bool Default: false Live update: no Condition: container

security.syscalls.intercept.mount.allowed

File systems that can be mounted Key: security.syscalls.intercept.mount.allowed Type: string Live update: yes Condition: container Specify a comma-separated list of file systems that are safe to mount for processes inside the instance.

security.syscalls.intercept.mount.fuse

File system that should be redirected to FUSE implementation Key: security.syscalls.intercept.mount.fuse Type: string Live update: yes Condition: container Specify the mounts of a given file system that should be redirected to their FUSE implementation (for example, ext4=fuse2fs).

security.syscalls.intercept.mount.shift

Whether to use idmapped mounts for syscall interception Key: security.syscalls.intercept.mount.shift Type: bool Default: false Live update: yes Condition: container

security.syscalls.intercept.sched_setscheduler

Whether to handle the sched_setscheduler system call Key: security.syscalls.intercept.sched_setscheduler Type: bool Default: false Live update: no Condition: container This system call allows increasing process priority.

security.syscalls.intercept.setxattr

Whether to handle the setxattr system call Key: security.syscalls.intercept.setxattr Type: bool Default: false Live update: no Condition: container This system call allows setting a limited subset of restricted extended attributes.

security.syscalls.intercept.sysinfo

Whether to handle the sysinfo system call Key: security.syscalls.intercept.sysinfo Type: bool Default: false Live update: no Condition: container This system call can be used to get cgroup-based resource usage information.

Snapshot scheduling and configuration

The following instance options control the creation and expiry of instance snapshots:

snapshots.expiry

When snapshots are to be deleted Key: snapshots.expiry Type: string Live update: no Specify an expression like 1M 2H 3d 4w 5m 6y.

snapshots.pattern

Template for the snapshot name Key: snapshots.pattern Type: string Default: snap%d Live update: no Specify a Pongo2 template string that represents the snapshot name. This template is used for scheduled snapshots and for unnamed snapshots. See Automatic snapshot names for more information.

snapshots.schedule

Schedule for automatic instance snapshots Key: snapshots.schedule Type: string Default: empty Live update: no Specify either a cron expression (<minute> <hour> <dom> <month> <dow>), a comma-separated list of schedule aliases (@hourly, @daily, @midnight, @weekly, @monthly, @annually, @yearly), or leave empty to disable automatic snapshots.

snapshots.schedule.stopped

Whether to automatically snapshot stopped instances Key: snapshots.schedule.stopped Type: bool Default: false Live update: no

Automatic snapshot names

The snapshots.pattern option takes a Pongo2 template string to format the snapshot name. To add a time stamp to the snapshot name, use the Pongo2 context variable creation_date. Make sure to format the date in your template string to avoid forbidden characters in the snapshot name. For example, set snapshots.pattern to {{ creation_date|date:'2006-01-02_15-04-05' }} to name the snapshots after their time of creation, down to the precision of a second. Another way to avoid name collisions is to use the placeholder %d in the pattern. For the first snapshot, the placeholder is replaced with 0. For subsequent snapshots, the existing snapshot names are taken into account to find the highest number at the placeholder’s position. This number is then incremented by one for the new name.

Volatile internal data

Warning The volatile.* keys cannot be manipulated by the user. Do not attempt to modify these keys in any way. LXD modifies these keys, and attempting to manipulate them yourself might break LXD in non-obvious ways.

The following volatile keys are currently used internally by LXD to store internal data specific to an instance:

volatile.<name>.apply_quota

Disk quota Key: volatile.<name>.apply_quota Type: string The disk quota is applied the next time the instance starts.

volatile.<name>.ceph_rbd

RBD device path for Ceph disk devices Key: volatile.<name>.ceph_rbd Type: string

volatile.<name>.host_name

Network device name on the host Key: volatile.<name>.host_name Type: string

volatile.<name>.hwaddr

Network device MAC address Key: volatile.<name>.hwaddr Type: string The network device MAC address is used when no hwaddr property is set on the device itself.

volatile.<name>.last_state.created

Whether the network device physical device was created Key: volatile.<name>.last_state.created Type: string Possible values are true or false.

volatile.<name>.last_state.hwaddr

Network device original MAC Key: volatile.<name>.last_state.hwaddr Type: string The original MAC that was used when moving a physical device into an instance.

volatile.<name>.last_state.mtu

Network device original MTU Key: volatile.<name>.last_state.mtu Type: string The original MTU that was used when moving a physical device into an instance.

volatile.<name>.last_state.vdpa.name

VDPA device name Key: volatile.<name>.last_state.vdpa.name Type: string The VDPA device name used when moving a VDPA device file descriptor into an instance.

volatile.<name>.last_state.vf.hwaddr

SR-IOV virtual function original MAC Key: volatile.<name>.last_state.vf.hwaddr Type: string The original MAC used when moving a VF into an instance.

volatile.<name>.last_state.vf.id

SR-IOV virtual function ID Key: volatile.<name>.last_state.vf.id Type: string The ID used when moving a VF into an instance.

volatile.<name>.last_state.vf.spoofcheck

SR-IOV virtual function original spoof check setting Key: volatile.<name>.last_state.vf.spoofcheck Type: string The original spoof check setting used when moving a VF into an instance.

volatile.<name>.last_state.vf.vlan

SR-IOV virtual function original VLAN Key: volatile.<name>.last_state.vf.vlan Type: string The original VLAN used when moving a VF into an instance.

volatile.apply_nvram

Whether to regenerate VM NVRAM the next time the instance starts Key: volatile.apply_nvram Type: bool

volatile.apply_template

Template hook Key: volatile.apply_template Type: string The template with the given name is triggered upon next startup.

volatile.base_image

Hash of the base image Key: volatile.base_image Type: string The hash of the image that the instance was created from (empty if the instance was not created from an image).

volatile.cloud-init.instance-id

instance-id (UUID) exposed to cloud-init Key: volatile.cloud-init.instance-id Type: string

volatile.evacuate.origin

The origin of the evacuated instance Key: volatile.evacuate.origin Type: string The cluster member that the instance lived on before evacuation.

volatile.idmap.base

The first ID in the container’s primary idmap range Key: volatile.idmap.base Type: integer Condition: container

volatile.idmap.current

The idmap currently in use by the container Key: volatile.idmap.current Type: string Condition: container

volatile.idmap.next

The idmap to use the next time the container starts Key: volatile.idmap.next Type: string Condition: container

volatile.last_state.idmap

On-disk UID/GID map for the container’s rootfs Key: volatile.last_state.idmap Type: string Condition: container The UID/GID map that has been applied to the container’s underlying storage. This is usually set for containers created on older kernels that don’t support idmapped mounts.

volatile.last_state.power

Instance state as of last host shutdown Key: volatile.last_state.power Type: string

volatile.uuid

Instance UUID Key: volatile.uuid Type: string The instance UUID is globally unique across all servers and projects.

volatile.uuid.generation

Instance generation UUID Key: volatile.uuid.generation Type: string The instance generation UUID changes whenever the instance’s place in time moves backwards. It is globally unique across all servers and projects.

volatile.vsock_id

Instance vsock ID used as of last start Key: volatile.vsock_id Type: string

Permissions