Sys::Virt::Domain - Represent & manage a libvirt guest domain
The Sys::Virt::Domain module represents a guest domain managed by the virtual machine monitor.
Sys::Virt::Domain
Returns an integer with a locally unique identifier for the domain.
Returns a 16 byte long string containing the raw globally unique identifier (UUID) for the domain.
Returns a printable string representation of the raw UUID, in the format 'XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX'.
Returns a string with a locally unique name of the domain
Returns a string representing the hostname of the guest. $flags can be zero or more of
$flags
Report the guest agent hostname
Report the DHCP lease hostname
Returns the metadata element of type $type associated with the domain. If $type is Sys::Virt::Domain::METADATA_ELEMENT then the $uri parameter specifies the XML namespace to retrieve, otherwise $uri should be undef. The optional $flags parameter defaults to zero.
$type
Sys::Virt::Domain::METADATA_ELEMENT
$uri
undef
Sets the metadata element of type $type to hold the value $val. If $type is Sys::Virt::Domain::METADATA_ELEMENT then the $key and $uri elements specify an XML namespace to use, otherwise they should both be undef. The optional $flags parameter defaults to zero.
$val
$key
Returns a true value if the domain is currently running
Returns a true value if the domain has a persistent configuration file defined
Returns a true value if the domain is running and has a persistent configuration file defined that is out of date compared to the current live config.
Returns an XML document containing a complete description of the domain's configuration. The optional $flags parameter controls generation of the XML document, defaulting to 0 if omitted. It can be one or more of the XML DUMP constants listed later in this document.
Returns a string containing the name of the OS type running within the domain.
Start a domain whose configuration was previously defined using the define_domain method in Sys::Virt. The $flags parameter accepts one of the DOMAIN CREATION constants documented later, and defaults to 0 if omitted.
define_domain
Start a domain whose configuration was previously defined using the define_domain method in Sys::Virt. The $fds parameter is an array of UNIX file descriptors which will be passed to the init process of the container. This is only supported with container based virtualization.The $flags parameter accepts one of the DOMAIN CREATION constants documented later, and defaults to 0 if omitted.
$fds
Remove the configuration associated with a domain previously defined with the define_domain method in Sys::Virt. If the domain is running, you probably want to use the shutdown or destroy methods instead.
shutdown
destroy
Temporarily stop execution of the domain, allowing later continuation by calling the resume method.
resume
Resume execution of a domain previously halted with the suspend method.
suspend
Wakeup the guest from power management suspend state
Tells the guest OS to enter the power management suspend state identified by $target. The $target parameter should be one of the NODE SUSPEND CONSTANTS listed in Sys::Virt. The $duration specifies when the guest should automatically wakeup. The $flags parameter is optional and defaults to zero.
$target
Sys::Virt
$duration
Take a snapshot of the domain's state and save the information to the file named in the $filename parameter. The domain can later be restored from this file with the restore_domain method on the Sys::Virt object. The optional $dxml parameter can be used to alter portions of the domain XML that will be used when performing restore. The $params parameter is a hash reference whose keys is a subset of the SAVE / RESTORE PARAMETER CONSTANTS. The $flags parameter accepts one of the SAVE / RESTORE FLAG CONSTANTS described later and defaults to zero.
$filename
restore_domain
$dxml
$params
Take a snapshot of the domain's state and save the information to a managed save location. The domain will be automatically restored with this state when it is next started. The $flags parameter is unused and defaults to zero.
Return a non-zero value if the domain has a managed save image that will be used at next start. The $flags parameter is unused and defaults to zero.
Remove the current managed save image, causing the guest to perform a full boot next time it is started. The $flags parameter is unused and defaults to zero.
Update the XML of the managed save image to $xml. The $flags parameter is unused and defaults to zero.
$xml
Get the XML in the managed save image. The $flags parameter accepts the following constants
Include security sensitive information in the XML dump, such as passwords.
Trigger a core dump of the guest virtual machine, saving its memory image to $filename so it can be analysed by tools such as crash. The optional $flags flags parameter is currently unused and if omitted will default to 0.
crash
Trigger a core dump of the guest virtual machine, saving its memory image to $filename so it can be analysed by tools such as crash. The $format parameter is one of the core dump format constants. The optional $flags flags parameter is currently unused and if omitted will default to 0.
$format
The raw ELF format
The zlib compressed ELF format
The snappy compressed ELF format
The lzo compressed ELF format
The Windows dump format
Immediately poweroff the machine. This is equivalent to removing the power plug. The guest OS is given no time to cleanup / save state. For a clean poweroff sequence, use the shutdown method instead.
Returns a hash reference summarising the execution state of the domain. The elements of the hash are as follows:
The maximum memory allowed for this domain, in kilobytes
The current memory allocated to the domain in kilobytes
The amount of CPU time used by the domain
The current number of virtual CPUs enabled in the domain
The execution state of the machine, which will be one of the constants &Sys::Virt::Domain::STATE_*.
Returns an array whose values specify the current state of the guest, and the reason for it being in that state. The $state values are the same as for the get_info API, and the $reason values come from:
$state
get_info
$reason
It is not known why the domain has crashed
The domain has crashed due to a kernel panic
It is not known why the domain has no state
The guest is paused due to a core dump operation
The guest is paused due to a snapshot
The guest is paused due to an I/O error
The guest is paused due to migration
The guest is paused due to a save operation
It is not known why the domain has paused
The guest is paused at admin request
The guest is paused due to the watchdog
The guest is paused while domain shutdown takes place
The guest is paused while a snapshot takes place
The guest is paused due to a kernel panic
The guest is paused as it is being started up.
The guest is paused as post-copy migration is taking place
The guest is paused as post-copy migration failed
The guest is paused as a result of failed resume operation at the end of some APIs which internally suspend guests.
The guest is running after being booted
The guest is running after restore from snapshot
The guest is running after migration
The guest is running after migration abort
The guest is running after restore from file
The guest is running after save cancel
It is not known why the domain has started
The guest is running after a resume
The guest is running after wakeup from power management suspend
The guest was restarted after crashing
The guest is running but post-copy is taking place
The guest is running, but migration failed in post-copy
The guest is blocked for an unknown reason
It is not known why the domain has shutdown
The guest is shutdown due to admin request
The guest is shutoff after a crash
The guest is shutoff after being destroyed
The guest is shutoff due to a virtualization failure
The guest is shutoff after a snapshot
The guest is shutoff after migration
The guest is shutoff after a save
The guest is shutoff due to controlled shutdown
It is not known why the domain has shutoff
The daemon stopped the guest due to a failure
It is not known why the domain was suspended to RAM
It is not known why the domain was suspended to disk
Returns a hash reference providing information about the control channel. The returned keys in the hash are
state
One of the CONTROL INFO constants listed later
details
Currently unused, always 0.
stateTime
The elapsed time since the control channel entered the current state.
Get the current time of the guest, in seconds and nanoseconds. The $flags parameter is currently unused and defaults to zero. The return value is an array ref with two elements, the first contains the time in seconds, the second contains the remaining nanoseconds.
Set the current time of the guest, in seconds and nanoseconds. The $flags parameter accepts one of
Sys::Virt::Domain::TIME_SYNC
Re-sync domain time from domain's RTC.
Update the password for account $username to be $password. $password is the clear-text password string unless the PASSWORD_ENCRYPTED flag is set.
$username
$password
Sys::Virt::Domain::PASSWORD_ENCRYPTED
The $password is encrypted with the password scheme required by the guest OS.
Change the name of an inactive guest to be $newname. The $flags parameter is currently unused and defaults to zero.
$newname
Returns a list of all disk errors that have occurred on the backing store for the guest's virtual disks. The returned array elements are hash references, containing two keys
path
The path of the disk with an error
error
The error type
Sends a sequence of keycodes to the guest domain. The $keycodeset should be one of the constants listed later in the KEYCODE SET section. $holdtiem is the duration, in milliseconds, to keep the key pressed before releasing it and sending the next keycode. @keycodes is an array reference containing the list of keycodes to send to the guest. The elements in the array should be keycode values from the specified keycode set. $flags is currently unused.
$keycodeset
$holdtiem
@keycodes
Returns a hash reference summarising the disk usage of the host backing store for a guest block device. The $dev parameter should be the path to the backing store on the host. $flags is currently unused and defaults to 0 if omitted. The returned hash contains the following elements
$dev
Logical size in bytes of the block device backing image *
Highest allocated extent in bytes of the block device backing image
Physical size in bytes of the container of the backing image
Set the maximum memory for the domain to the value $mem. The value of the $mem parameter is specified in kilobytes.
$mem
Returns the current maximum memory allowed for this domain in kilobytes.
Set the current memory for the domain to the value $mem. The value of the $mem parameter is specified in kilobytes. This must be less than, or equal to the domain's max memory limit. The $flags parameter can control whether the update affects the live guest, or inactive config, defaulting to modifying the current state.
Set the period on which guests memory stats are refreshed, with $period being a value in seconds. The $flags parameter is currently unused.
$period
Request that the guest OS perform a graceful shutdown and poweroff. This usually requires some form of cooperation from the guest operating system, such as responding to an ACPI signal, or a guest agent process. For an immediate, forceful poweroff, use the destroy method instead.
Request that the guest OS perform a graceful shutdown and optionally restart. The optional $flags parameter is currently unused and if omitted defaults to zero.
Perform a hardware reset of the virtual machine. The guest OS is given no opportunity to shutdown gracefully. The optional $flags parameter is currently unused and if omitted defaults to zero.
Return the maximum number of vcpus that are configured for the domain
Hotplug a new device whose configuration is given by $xml, to the running guest. The optional <$flags> parameter defaults to 0, but can accept one of the device hotplug flags described later.
Hotunplug an existing device whose configuration is given by $xml, from the running guest. The optional <$flags> parameter defaults to 0, but can accept one of the device hotplug flags described later.
Hotunplug an existing device which is identified by $alias. The optional <$flags> parameter defaults to 0, but can accept one of the device hotplug flags described later.
$alias
Update the configuration of an existing device. The new configuration is given by $xml. The optional <$flags> parameter defaults to 0 but can accept one of the device hotplug flags described later.
Peek into the guest disk $path, at byte $offset capturing $size bytes of data. The returned scalar may contain embedded NULLs. The optional $flags parameter is currently unused and if omitted defaults to zero.
$path
$offset
$size
Peek into the guest memory at byte $offset virtual address, capturing $size bytes of memory. The return scalar may contain embedded NULLs. The optional $flags parameter is currently unused and if omitted defaults to zero.
Return a true value if the guest domain is configured to automatically start upon boot. Return false, otherwise
Set the state of the autostart flag, which determines whether the guest will automatically start upon boot of the host OS
Set the number of virtual CPUs in the guest VM to $count. The optional $flags parameter can be used to control whether the setting changes the live config or inactive config.
$count
Set the state of the CPUs in $cpumap to $state. The $flags parameter defaults to zero if not present.
$cpumap
Get the number of virtual CPUs in the guest VM. The optional $flags parameter can be used to control whether to query the setting of the live config or inactive config.
Set the online status of the guest OS CPUs. The $cpumap parameter describes the set of CPUs to modify (e.g. "0-3,^1"). $state is either 1 to set the CPUs online, or 0 to set them offline. The $flags parameter is currently unused and defaults to 0.
Query information about the guest OS CPUs. The returned data is a hash reference with the following keys.
String containing bitmap representing CPU ids reported currently known to the guest.
String containing bitmap representing CPU ids that are currently online in the guest.
String containing bitmap representing CPU ids that can be offlined in the guest.
The $flags parameter is currently unused and defaults to 0.
Return the scheduler type for the guest domain
Fetch the current I/O statistics for the block device given by $path. The returned hash reference contains keys for
rd_req
Number of read requests
rd_bytes
Number of bytes read
wr_req
Number of write requests
wr_bytes
Number of bytes written
errs
Some kind of error count
Return the set of scheduler tunable parameters for the guest, as a hash reference. The precise set of keys in the hash are specific to the hypervisor.
Update the set of scheduler tunable parameters. The value names for tunables vary, and can be discovered using the get_scheduler_params call
get_scheduler_params
Return a hash reference containing the set of memory tunable parameters for the guest. The keys in the hash are one of the constants MEMORY PARAMETERS described later. The $flags parameter accepts one or more the CONFIG OPTION constants documented later, and defaults to 0 if omitted.
Update the memory tunable parameters for the guest. The $params should be a hash reference whose keys are one of the MEMORY PARAMETERS constants. The $flags parameter accepts one or more the CONFIG OPTION constants documented later, and defaults to 0 if omitted.
Return a hash reference containing the set of blkio tunable parameters for the guest. The keys in the hash are one of the constants BLKIO PARAMETERS described later. The $flags parameter accepts one or more the CONFIG OPTION constants documented later, and defaults to 0 if omitted.
Update the blkio tunable parameters for the guest. The $params should be a hash reference whose keys are one of the BLKIO PARAMETERS constants. The $flags parameter accepts one or more the CONFIG OPTION constants documented later, and defaults to 0 if omitted.
Return a hash reference containing the set of blkio tunable parameters for the guest disk $disk. The keys in the hash are one of the constants BLOCK IOTUNE PARAMETERS described later.
$disk
Update the blkio tunable parameters for the guest disk $disk. The $params should be a hash reference whose keys are one of the BLOCK IOTUNE PARAMETERS constants.
Return a hash reference containing the set of interface tunable parameters for the guest. The keys in the hash are one of the constants INTERFACE PARAMETERS described later.
Update the interface tunable parameters for the guest. The $params should be a hash reference whose keys are one of the INTERFACE PARAMETERS constants.
Return a hash reference containing the set of numa tunable parameters for the guest. The keys in the hash are one of the constants NUMA PARAMETERS described later. The $flags parameter accepts one or more the CONFIG OPTION constants documented later, and defaults to 0 if omitted.
Update the numa tunable parameters for the guest. The $params should be a hash reference whose keys are one of the NUMA PARAMETERS constants. The $flags parameter accepts one or more the CONFIG OPTION constants documented later, and defaults to 0 if omitted.
Return a hash reference containing the set of performance events that are available for the guest. The keys in the hash are one of the constants PERF EVENTS described later. The $flags parameter accepts one or more the CONFIG OPTION constants documented later, and defaults to 0 if omitted.
Update the enabled state for performance events for the guest. The $params should be a hash reference whose keys are one of the PERF EVENTS constants. The $flags parameter accepts one or more the CONFIG OPTION constants documented later, and defaults to 0 if omitted.
Resize the disk $disk to have new size $newsize KB. If the disk is backed by a special image format, the actual resize is done by the hypervisor. If the disk is backed by a raw file, or block device, the resize must be done prior to invoking this API call, and it merely updates the hypervisor's view of the disk size. The following flags may be used
$newsize
Treat $newsize as if it were in bytes, rather than KB.
Resize to full the capacity of the source
Fetch the current I/O statistics for the block device given by $path. The returned hash contains keys for
rx_bytes
Total bytes received
rx_packets
Total packets received
rx_errs
Total packets received with errors
rx_drop
Total packets drop at reception
tx_bytes
Total bytes transmitted
tx_packets
Total packets transmitted
tx_errs
Total packets transmitted with errors
tx_drop
Total packets dropped at transmission.
Fetch the current memory statistics for the guest domain. The $flags parameter is currently unused and can be omitted. The returned hash contains keys for
swap_in
Data read from swap space
swap_out
Data written to swap space
major_fault
Page fault involving disk I/O
minor_fault
Page fault not involving disk I/O
unused
Memory not used by the system
available
Total memory seen by guest
rss
Resident set size. Size of memory resident in host RAM.
Fetch information about the security label assigned to the guest domain. The returned hash reference has two keys, model gives the name of the security model in effect (e.g. selinux), while label provides the name of the security label applied to the domain. This method only returns information about the first security label. To retrieve all labels, use get_security_label_list.
model
selinux
label
get_security_label_list
Fetches information about all security labels assigned to the guest domain. The elements in the returned array are all hash references, whose keys are as described for get_security_label.
get_security_label
Migrate a domain to an alternative host. The destcon parameter should be a Sys::Virt connection to the remote target host. The flags parameter takes one or more of the Sys::Virt::Domain::MIGRATE_XXX constants described later in this document. The %params parameter is a hash reference used to set various parameters for the migration operation, with the following valid keys.
destcon
flags
Sys::Virt::Domain::MIGRATE_XXX
%params
Sys::Virt::Domain::MIGRATE_PARAM_URI
The URI to use for initializing the domain migration. It takes a hypervisor specific format. The uri_transports element of the hypervisor capabilities XML includes details of the supported URI schemes. When omitted libvirt will auto-generate suitable default URI. It is typically only necessary to specify this URI if the destination host has multiple interfaces and a specific interface is required to transmit migration data.
Sys::Virt::Domain::MIGRATE_PARAM_DEST_NAME
The name to be used for the domain on the destination host. Omitting this parameter keeps the domain name the same. This field is only allowed to be used with hypervisors that support domain renaming during migration.
Sys::Virt::Domain::MIGRATE_PARAM_DEST_XML
The new configuration to be used for the domain on the destination host. The configuration must include an identical set of virtual devices, to ensure a stable guest ABI across migration. Only parameters related to host side configuration can be changed in the XML. Hypervisors which support this field will forbid migration if the provided XML would cause a change in the guest ABI. This field cannot be used to rename the domain during migration (use VIR_MIGRATE_PARAM_DEST_NAME field for that purpose). Domain name in the destination XML must match the original domain name.
Omitting this parameter keeps the original domain configuration. Using this field with hypervisors that do not support changing domain configuration during migration will result in a failure.
Sys::Virt::Domain::MIGRATE_PARAM_GRAPHICS_URI
URI to use for migrating client's connection to domain's graphical console as VIR_TYPED_PARAM_STRING. If specified, the client will be asked to automatically reconnect using these parameters instead of the automatically computed ones. This can be useful if, e.g., the client does not have a direct access to the network virtualization hosts are connected to and needs to connect through a proxy. The URI is formed as follows:
protocol://hostname[:port]/[?parameters]
where protocol is either "spice" or "vnc" and parameters is a list of protocol specific parameters separated by '&'. Currently recognized parameters are "tlsPort" and "tlsSubject". For example,
spice://target.host.com:1234/?tlsPort=4567
Sys::Virt::Domain::MIGRATE_PARAM_BANDWIDTH
The maximum bandwidth (in MiB/s) that will be used for migration. If set to 0 or omitted, libvirt will choose a suitable default. Some hypervisors do not support this feature and will return an error if this field is used and is not 0.
Sys::Virt::Domain::MIGRATE_PARAM_BANDWIDTH_POSTCOPY
The maximum bandwidth (in MiB/s) that will be used for migration during post-copy phase. If set to 0 or omitted, libvirt will choose a suitable default. Some hypervisors do not support this feature and return an error if this field is used and is not 0.
Sys::Virt::Domain::MIGRATE_PARAM_LISTEN_ADDRESS
The address on which to listen for incoming migration connections. If omitted, libvirt will listen on the wildcard address (0.0.0.0 or ::). This default may be a security risk if guests, or other untrusted users have the ability to connect to the virtualization host, thus use of an explicit restricted listen address is recommended.
Sys::Virt::Domain::MIGRATE_PARAM_DISK_PORT
Port that destination server should use for incoming disks migration. Type is VIR_TYPED_PARAM_INT. If set to 0 or omitted, libvirt will choose a suitable default. At the moment this is only supported by the QEMU driver.
Sys::Virt::Domain::MIGRATE_PARAM_MIGRATE_DISKS
The list of disks to migrate when doing block storage migration. In contrast to other parameters whose values are plain strings, the parameter value should be an array reference, whose elements are in turn strings representing the disk target names.
Sys::Virt::Domain::MIGRATE_PARAM_COMPRESSION
The type of compression method use use, either xbzrle or mt.
xbzrle
mt
Sys::Virt::Domain::MIGRATE_PARAM_COMPRESSION_MT_THREADS
The number of compression threads to use
Sys::Virt::Domain::MIGRATE_PARAM_COMPRESSION_MT_DTHREADS
The number of decompression threads
Sys::Virt::Domain::MIGRATE_PARAM_COMPRESSION_MT_LEVEL
The compression level from 0 (no compression) to 9 (maximum compression)
Sys::Virt::Domain::MIGRATE_PARAM_COMPRESSION_XBZRLE_CACHE
The size of the cache for xbzrle compression
Sys::Virt::Domain::MIGRATE_PARAM_PERSIST_XML
The alternative persistent XML config to copy
Sys::Virt::Domain::MIGRATE_PARAM_AUTO_CONVERGE_INITIAL
The initial percentage to throttle guest vCPUs
Sys::Virt::Domain::MIGRATE_PARAM_AUTO_CONVERGE_INCREMENT
The additional percentage step size to throttle guest vCPUs if progress is not made
Sys::Virt::Domain::MIGRATE_PARAM_PARALLEL_CONNECTIONS
The number of connections used during parallel migration.
Sys::Virt::Domain::MIGRATE_PARAM_TLS_DESTINATION
Override the destination host name used for TLS verification. Normally the TLS certificate from the destination host must match the host's name for TLS verification to succeed. When the certificate does not match the destination hostname and the expected certificate's hostname is known, this parameter can be used to pass this expected hostname when starting the migration.
Sys::Virt::Domain::MIGRATE_PARAM_DISKS_URI
The URI to use for initializing the domain migration for storage. It takes a hypervisor specific format. The uri_transports element of the hypervisor capabilities XML includes details of the supported URI schemes. When omitted libvirt will auto-generate suitable default URI. It is typically only necessary to specify this URI if the destination host has multiple interfaces and a specific interface is required to transmit storage data.
Sys::Virt::Domain::MIGRATE_PARAM_ZLIB_LEVEL
The level of compression for zlib. Accepted values are in range 0-9. 0 is no compression, 1 is maximum speed and 9 is maximum compression.
Sys::Virt::Domain::MIGRATE_PARAM_ZSTD_LEVEL
The level of compression for zstd. Accepted values are in range 0-20. 0 is no compression, 1 is maximum speed and 20 is maximum compression.
Migrate a domain to an alternative host. Use of positional parameters with migrate is deprecated in favour of passing a hash reference as described above.
migrate
Migrate a domain to an alternative host. This method is deprecated in favour of passing a hash ref to migrate.
Migrate a domain to an alternative host. The desturi parameter should be a valid libvirt connection URI for the remote target host. The flags parameter takes one or more of the Sys::Virt::Domain::MIGRATE_XXX constants described later in this document. The %params parameter is a hash reference used to set various parameters for the migration operation, with the same keys described for the migrate API.
desturi
Migrate a domain to an alternative host. Use of positional parameters with migrate_to_uri is deprecated in favour of passing a hash reference as described above.
migrate_to_uri
Migrate a domain to an alternative host. This method is deprecated in favour of passing a hash ref to migrate_to_uri.
Set the maximum allowed downtime during migration of the guest. A longer downtime makes it more likely that migration will complete, at the cost of longer time blackout for the guest OS at the switch over point. The downtime parameter is measured in milliseconds. The $flags parameter is currently unused and defaults to zero.
downtime
Get the current value of the maximum downtime allowed during a migration of a guest. The returned <downtime> value is measured in milliseconds. The $flags parameter is currently unused and defaults to zero.
Set the maximum allowed bandwidth during migration of the guest. The bandwidth parameter is measured in MB/second. The $flags parameter takes zero or more of the constants:
bandwidth
Set the post-copy speed instead of the pre-copy speed.
Get the maximum allowed bandwidth during migration of the guest. The returned <bandwidth> value is measured in MB/second. The $flags parameter is accepts the same constants as migrate_set_max_speed.
migrate_set_max_speed
Set the maximum allowed compression cache size during migration of the guest. The cacheSize parameter is measured in bytes. The $flags parameter is currently unused and defaults to zero.
cacheSize
Get the maximum allowed compression cache size during migration of the guest. The returned <bandwidth> value is measured in bytes. The $flags parameter is currently unused and defaults to zero.
Switch the domain from pre-copy to post-copy mode. This requires that the original migrate command had the Sys::Virt::Domain::MIGRATE_POST_COPY flag specified.
Sys::Virt::Domain::MIGRATE_POST_COPY
Trigger an NMI in the guest virtual machine. The $flags parameter is currently unused and defaults to 0.
Open the text console for a serial, parallel or paravirt console device identified by $devname, connecting it to the stream $st. If $devname is undefined, the default console will be opened. $st must be a Sys::Virt::Stream object used for bi-directional communication with the console. $flags is currently unused, defaulting to 0.
$devname
$st
Sys::Virt::Stream
Open the text console for a data channel device identified by $devname, connecting it to the stream $st. $st must be a Sys::Virt::Stream object used for bi-directional communication with the channel. $flags is currently unused, defaulting to 0.
Open the graphics console for a guest, identified by $idx, counting from 0. The $fd should be a file descriptor for an anonymous socket pair. The $flags argument should be one of the constants listed at the end of this document, and defaults to 0.
$idx
$fd
Open the graphics console for a guest, identified by $idx, counting from 0. The $flags argument should be one of the constants listed at the end of this document, and defaults to 0. The return value will be a file descriptor connected to the console which must be closed when no longer needed. This method is preferred over open_graphics since it will work correctly under sVirt mandatory access control policies.
open_graphics
Capture a screenshot of the virtual machine's monitor. The $screen parameter controls which monitor is captured when using a multi-head or multi-card configuration. $st must be a Sys::Virt::Stream object from which the data can be read. $flags is currently unused and defaults to 0. The mimetype of the screenshot is returned
$screen
Obtain information about the state of all virtual CPUs in a running guest domain. The returned list will have one element for each vCPU, where each elements contains a hash reference. The keys in the hash are, number the vCPU number, cpu the physical CPU on which the vCPU is currently scheduled, cpuTime the cumulative execution time of the vCPU, state the running state and affinity giving the allowed scheduler placement. The value for affinity is a string representing a bitmask against physical CPUs, 8 cpus per character. To extract the bits use the unpack function with the b* template. NB The state, cpuTime, cpu values are only available if using $flags value of 0, and the domain is currently running; otherwise they will all be set to zero.
number
cpu
cpuTime
affinity
unpack
b*
Pin the virtual CPU given by index $vcpu to physical CPUs given by $mask. The $mask is a string representing a bitmask against physical CPUs, 8 cpus per character.
$vcpu
$mask
Obtain information about the CPU affinity of the emulator process. The returned $mask is a bitstring against physical CPUs, 8 cpus per character. To extract the bits use the unpack function with the b* template.
Pin the emulator threads to the physical CPUs identified by the affinity in $newmask. The $newmask is a bitstring against the physical CPUa, 8 cpus per character. To create a suitable bitstring, use the vec function with a value of 1 for the BITS parameter.
$newmask
vec
1
BITS
Obtain information about the state of all IOThreads in a running guest domain. The returned list will have one element for each IOThread, where each elements contains a hash reference. The keys in the hash are, number the IOThread number and affinity giving the allowed scheduler placement. The value for affinity is a string representing a bitmask against physical CPUs, 8 cpus per character. To extract the bits use the unpack function with the b* template.
Pin the IOThread given by index $iothread to physical CPUs given by $mask. The $mask is a string representing a bitmask against physical CPUs, 8 cpus per character.
$iothread
Add a new IOThread by the $iothread value to the guest domain. The $flags parameter accepts one or more the CONFIG OPTION constants documented later, and defaults to 0 if omitted.
Delete an existing IOThread by the $iothread value from the guest domain. The $flags parameter accepts one or more the CONFIG OPTION constants documented later, and defaults to 0 if omitted.
Set parameters for the IOThread by the $iothread value on the guest domain. The $params parameter is a hash reference whose keys are the IOTHREAD STATS constants documented later. The $flags parameter accepts one or more the CONFIG OPTION constants documented later, and defaults to 0 if omitted.
IOTHREAD STATS
Requests the guests host physical CPU usage statistics, starting from host CPU <$startCpu> counting up to $numCpus. If $startCpu is -1 and $numCpus is 1, then the utilization across all CPUs is returned. Returns an array of hash references, each element containing stats for one CPU.
$numCpus
$startCpu
Returns a hash reference summarising the execution state of the background job. The elements of the hash are as follows:
The type of job, one of the JOB TYPE constants listed later in this document.
The elapsed time in milliseconds
The expected remaining time in milliseconds. Only set if the type is JOB_UNBOUNDED.
type
The total amount of data expected to be processed by the job, in bytes.
The current amount of data processed by the job, in bytes.
The expected amount of data remaining to be processed by the job, in bytes.
The total amount of memory expected to be processed by the job, in bytes.
The current amount of memory processed by the job, in bytes.
The expected amount of memory remaining to be processed by the job, in bytes.
The total amount of file expected to be processed by the job, in bytes.
The current amount of file processed by the job, in bytes.
The expected amount of file remaining to be processed by the job, in bytes.
Returns an array summarising the execution state of the background job. The $type value is one of the JOB TYPE constants listed later in this document. The $stats value is a hash reference, whose elements are one of the following constants.
$stats
The $flags parameter defaults to zero and can take one of the following constants.
Return the stats of the most recently completed job.
Don't clear the completed stats after reading them.
Time in milliseconds since the beginning of the migration job NOT including the time required to transfer control flow from the source host to the destination host.
The number of pages filled with a constant byte which have been transferred
The number of pages transferred without any compression
The number of bytes transferred without any compression
The bytes per second transferred
The number of memory pages dirtied per second
The memory page size in bytes
The total number of iterations over guest memory
The number of page requests received from the destination host during post-copy migration.
The percentage by which vCPUs are currently throttled
The size of the compression cache in bytes
The number of compressed bytes transferred
The number of compressed pages transferred
The number of changing pages not in compression cache
The number of changing pages in the compression cache but sent uncompressed since the compressed page was larger than the non-compressed page.
The number of milliseconds of downtime expected during migration switchover.
Real measured downtime (ms) NOT including the time required to transfer control flow from the source host to the destination host.
The number of milliseconds of time doing setup of the job
The type of operation associated with the job
Whether the job was successfully completed.
Possible total temporary disk space for the job in bytes
Current total temporary disk space for the job in bytes
The error message from a failed job
The values for the Sys::Virt::Domain::JOB_OPERATION field are
No known job type
The guest is starting
The guest is saving to disk
The guest is restoring from disk
The guest is migrating in from another host
The guest is migrating out to another host
The guest is saving a snapshot
The guest is reverting to a snapshot
The guest is saving a crash dump
The guest is performing a block backup
The guest is deleting a snapshot
Aborts the currently executing job. Valid $flags include:
Interrupt post-copy migration. It can later be resumed using migrate APIs with MIGRATE_POSTCOPY_RESUME flag.
MIGRATE_POSTCOPY_RESUME
Returns a hash reference summarising the execution state of the block job. The $path parameter should be the fully qualified path of the block device being changed. Valid $flags include:
Treat bandwidth value as bytes instead of MiB.
Change the maximum I/O bandwidth used by the block job that is currently executing for $path. The $bandwidth argument is specified in MB/s. The $flags parameter can take the bitwise union of the values:
$bandwidth
The $bandwidth parameter value is measured in bytes/s instead of MB/s.
Abort the current job that is executing for the block device associated with $path
Merge the backing files associated with $path into the top level file. The $bandwidth parameter specifies the maximum I/O rate to allow in MB/s. The $flags parameter can take the bitwise union of the values:
Switch the backing path associated with $path to instead use $base. The $bandwidth parameter specifies the maximum I/O rate to allow in MB/s. The $flags parameter can take the bitwise union of the values:
$base
Copy contents of a disk image <$path> into the target volume described by $destxml which follows the schema of the <disk> element in the domain XML. The $params parameter is a hash of optional parameters to control the process
$destxml
The maximum bandwidth in bytes per second.
The granularity in bytes of the copy process
The maximum amount of data in flight in bytes.
Commit changes there were made to the temporary top level file $top. Takes all the differences between $top and $base and merge them into $base. The $bandwidth parameter specifies the maximum I/O rate to allow in MB/s. The $flags parameter can take the bitwise union of the values:
$top
The $bandwidth parameter value is measured in bytes instead of MB/s.
Return the number of saved snapshots of the domain
List the names of all saved snapshots. The names can be used with the lookup_snapshot_by_name
lookup_snapshot_by_name
Return a list of all snapshots currently known to the domain. The elements in the returned list are instances of the Sys::Virt::DomainSnapshot class. This method requires O(n) RPC calls, so the list_all_snapshots method is recommended as a more efficient alternative.
list_all_snapshots
Return a list of all domain snapshots associated with this domain. The elements in the returned list are instances of the Sys::Virt::DomainSnapshot class. The $flags parameter can be used to filter the list of return domain snapshots.
Return the domain snapshot with a name of $name. The returned object is an instance of the Sys::Virt::DomainSnapshot class.
$name
Returns a true value if the domain has a currently active snapshot
Returns the currently active snapshot for the domain.
Create a new snapshot from the $xml. The $flags parameter accepts the SNAPSHOT CREATION constants listed in Sys::Virt::DomainSnapshots.
Sys::Virt::DomainSnapshots
Return a list of all domain checkpoints associated with this domain. The elements in the returned list are instances of the Sys::Virt::DomainCheckpoint class. The $flags parameter can be used to filter the list of return domain checkpoints.
Return the domain checkpoint with a name of $name. The returned object is an instance of the Sys::Virt::DomainCheckpoint class.
Create a new checkpoint from the $xml. The $flags parameter accepts the CHECKPOINT CREATION constants listed in Sys::Virt::DomainCheckpoints.
Sys::Virt::DomainCheckpoints
Start a point-in-time backup job for the specified disks of a running domain. The $backupxml parameter describes the backup operation, including which disks to use. The optional $checkpointxml parameter can be used to create a checkpoint covering to the same point in time as the backup. The optional $flags parameter can be one of the following constants:
$backupxml
$checkpointxml
Assume that the output/temporary files for the backup have been precreated by the caller with the correct size and format.
Get the XML description of the currently executing backup job. If there is no backup job then an error is raised.
Issue an FS_TRIM command to the device at $mountPoint to remove chunks of unused space that are at least $minimum bytes in length. $flags is currently unused and defaults to zero.
$mountPoint
$minimum
Freeze all the filesystems associated with the @mountPoints array reference. If <@mountPoints> is an empty list, then all filesystems will be frozen. $flags is currently unused and defaults to zero.
@mountPoints
Thaw all the filesystems associated with the @mountPoints array reference. If <@mountPoints> is an empty list, then all filesystems will be thawed. $flags is currently unused and defaults to zero.
Obtain a list of all guest filesystems. The returned list will contain one element for each filesystem, whose value will be a hash reference with the following keys
The name of the guest device that is mounted
The filesystem type (e.g. 'ext4', 'fat', 'ntfs', etc)
The location in the filesystem tree of the mount
An array reference containing list of device aliases associated with the guest device. The device aliases correspond to disk target names in the guest XML configuration
Obtain a list of all guest network interfaces. The $src parameter is one of the constants
$src
Extract the DHCP server lease information
Query the guest OS via an agent
Extract from the local ARP tables
The returned list will contain one element for each interface. The values in the list will be a hash reference with the following keys
The name of the guest interface that is mounted
The hardware address, aka MAC, if available.
An array reference containing list of IP addresses associated with the guest NIC. Each element in the array is a further hash containing
The IP address string
The IP address network prefix
The IP address type (IPv4 vs IPv6)
Send the process $pid the signal $signum. The $signum value must be one of the constants listed later, not a POSIX or Linux signal value. $flags is currently unused and defaults to zero.
$pid
$signum
Set the threshold level for delivering the EVENT_ID_BLOCK_THRESHOLD if the device or backing chain element described by $dev is written beyond the set $threshold level. The threshold level is unset once the event fires. The event might not be delivered at all if libvirtd was not running at the moment when the threshold was reached.
$threshold
Changes the actions of lifecycle events for domain represented as <on_$type>$action</on_$type> in the domain XML.
Get information about the domain launch security policy. $flags is currently unused and defaults to zero. The returned hash may contain the following keys
The AMD SEV launch measurement
The host SEV API major version
The host SEV API minor version
The host SEV firmware build ID
The guest SEV policy
Set information about the domain launch security state. $flags is currently unused and defaults to zero. The provided hash may contain the following keys
The SEV secret string to inject
The SEV secret header string to inject
The address at which to inject the SEV secret. If omitted it can be automatically determined from the firmware
Get information about the domain guest configuration. The $types parameter determines what pieces of information are returned and should be the bitwise or of the following constants:
$types
Active user information
Misc operating system information
The guest timezone
The guest hostname
Filesystem mount information
Block device information
Network interfaces information
$flags is currently unused and defaults to zero.
Set the amount of time to wait for the agent to respond to a command. $timeout is a positive integer representing the number of seconds to wait, or one of the constants:
$timeout
Wait forever with no timeout
Don't want at all, return immediately
Use the hypervisor driver's default timeout
The $flags parameter is currently unused and defaults to zero.
Retrieve the list of authorized SSH keys for the user account $user. The $flags parameter is currently unused and defaults to zero.
$user
Update the list of authorized SSH keys for the user account $user. The @keys parameter should be an array reference containing the new keys, if any. The default behaviour is to set the authorized SSH keys to the exact set specified in @keys. This can be modified via the $flags parameter which takes the following constants
@keys
Append @keys to the current set of keys, rather than replacing the existing keys.
Remove @keys from the current set of keys, rather than replacing the existing keys. It is not an error if the keys do not exist.
Retrieve a list of messages associated with the domain. The optional $flags parameter can accept zero or more of
Warnings about use of deprecated features
Warnings about actions that have tainting the domain
Request calculation of the domain's memory dirty rate over the next $secs seconds. $flags accepts one or more of
$secs
Page sampling mode
Dirty bitmap mode
Dirty ring mode
Associate a set of file descriptors in @fds with the domain, giving them the name $name. If $flags is non-zero, it can take one or more of:
@fds
Attempt to restore the label when shutting down
Label the FD to permit the guest to write to it.
Request a reload of graphics backend state. The $type parameter specifies what type of graphics backend to reload if there are multiple present. It takes one of the constants:
Reload state for any graphics backend type
Reload state for only VNC graphics backends
A number of the APIs take a flags parameter. In most cases passing a value of zero will be satisfactory. Some APIs, however, accept named constants to alter their behaviour. This section documents the current known constants.
The domain state constants are useful in interpreting the state key in the hash returned by the get_info method.
The domain is active, but is not running / blocked (e.g. idle)
The domain is active and running
The domain is active, but execution is blocked
The domain is active, but execution has been paused
The domain is active, but in the shutdown phase
The domain is inactive, and shut down.
The domain is inactive, and crashed.
The domain is active, but in power management suspend state
The following constants can be used to determine what the guest domain control channel status is
The control channel has a fatal error
The control channel is ready for jobs
The control channel is busy
The control channel is busy with a job
If the status is Sys::Virt::Domain::CONTROL_ERROR, then one of the following constants describes the reason
Sys::Virt::Domain::CONTROL_ERROR
There is no reason for the error available
The reason for the error is unknown
There was an internal error in libvirt
There was an error speaking to the monitor
The following constants can be used to control the behaviour of domain creation
Keep the guest vCPUs paused after starting the guest
Automatically destroy the guest when the connection is closed (or fails)
Do not use OS I/O cache if starting a domain with a saved state image
Boot the guest, even if there was a saved snapshot
Validate the XML document against the XML schema
Reset the firmware NVRAM state from its original template
The following constants can be used to control the behaviour of domain define operations
The following constants define the set of supported keycode sets
The Linux event subsystem keycodes
The original XT keycodes
The AT Set1 keycodes (aka XT)
The AT Set2 keycodes (aka AT)
The AT Set3 keycodes (aka PS2)
The OS-X keycodes
The XT keycodes from the Linux Keyboard driver
The USB HID keycode set
The Windows keycode set
The XT keycode set, with the extended scancodes using the high bit of the first byte, instead of the low bit of the second byte.
A deprecated alias for Sys::Virt::Domain::KEYCODE_SET_QNUM
Sys::Virt::Domain::KEYCODE_SET_QNUM
The following constants can be used with the memory_peek method's flags parameter
memory_peek
Indicates that the offset is using virtual memory addressing.
Indicates that the offset is using physical memory addressing.
The following constants are useful when interpreting the virtual CPU run state
The virtual CPU is not online
The virtual CPU is executing code
The virtual CPU is waiting to be scheduled
The following constants are useful when interpreting the virtual CPU host placement
The virtual CPU placement is not available from this hypervisor
The following constants are used when opening a connection to the guest graphics server
Skip authentication of the client
The following constants are used when opening a connection to the guest console
Force opening of the console, disconnecting any other open session
Check if the console driver supports safe operations
The following constants are used when opening a connection to the guest channel
Force opening of the channel, disconnecting any other open session
The following constants are used to control the information included in the XML configuration dump
Report the persistent inactive configuration for the guest, even if it is currently running.
Update the CPU model definition to match the current executing state.
Update the XML to allow migration to older versions of libvirt
The following constants are used to control device hotplug operations
Modify the domain in its current state
Modify only the live state of the domain
Modify only the persistent config of the domain
Force the device to be modified
The following constants are used to control memory change operations
Modify the current state
Modify the maximum memory value
The following constants are used to control what configuration a domain update changes
The following constants are used to control how migration is performed
Migrate the guest without interrupting its execution on the source host.
Manage the migration process over a direct peer-2-peer connection between the source and destination host libvirtd daemons.
Tunnel the migration data over the libvirt daemon connection, rather than the native hypervisor data transport. Requires PEER2PEER flag to be set.
Make the domain persistent on the destination host, defining its configuration file upon completion of migration.
Remove the domain's persistent configuration after migration completes successfully.
Do not re-start execution of the guest CPUs on the destination host after migration completes.
Copy the complete contents of the disk images during migration
Copy the incrementally changed contents of the disk images during migration
Do not allow changes to the virtual domain configuration while migration is taking place. This option is automatically implied if doing a peer-2-peer migration.
Migrate even if the compatibility check indicates the migration will be unsafe to the guest.
Migrate the guest config if the guest is not currently running
Enable compression of the migration data stream
Abort if an I/O error occurs on the disk
Force convergence of the migration operation by throttling guest runtime
Pin memory for RDMA transfer
Enable support for post-copy migration
Setting this flag will cause the migration to attempt to use the TLS environment configured by the hypervisor in order to perform the migration. If incorrectly configured on either source or destination, the migration will fail.
Send memory pages to the destination host through several network connections. See Sys::Virt::Domain::MIGRATE_PARAM_PARALLEL_* parameters for configuring the parallel migration.
Sys::Virt::Domain::MIGRATE_PARAM_PARALLEL_*
Force the guest writes which happen when copying disk images for non-shared storage migration to be synchronously written to the destination. This ensures the storage migration converges for VMs doing heavy I/O on fast local storage and slow mirror.
Resume migration which failed in post-copy phase.
Attempt to avoid copying data. This is a request, not a guarantee and may apply to either one or both directions of data transfer at discretion of the hypervisor implementation.
The following constants can be used when undefining virtual domain configurations
Also remove any managed save image when undefining the virtual domain
Also remove any snapshot metadata when undefining the virtual domain.
Also remove any NVRAM state file when undefining the virtual domain.
keep NVRAM state file when undefining the virtual domain.
Also remove any checkpoint metadata when undefining the virtual domain.
Also remove any TPM state when undefining the virtual domain.
Keep TPM state when undefining the virtual domain.
The following constants describe the different background job types.
No job is active
A job with a finite completion time is active
A job with an unbounded completion time is active
The job has finished, but isn't cleaned up
The job has hit an error, but isn't cleaned up
The job was aborted at user request, but isn't cleaned up
The following constants are useful when getting/setting memory parameters for guests
The maximum memory the guest can use.
The memory upper limit enforced during memory contention.
The minimum memory guaranteed to be reserved for the guest.
The maximum swap the guest can use.
The value of an unlimited memory parameter
The following parameters control I/O tuning for the domain as a whole
The I/O weight parameter
The per-device I/O weight parameter
The per-device I/O bytes read per second
The per-device I/O operations read per second
The per-device I/O bytes write per second
The per-device I/O operations write per second
The following parameters control I/O tuning for an individual guest disk.
The total bytes processed per second.
The bytes read per second.
The bytes written per second.
The total I/O operations processed per second.
The I/O operations read per second.
The I/O operations written per second.
The maximum total bytes processed per second.
The maximum bytes read per second.
The maximum bytes written per second.
The maximum total I/O operations processed per second.
The maximum I/O operations read per second.
The maximum I/O operations written per second.
The maximum I/O operations per second
A string representing a group name to allow sharing of I/O throttling quota between multiple drives
The duration in seconds allowed for maximum total bytes processed per second.
The duration in seconds allowed for maximum bytes read per second.
The duration in seconds allowed for maximum bytes written per second.
The duration in seconds allowed for maximum total I/O operations processed per second.
The duration in seconds allowed for maximum I/O operations read per second.
The duration in seconds allowed for maximum I/O operations written per second.
The VM cap tunable
The CPU shares tunable
The VM limit tunable
The VM reservation tunable
The VM shares tunable
The VCPU period tunable
The VCPU quota tunable
The VM global period tunable
The VM global quota tunable
The VM weight tunable
The following constants are useful when getting/setting the guest NUMA memory policy
The NUMA policy mode
The NUMA nodeset mask
The following constants are useful when interpreting the Sys::Virt::Domain::NUMA_MODE parameter value
Sys::Virt::Domain::NUMA_MODE
Allocation is mandatory from the mask nodes
Allocation is preferred from the masked nodes
Allocation is interleaved across all masked nodes
Allocation is determined by the host using the masked nodes.
The following constants are useful when getting/setting the per network interface tunable parameters
The average inbound bandwidth
The peak inbound bandwidth
The burstable inbound bandwidth
The minimum inbound bandwidth
The average outbound bandwidth
The peak outbound bandwidth
The burstable outbound bandwidth
The following constants defined performance events which can be monitored for a guest
The CMT event counter which can be used to measure the usage of cache (bytes) by applications running on the platform. It corresponds to the "perf.cmt" field in the *Stats APIs.
The MBML event counter which can be used to monitor the amount of data (bytes/s) sent through the memory controller on the socket. It corresponds to the "perf.mbml" field in the *Stats APIs.
The MBMT event counter which can be used to monitor total system bandwidth (bytes/s) from one level of cache to another. It corresponds to the "perf.mbmt" field in the *Stats APIs.
The cache_misses perf event counter which can be used to measure the count of cache misses by applications running on the platform. It corresponds to the "perf.cache_misses" field in the *Stats APIs.
The cache_references perf event counter which can be used to measure the count of cache hits by applications running on the platform. It corresponds to the "perf.cache_references" field in the *Stats APIs.
The cpu_cycles perf event counter which can be used to measure how many cpu cycles one instruction needs. It corresponds to the "perf.cpu_cycles" field in the *Stats APIs.
The instructions perf event counter which can be used to measure the count of instructions by applications running on the platform. It corresponds to the "perf.instructions" field in the *Stats APIs.
The branch_instructions perf event counter which can be used to measure the count of instructions by applications running on the platform. It corresponds to the "perf.branch_instructions" field in the *Stats APIs.
The branch_misses perf event which can be used to measure the count of branch misses by applications running on the platform. It corresponds to the "perf.branch_misses" field in the *Stats APIs.
The bus_cycles perf event counter which can be used to measure the count of bus cycles by applications running on the platform. It corresponds to the "perf.bus_cycles" field in the *Stats APIs.
The stalled_cycles_frontend perf event counter which can be used to measure the count of stalled cpu cycles in the frontend of the instruction processor pipeline by applications running on the platform. It corresponds to the "perf.stalled_cycles_frontend" field in the *Stats APIs.
The stalled_cycles_backend perf event counter which can be used to measure the count of stalled cpu cycles in the backend of the instruction processor pipeline by application running on the platform. It corresponds to the "perf.stalled_cycles_backend" field in the *Stats APIs.
The ref_cpu_cycles perf event counter which can be used to measure the count of total cpu cycles not affected by CPU frequency scaling by applications running on the platform. It corresponds to the "perf.ref_cpu_cycles" field in the *Stats APIs.
The cpu_clock perf event counter which can be used to measure the count of cpu clock time by applications running on the platform. It corresponds to the "perf.cpu_clock" field in the *Stats APIs.
The task_clock perf event counter which can be used to measure the count of task clock time by applications running on the platform. It corresponds to the "perf.task_clock" field in the *Stats APIs.
The page_faults perf event counter which can be used to measure the count of page faults by applications running on the platform. It corresponds to the "perf.page_faults" field in the *Stats APIs.
The context_switches perf event counter which can be used to measure the count of context switches by applications running on the platform. It corresponds to the "perf.context_switches" field in the *Stats APIs.
The cpu_migrations perf event counter which can be used to measure the count of cpu migrations by applications running on the platform. It corresponds to the "perf.cpu_migrations" field in the *Stats APIs.
The page_faults_min perf event counter which can be used to measure the count of minor page faults by applications running on the platform. It corresponds to the "perf.page_faults_min" field in the *Stats APIs.
The page_faults_maj perf event counter which can be used to measure the count of major page faults by applications running on the platform. It corresponds to the "perf.page_faults_maj" field in the *Stats APIs.
The alignment_faults perf event counter which can be used to measure the count of alignment faults by applications running on the platform. It corresponds to the "perf.alignment_faults" field in the *Stats APIs.
The emulation_faults perf event counter which can be used to measure the count of emulation faults by applications running on the platform. It corresponds to the "perf.emulation_faults" field in the *Stats APIs.
The following constants defined IOThread statistics which can be monitored for a guest
The maximum polling time that can be used by polling algorithm in ns. The polling time starts at 0 (zero) and is the time spent by the guest to process IOThread data before returning the CPU to the host. The polling time will be dynamically modified over time based on the poll_grow and poll_shrink parameters provided.
This provides a value for the dynamic polling adjustment algorithm to use to grow its polling interval up to the poll_max_ns value.
This provides a value for the dynamic polling adjustment algorithm to use to shrink its polling interval when the polling interval exceeds the poll_max_ns value.
Sets the lower bound for thread pool size. A value of -1 disables this bound leaving hypervisor use its default value, though this value is not accepted for running domains. Due to internal implementation it's recommended to set lower and upper bounds separately.
Sets the upper bound for thread pool size. A value of -1 disables this bound leaving hypervisor use its default value, though this value is not accepted for running domains. Since the upper band has to be equal to or greater than lower bound value of 0 is not accepted. Due to internal implementation it's recommended to set lower and upper bounds separately.
The following constants are useful when getting/setting the VCPU count for a guest
Flag to request the live value
Flag to request the persistent config value
Flag to request the current config value
Flag to request adjustment of the maximum vCPU value
Flag to request the guest VCPU mask
Flag to make vcpus added hot(un)pluggable
The following constants allow domain state change events to be interpreted. The events contain both a state change, and a reason.
Indicates that a persistent configuration has been defined for the domain.
The defined configuration is newly added
The defined configuration is an update to an existing configuration
The defined configuration is a rename of an existing configuration
The defined configuration was restored from a snapshot
The domain has resumed execution
The domain resumed because migration has completed. This is emitted on the destination host.
The domain resumed because the admin unpaused it.
The domain resumed because it was restored from a snapshot
The domain resumed but post-copy is running in background
The domain is running, but migration failed in post-copy.
The domain has started running
The domain was booted from shutoff state
The domain started due to an incoming migration
The domain was restored from saved state file
The domain was restored from a snapshot
The domain was woken up from suspend
The domain has stopped running
The domain stopped because guest operating system has crashed
The domain stopped because administrator issued a destroy command.
The domain stopped because of a fault in the host virtualization environment.
The domain stopped because it was migrated to another machine.
The domain was saved to a state file
The domain stopped due to graceful shutdown of the guest.
The domain was stopped due to a snapshot
The domain has shutdown but is not yet stopped
The domain finished shutting down
The domain shutdown due to host trigger
The domain shutdown due to guest trigger
The domain has stopped executing, but still exists
The domain has been suspended due to offline migration
The domain has been suspended due to administrator pause request.
The domain has been suspended due to a block device I/O error.
The domain has been suspended due to resume from snapshot
The domain has been suspended due to the watchdog triggering
The domain has been suspended due to restore from saved state
The domain has been suspended due to an API error
The domain has been suspended for post-copy migration
The domain has been suspended due post-copy migration failing
The persistent configuration has gone away
The domain configuration has gone away due to it being removed by administrator.
The undefined configuration is a rename of an existing configuration
The domain has suspend to RAM.
The domain has suspend to Disk.
The domain has crashed
The domain has crashed and reloaded itself
Domain lifecycle events
Soft / warm reboot events
RTC clock adjustments
File IO errors, typically from disks
Watchdog device triggering
Graphics client connections.
File IO errors, typically from disks, with a root cause
Errors from the virtualization control channel
Completion status of asynchronous block jobs, identified by source file name.
Completion status of asynchronous block jobs, identified by target device name.
Changes in disk media
CDROM media tray state
Power management initiated suspend to RAM
Power management initiated suspend to Disk
Power management initiated wakeup
Balloon target changes
Asynchronous guest device addition
Asynchronous guest device removal
Changes of any domain tuning parameters. The callback will be provided with a hash listing all changed parameters. The later DOMAIN TUNABLE constants can be useful when accessing the hash keys
Domain guest agent lifecycle events. The state parameter to the callback will match one of the constants
The agent is now connected
The agent is now disconnected
The second parameter, reason, matches one of the following constants
reason
Domain migration progress iteration. The iteration parameter to the callback will specify the number of iterations migration has made over guest RAM.
iteration
The reason is unknown
The domain was initially booted
The channel on a running guest changed state
Domain background job completion notification. The callback provides a hash containing the job stats. The keys in the hash are the same as those used with the Sys::Virt::Domain::get_job_stats() method.
Sys::Virt::Domain::get_job_stats()
Guest device removal has failed.
The domain metadata has changed
The event occurs when the hypervisor detects that the given storage element was written beyond the point specified by threshold. The event is useful for thin-provisioned storage.
The event occurs when the hypervisor detects hardware memory corruption.
The event occurs when the guest accepts a request to change the memory device size.
These constants describe what action was taken due to the IO error.
No action was taken, the error was ignored & reported as success to guest
The guest is paused since the error occurred
The error has been reported to the guest OS
These constants describe what action was taken due to the watchdog firing
No action was taken, the watchdog was ignored
The guest is paused since the watchdog fired
The guest is powered off after the watchdog fired
The guest is reset after the watchdog fired
The guest attempted to gracefully shutdown after the watchdog fired
No action was taken, the watchdog was logged
An NMI was injected into the guest after the watchdog fired
These constants describe the phase of the graphics connection
The initial client connection
The client has been authenticated & the connection is running
The client has disconnected
These constants describe the format of the address
An IPv4 address
An IPv6 address
An UNIX socket path address
These constants describe the reason for a disk change event
The disk media was cleared, as its source was missing when attempting to start the guest
The disk device was dropped, as its source was missing when attempting to start the guest
These constants describe the reason for a tray change event
The tray was closed
The tray was opened
The following constants identify the different types of domain block jobs
An unknown block job type
The block pull job type
The block copy job type
The block commit job type
The block active commit job type
The block backup job type
The following constants can be used to determine the completion status of a block job
A successfully completed block job
An unsuccessful block job
A block job canceled by the user
A block job is running
The following constants are useful when rebasing block devices
Limit copy to top of source backing chain
Reuse existing external file for copy
Make destination file raw
Start a copy job
Treat destination as a block device instead of file
Keep backing chain referenced using relative names
The following constants are useful when copying block devices
Don't force usage of recoverable job for the copy operation
Force the copy job to synchronously propagate guest writes into the destination image, so that the copy is guaranteed to converge
The following constants are useful when aborting job copy jobs
Request only, do not wait for completion
Pivot to mirror when ending a copy job
The following constants are useful with block commit job types
Delete any files that are invalid after commit
NULL base means next backing file, not whole chain
Allow two phase commit when top is active layer
The following constants can be used when saving or restoring virtual machines
Do not use OS I/O cache when saving state.
Mark the saved state as paused to prevent the guest CPUs starting upon restore.
Mark the saved state as running to allow the guest CPUs to start upon restore.
Specify the save state file to save to or restore from.
Used to adjust guest xml on restore, e.g. to alter a device while domain is stopped.
The following constants can be used when triggering domain core dumps
Do not pause execution while dumping the guest
Crash the guest after completing the core dump
Do not use OS I/O cache when writing core dump
Reset the virtual machine after finishing the dump
Only include guest RAM in the dump, not the device state
The following constants are useful when terminating guests using the destroy API.
Destroy the guest using the default approach
Destroy the guest in a graceful manner
Delete log files associated with the guest
The following constants are useful when requesting that a guest terminate using the shutdown API
Shutdown using the hypervisor's default mechanism
Shutdown by issuing a command to a guest agent
Shutdown by injecting an ACPI power button press
Shutdown by talking to initctl (containers only)
Shutdown by sending SIGTERM to the init process
Shutdown by issuing a paravirt power control command
The following constants are useful when requesting that a guest terminate using the reboot API
reboot
Reboot using the hypervisor's default mechanism
Reboot by issuing a command to a guest agent
Reboot by injecting an ACPI power button press
Reboot by talking to initctl (containers only)
Reboot by sending SIGHUP to the init process
Reboot by issuing a paravirt power control command
The following constants are useful when reading/writing metadata about a guest
The short human friendly title of the guest
The long free text description of the guest
The structured metadata elements for the guest
The following constants are useful when interpreting disk error codes
No error
The host storage has run out of free space
An unspecified error has occurred.
Swap in
Swap out
Minor faults
Major faults
Resident memory
Unused memory
Available memory
Actual balloon limit
Amount of usable memory
Time of last stats refresh from guest
Disk cache size
The amount of successful huge page allocations
The amount of failed huge page allocations
The following constants can be used when listing domains
Only list domains that are currently active (running, or paused)
Only list domains that are set to automatically start on boot
Only list domains that have a stored snapshot
Only list domains that are currently inactive (shutoff, saved)
Only list domains that have current managed save state
Only list domains that are not set to automatically start on boto
Only list domains that do not have any managed save state
Only list domains that do not have a stored snapshot
Only list domains that are not running, paused or shutoff
Only list domains that are paused
Only list domains which have a persistent config
Only list domains that are currently running
Only list domains that are currently shutoff
Only list domains that do not have a persistent config
Only list domains that have a stored checkpoint
Only list domains that do not have a stored checkpoint
The following constants are to be used with the send_key API
send_key
The maximum number of keys that can be sent in a single call to send_key
The following constants provide the names of well known block stats fields
The number of I/O errors
The number of flush requests
The time spent processing flush requests
The amount of data read
The number of read requests
The time spent processing read requests
The amount of data written
The number of write requests
The time spent processing write requests
The following constants provide the names of well known cpu stats fields
The total CPU time, including both hypervisor and vCPU time.
THe total time in kernel
The total time in userspace
The total vCPU time.
The following constants provide the names of well known schedular parameters
The duration of the time period for scheduling the emulator
The quota for the emulator in one schedular time period
The duration of the time period for scheduling the iothread
The quota for the iothread in one schedular time period
The following constants are used as flags when requesting bulk domain stats from Sys::Virt::get_all_domain_stats.
Sys::Virt::get_all_domain_stats
Include stats for active domains
Include stats for inactive domains
Include stats for other domains
Include stats for paused domains
Include stats for persistent domains
Include stats for running domains
Include stats for shutoff domains
Include stats for transient domains
Require that all requested stats fields are returned
Get stats for image backing files too
Skip stats if they can't be acquired without waiting
The following constants are used to control which fields are returned for stats queries.
Balloon statistics
Block device info
CPU usage info
Network interface info
General lifecycle state
Virtual CPU info
Performance event counter values
IOThread performance statistics values
Memory bandwidth statistics values
Memory dirty rate statistics
Hypervisor-specific stats fields for given VM
The following constants provide the names of signals which can be sent to guest processes. They mostly correspond to POSIX signal names.
SIGNOP
SIGHUP
SIGINT
SIGQUIT
SIGILL
SIGTRAP
SIGABRT
SIGBUS
SIGFPE
SIGKILL
SIGUSR1
SIGSEGV
SIGUSR2
SIGPIPE
SIGALRM
SIGTERM
SIGSTKFLT
SIGCHLD
SIGCONT
SIGSTOP
SIGTSTP
SIGTTIN
SIGTTOU
SIGURG
SIGXCPU
SIGXFSZ
SIGVTALRM
SIGPROF
SIGWINCH
SIGPOLL
SIGPWR
SIGSYS
SIGRT0
SIGRT1
SIGRT2
SIGRT3
SIGRT4
SIGRT5
SIGRT6
SIGRT7
SIGRT8
SIGRT9
SIGRT10
SIGRT11
SIGRT12
SIGRT13
SIGRT14
SIGRT15
SIGRT16
SIGRT17
SIGRT18
SIGRT19
SIGRT20
SIGRT21
SIGRT22
SIGRT23
SIGRT24
SIGRT25
SIGRT26
SIGRT27
SIGRT28
SIGRT29
SIGRT30
SIGRT31
SIGRT32
The following constants are useful when accessing domain tuning parameters in APIs and events
Proportional CPU weight
Emulator thread CPU pinning mask
Emulator thread CPU period
Emulator thread CPU quota
Iothread thread CPU period
Iothread thread CPU quota
VCPU thread pinning mask
VCPU thread period
VCPU thread quota
VM global period
VM global quota
The name of guest disks
Read throughput in bytes per sec
Read throughput in I/O operations per sec
Total throughput in bytes per sec
Total throughput in I/O operations per sec
Write throughput in bytes per sec
Write throughput in I/O operations per sec
Maximum read throughput in bytes per sec
Maximum read throughput in I/O operations per sec
Maximum total throughput in bytes per sec
Maximum total throughput in I/O operations per sec
Maximum write throughput in bytes per sec
Maximum write throughput in I/O operations per sec
The name of the blkdev group
The I/O threads pinning
The following constants are useful when setting action for lifecycle events.
The poweroff lifecycle event type
The reboot lifecycle event type
The crash lifecycle event type
The destroy lifecycle action
The restart lifecycle action
The restart-rename lifecycle action
The preserve lifecycle action
The coredump-destroy lifecycle action
The coredump-restart lifecycle action
The failure could be ignored
An MCE was injected to the guest
The failure is non-recoverable and the hypervisor was not able to handle it
The failure is non-recoverable and the guest was not able to handle it.
The memory failure was in hypervisor address space
The memory failure was in guest address space
Whether the flag is action-required or action-optional
The failure occurred while the previous fault was being handled.
The dirty rate is not being measured currently.
The dity rate is in the process of being measured
The dirty rate has been measured
Daniel P. Berrange <berrange@redhat.com>
Copyright (C) 2006 Red Hat Copyright (C) 2006-2007 Daniel P. Berrange
This program is free software; you can redistribute it and/or modify it under the terms of either the GNU General Public License as published by the Free Software Foundation (either version 2 of the License, or at your option any later version), or, the Artistic License, as specified in the Perl README file.
Sys::Virt, Sys::Virt::Error, http://libvirt.org
http://libvirt.org
To install Sys::Virt, copy and paste the appropriate command in to your terminal.
cpanm
cpanm Sys::Virt
CPAN shell
perl -MCPAN -e shell install Sys::Virt
For more information on module installation, please visit the detailed CPAN module installation guide.