qemu/qapi/job.json

277 lines
7.3 KiB
JSON
Raw Normal View History

# -*- Mode: Python -*-
# vim: filetype=python
##
qapi storage-daemon/qapi: Fix documentation section structure In the QEMU QMP Reference Manual, subsection "Block core (VM unrelated)" is empty. Its contents is at the end of subsection "Background jobs" instead. That's because qapi/job.json is included first from qapi/block-core.json, which makes qapi/job.json's documentation go between qapi/block-core.json's subsection heading and contents. In the QEMU Storage Daemon QMP Reference Manual, section "Block Devices" contains nothing but an empty subsection "Block core (VM unrelated)". The latter's contents is at the end section "Socket data types", along with subsection "Block device exports". Subsection "Background jobs" is at the end of section "Cryptography". All this is because storage-daemon/qapi/qapi-schema.json includes modules in a confused order. Fix both as follows. Turn subsection "Background jobs" into a section. Move it before section "Block devices" in the QEMU QMP Reference Manual, by including qapi/jobs.json right before qapi/block.json. Reorder include directives in storage-daemon/qapi/qapi-schema.json to match the order in qapi/qapi-schema.json, so that the QEMU Storage Daemon QMP Reference Manual's section structure the QEMU QMP Reference Manual's. In the QEMU QMP Reference Manual, qapi/cryptodev.json's documentation is at the end of section "Virtio devices". That's because it lacks a section heading, and therefore gets squashed into whatever section happens to precede it. Add section heading so it's in section "Cryptography devices". Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru> Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Acked-by: zhenwei pi <pizhenwei@bytedance.com> Message-Id: <20230425064223.820979-17-armbru@redhat.com>
2023-04-25 09:42:23 +03:00
# = Background jobs
##
##
# @JobType:
#
# Type of a background job.
#
# @commit: block commit job type, see "block-commit"
#
# @stream: block stream job type, see "block-stream"
#
# @mirror: drive mirror job type, see "drive-mirror"
#
# @backup: drive backup job type, see "drive-backup"
#
# @create: image creation job type, see "blockdev-create" (since 3.0)
#
# @amend: image options amend job type, see "x-blockdev-amend" (since
# 5.1)
#
# @snapshot-load: snapshot load job type, see "snapshot-load" (since
# 6.0)
migration: introduce snapshot-{save, load, delete} QMP commands savevm, loadvm and delvm are some of the few HMP commands that have never been converted to use QMP. The reasons for the lack of conversion are that they blocked execution of the event thread, and the semantics around choice of disks were ill-defined. Despite this downside, however, libvirt and applications using libvirt have used these commands for as long as QMP has existed, via the "human-monitor-command" passthrough command. IOW, while it is clearly desirable to be able to fix the problems, they are not a blocker to all real world usage. Meanwhile there is a need for other features which involve adding new parameters to the commands. This is possible with HMP passthrough, but it provides no reliable way for apps to introspect features, so using QAPI modelling is highly desirable. This patch thus introduces new snapshot-{load,save,delete} commands to QMP that are intended to replace the old HMP counterparts. The new commands are given different names, because they will be using the new QEMU job framework and thus will have diverging behaviour from the HMP originals. It would thus be misleading to keep the same name. While this design uses the generic job framework, the current impl is still blocking. The intention that the blocking problem is fixed later. None the less applications using these new commands should assume that they are asynchronous and thus wait for the job status change event to indicate completion. In addition to using the job framework, the new commands require the caller to be explicit about all the block device nodes used in the snapshot operations, with no built-in default heuristics in use. Note that the existing "query-named-block-nodes" can be used to query what snapshots currently exist for block nodes. Acked-by: Markus Armbruster <armbru@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> Message-Id: <20210204124834.774401-13-berrange@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com> dgilbert: removed tests for now, the output ordering isn't deterministic
2021-02-04 15:48:34 +03:00
#
# @snapshot-save: snapshot save job type, see "snapshot-save" (since
# 6.0)
migration: introduce snapshot-{save, load, delete} QMP commands savevm, loadvm and delvm are some of the few HMP commands that have never been converted to use QMP. The reasons for the lack of conversion are that they blocked execution of the event thread, and the semantics around choice of disks were ill-defined. Despite this downside, however, libvirt and applications using libvirt have used these commands for as long as QMP has existed, via the "human-monitor-command" passthrough command. IOW, while it is clearly desirable to be able to fix the problems, they are not a blocker to all real world usage. Meanwhile there is a need for other features which involve adding new parameters to the commands. This is possible with HMP passthrough, but it provides no reliable way for apps to introspect features, so using QAPI modelling is highly desirable. This patch thus introduces new snapshot-{load,save,delete} commands to QMP that are intended to replace the old HMP counterparts. The new commands are given different names, because they will be using the new QEMU job framework and thus will have diverging behaviour from the HMP originals. It would thus be misleading to keep the same name. While this design uses the generic job framework, the current impl is still blocking. The intention that the blocking problem is fixed later. None the less applications using these new commands should assume that they are asynchronous and thus wait for the job status change event to indicate completion. In addition to using the job framework, the new commands require the caller to be explicit about all the block device nodes used in the snapshot operations, with no built-in default heuristics in use. Note that the existing "query-named-block-nodes" can be used to query what snapshots currently exist for block nodes. Acked-by: Markus Armbruster <armbru@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> Message-Id: <20210204124834.774401-13-berrange@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com> dgilbert: removed tests for now, the output ordering isn't deterministic
2021-02-04 15:48:34 +03:00
#
# @snapshot-delete: snapshot delete job type, see "snapshot-delete"
# (since 6.0)
migration: introduce snapshot-{save, load, delete} QMP commands savevm, loadvm and delvm are some of the few HMP commands that have never been converted to use QMP. The reasons for the lack of conversion are that they blocked execution of the event thread, and the semantics around choice of disks were ill-defined. Despite this downside, however, libvirt and applications using libvirt have used these commands for as long as QMP has existed, via the "human-monitor-command" passthrough command. IOW, while it is clearly desirable to be able to fix the problems, they are not a blocker to all real world usage. Meanwhile there is a need for other features which involve adding new parameters to the commands. This is possible with HMP passthrough, but it provides no reliable way for apps to introspect features, so using QAPI modelling is highly desirable. This patch thus introduces new snapshot-{load,save,delete} commands to QMP that are intended to replace the old HMP counterparts. The new commands are given different names, because they will be using the new QEMU job framework and thus will have diverging behaviour from the HMP originals. It would thus be misleading to keep the same name. While this design uses the generic job framework, the current impl is still blocking. The intention that the blocking problem is fixed later. None the less applications using these new commands should assume that they are asynchronous and thus wait for the job status change event to indicate completion. In addition to using the job framework, the new commands require the caller to be explicit about all the block device nodes used in the snapshot operations, with no built-in default heuristics in use. Note that the existing "query-named-block-nodes" can be used to query what snapshots currently exist for block nodes. Acked-by: Markus Armbruster <armbru@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> Message-Id: <20210204124834.774401-13-berrange@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com> dgilbert: removed tests for now, the output ordering isn't deterministic
2021-02-04 15:48:34 +03:00
#
# Since: 1.7
##
{ 'enum': 'JobType',
migration: introduce snapshot-{save, load, delete} QMP commands savevm, loadvm and delvm are some of the few HMP commands that have never been converted to use QMP. The reasons for the lack of conversion are that they blocked execution of the event thread, and the semantics around choice of disks were ill-defined. Despite this downside, however, libvirt and applications using libvirt have used these commands for as long as QMP has existed, via the "human-monitor-command" passthrough command. IOW, while it is clearly desirable to be able to fix the problems, they are not a blocker to all real world usage. Meanwhile there is a need for other features which involve adding new parameters to the commands. This is possible with HMP passthrough, but it provides no reliable way for apps to introspect features, so using QAPI modelling is highly desirable. This patch thus introduces new snapshot-{load,save,delete} commands to QMP that are intended to replace the old HMP counterparts. The new commands are given different names, because they will be using the new QEMU job framework and thus will have diverging behaviour from the HMP originals. It would thus be misleading to keep the same name. While this design uses the generic job framework, the current impl is still blocking. The intention that the blocking problem is fixed later. None the less applications using these new commands should assume that they are asynchronous and thus wait for the job status change event to indicate completion. In addition to using the job framework, the new commands require the caller to be explicit about all the block device nodes used in the snapshot operations, with no built-in default heuristics in use. Note that the existing "query-named-block-nodes" can be used to query what snapshots currently exist for block nodes. Acked-by: Markus Armbruster <armbru@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> Message-Id: <20210204124834.774401-13-berrange@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com> dgilbert: removed tests for now, the output ordering isn't deterministic
2021-02-04 15:48:34 +03:00
'data': ['commit', 'stream', 'mirror', 'backup', 'create', 'amend',
'snapshot-load', 'snapshot-save', 'snapshot-delete'] }
##
# @JobStatus:
#
# Indicates the present state of a given job in its lifetime.
#
# @undefined: Erroneous, default state. Should not ever be visible.
#
# @created: The job has been created, but not yet started.
#
# @running: The job is currently running.
#
# @paused: The job is running, but paused. The pause may be requested
# by either the QMP user or by internal processes.
#
# @ready: The job is running, but is ready for the user to signal
# completion. This is used for long-running jobs like mirror that
# are designed to run indefinitely.
#
# @standby: The job is ready, but paused. This is nearly identical to
# @paused. The job may return to @ready or otherwise be canceled.
#
# @waiting: The job is waiting for other jobs in the transaction to
# converge to the waiting state. This status will likely not be
# visible for the last job in a transaction.
#
# @pending: The job has finished its work, but has finalization steps
# that it needs to make prior to completing. These changes will
# require manual intervention via @job-finalize if auto-finalize
# was set to false. These pending changes may still fail.
#
# @aborting: The job is in the process of being aborted, and will
# finish with an error. The job will afterwards report that it is
# @concluded. This status may not be visible to the management
# process.
#
# @concluded: The job has finished all work. If auto-dismiss was set
# to false, the job will remain in the query list until it is
# dismissed via @job-dismiss.
#
# @null: The job is in the process of being dismantled. This state
# should not ever be visible externally.
#
# Since: 2.12
##
{ 'enum': 'JobStatus',
'data': ['undefined', 'created', 'running', 'paused', 'ready', 'standby',
'waiting', 'pending', 'aborting', 'concluded', 'null' ] }
##
# @JobVerb:
#
# Represents command verbs that can be applied to a job.
#
# @cancel: see @job-cancel
#
# @pause: see @job-pause
#
# @resume: see @job-resume
#
# @set-speed: see @block-job-set-speed
#
# @complete: see @job-complete
#
# @dismiss: see @job-dismiss
#
# @finalize: see @job-finalize
#
# @change: see @block-job-change (since 8.2)
#
# Since: 2.12
##
{ 'enum': 'JobVerb',
'data': ['cancel', 'pause', 'resume', 'set-speed', 'complete', 'dismiss',
'finalize', 'change' ] }
##
# @JOB_STATUS_CHANGE:
#
# Emitted when a job transitions to a different status.
#
# @id: The job identifier
#
# @status: The new job status
#
# Since: 3.0
##
{ 'event': 'JOB_STATUS_CHANGE',
'data': { 'id': 'str',
'status': 'JobStatus' } }
##
# @job-pause:
#
# Pause an active job.
#
# This command returns immediately after marking the active job for
# pausing. Pausing an already paused job is an error.
#
# The job will pause as soon as possible, which means transitioning
# into the PAUSED state if it was RUNNING, or into STANDBY if it was
# READY. The corresponding JOB_STATUS_CHANGE event will be emitted.
#
# Cancelling a paused job automatically resumes it.
#
# @id: The job identifier.
#
# Since: 3.0
##
{ 'command': 'job-pause', 'data': { 'id': 'str' } }
##
# @job-resume:
#
# Resume a paused job.
#
# This command returns immediately after resuming a paused job.
# Resuming an already running job is an error.
#
# @id: The job identifier.
#
# Since: 3.0
##
{ 'command': 'job-resume', 'data': { 'id': 'str' } }
##
# @job-cancel:
#
# Instruct an active background job to cancel at the next opportunity.
# This command returns immediately after marking the active job for
# cancellation.
#
# The job will cancel as soon as possible and then emit a
# JOB_STATUS_CHANGE event. Usually, the status will change to
# ABORTING, but it is possible that a job successfully completes (e.g.
# because it was almost done and there was no opportunity to cancel
# earlier than completing the job) and transitions to PENDING instead.
#
# @id: The job identifier.
#
# Since: 3.0
##
{ 'command': 'job-cancel', 'data': { 'id': 'str' } }
##
# @job-complete:
#
# Manually trigger completion of an active job in the READY state.
#
# @id: The job identifier.
#
# Since: 3.0
##
{ 'command': 'job-complete', 'data': { 'id': 'str' } }
##
# @job-dismiss:
#
# Deletes a job that is in the CONCLUDED state. This command only
# needs to be run explicitly for jobs that don't have automatic
# dismiss enabled.
#
# This command will refuse to operate on any job that has not yet
# reached its terminal state, JOB_STATUS_CONCLUDED. For jobs that make
# use of JOB_READY event, job-cancel or job-complete will still need
# to be used as appropriate.
#
# @id: The job identifier.
#
# Since: 3.0
##
{ 'command': 'job-dismiss', 'data': { 'id': 'str' } }
##
# @job-finalize:
#
# Instructs all jobs in a transaction (or a single job if it is not
# part of any transaction) to finalize any graph changes and do any
# necessary cleanup. This command requires that all involved jobs are
# in the PENDING state.
#
# For jobs in a transaction, instructing one job to finalize will
# force ALL jobs in the transaction to finalize, so it is only
# necessary to instruct a single member job to finalize.
#
# @id: The identifier of any job in the transaction, or of a job that
# is not part of any transaction.
#
# Since: 3.0
##
{ 'command': 'job-finalize', 'data': { 'id': 'str' } }
##
# @JobInfo:
#
# Information about a job.
#
# @id: The job identifier
#
# @type: The kind of job that is being performed
#
# @status: Current job state/status
#
# @current-progress: Progress made until now. The unit is arbitrary
# and the value can only meaningfully be used for the ratio of
# @current-progress to @total-progress. The value is
# monotonically increasing.
#
# @total-progress: Estimated @current-progress value at the completion
# of the job. This value can arbitrarily change while the job is
# running, in both directions.
#
# @error: If this field is present, the job failed; if it is still
# missing in the CONCLUDED state, this indicates successful
# completion.
#
# The value is a human-readable error message to describe the
# reason for the job failure. It should not be parsed by
# applications.
#
# Since: 3.0
##
{ 'struct': 'JobInfo',
'data': { 'id': 'str', 'type': 'JobType', 'status': 'JobStatus',
'current-progress': 'int', 'total-progress': 'int',
'*error': 'str' } }
##
# @query-jobs:
#
# Return information about jobs.
#
# Returns: a list with a @JobInfo for each active job
#
# Since: 3.0
##
{ 'command': 'query-jobs', 'returns': ['JobInfo'] }