package mesosproto

Get desktop application:
View/edit binary Protocol Buffers messages

Represents an "action" performed on the log. Each action has an associated position in the log. In addition, each action (i.e., position) will have been "promised" to a specific proposer (implicitly or explicitly) and may have been "performed" from a specific proposer. An action may also be "learned" to have reached consensus. There are three types of possible actions that can be performed on the log: nop (no action), append, and truncate.

Used in: LearnedMessage, PromiseResponse, Record

• required uint64 position = 1

• required uint64 promised = 2

• optional uint64 performed = 3

• optional bool learned = 4

• optional Action.Type type = 5

  Set iff performed is set.

• optional Action.Nop nop = 6

• optional Action.Append append = 7

• optional Action.Truncate truncate = 8

Used in: Action, WriteRequest

• required bytes bytes = 1

• optional bytes cksum = 2

Used in: Action, WriteRequest

(message has no fields)

Used in: Action, WriteRequest

• required uint64 to = 1

  All positions before and exclusive of 'to'.

Used in: Action, WriteRequest

• NOP = 1

• APPEND = 2

• TRUNCATE = 3

* Describes Completed Frameworks, etc. for archival.

• repeated Archive.Framework frameworks = 1

Used in: Archive, ReregisterSlaveMessage

• required FrameworkInfo framework_info = 1

• optional string pid = 2

• repeated Task tasks = 3

* Describes an attribute that can be set on a machine. For now, attributes and resources share the same "value" type, but this may change in the future and attributes may only be string based.

Used in: Offer, SlaveInfo

• required string name = 1

• required Value.Type type = 2

• optional Value.Scalar scalar = 3

• optional Value.Ranges ranges = 4

• optional Value.Set set = 6

• optional Value.Text text = 5

• required string pid = 1

  PID that needs to be authenticated.

(message has no fields)

• optional string error = 1

(message has no fields)

• repeated string mechanisms = 1

  List of available SASL mechanisms.

• required string mechanism = 1

• optional string data = 2

• required bytes data = 1

* Describes a command, executed via: '/bin/sh -c value'. Any URIs specified are fetched before executing the command. If the executable field for an uri is set, executable file permission is set on the downloaded file. Otherwise, if the downloaded file has a recognized archive extension (currently [compressed] tar and zip) it is extracted into the executor's working directory. In addition, any environment variables are set before executing the command (so they can be used to "parameterize" your command).

Used in: ExecutorInfo, TaskInfo

• repeated CommandInfo.URI uris = 1

• optional Environment environment = 2

• required string value = 3

Used in: CommandInfo

• required string value = 1

• optional bool executable = 2

* A slave generated ID to distinguish a container. The ID must be unique between any active or completed containers on the slave. In particular, containers for different runs of the same (framework, executor) pair must be unique.

• required string value = 1

* Credential used for authentication. NOTE: The 'principal' is used for authenticating the framework with the master. This is different from 'FrameworkInfo.user' which is used to determine the user under which the framework's executors/tasks are run.

• required string principal = 1

• optional bytes secret = 2

• required FrameworkID framework_id = 1

* Describes a collection of environment variables. This is used with CommandInfo in order to set environment variables before running a command.

Used in: CommandInfo

• repeated Environment.Variable variables = 1

Used in: Environment

• required string name = 1

• required string value = 2

* A framework generated ID to distinguish an executor. Only one executor with the same ID can be active on the same slave at a time.

Used in: ExecutorInfo, ExecutorToFrameworkMessage, ExitedExecutorMessage, FrameworkToExecutorMessage, Offer, RegisterExecutorMessage, ReregisterExecutorMessage, ResourceUsage, StatusUpdate, Task

• required string value = 1

* Describes information about an executor. The 'data' field can be used to pass arbitrary bytes to an executor.

Used in: ExecutorRegisteredMessage, ReregisterSlaveMessage, TaskInfo

• required ExecutorID executor_id = 1

• optional FrameworkID framework_id = 8

  TODO(benh): Make this required.

• required CommandInfo command = 7

• repeated Resource resources = 5

• optional string name = 9

• optional string source = 10

  Source is an identifier style string used by frameworks to track the source of an executor. This is useful when it's possible for different executor ids to be related semantically. NOTE: Source is exposed alongside the resource usage of the executor via JSON on the slave. This allows users to import usage information into a time series database for monitoring.

• optional bytes data = 4

• required ExecutorInfo executor_info = 2

• required FrameworkID framework_id = 3

• required FrameworkInfo framework_info = 4

• required SlaveID slave_id = 5

• required SlaveInfo slave_info = 6

• required SlaveID slave_id = 1

• required SlaveInfo slave_info = 2

• required SlaveID slave_id = 1

• required FrameworkID framework_id = 2

• required ExecutorID executor_id = 3

• required bytes data = 4

• required SlaveID slave_id = 1

• required FrameworkID framework_id = 2

• required ExecutorID executor_id = 3

• required int32 status = 4

* Describes possible filters that can be applied to unused resources (see SchedulerDriver::launchTasks) to influence the allocator.

Used in: LaunchTasksMessage

• optional double refuse_seconds = 1

  Time to consider unused resources refused. Note that all unused resources will be considered refused and use the default value (below) regardless of whether Filters was passed to SchedulerDriver::launchTasks. You MUST pass Filters with this field set to change this behavior (i.e., get another offer which includes unused resources sooner or later than the default).

• required string message = 2

• required FrameworkID framework_id = 1

* A unique ID assigned to a framework. A framework can reuse this ID in order to do failover (see MesosSchedulerDriver).

Used in: DeactivateFrameworkMessage, ExecutorInfo, ExecutorRegisteredMessage, ExecutorToFrameworkMessage, ExitedExecutorMessage, FrameworkExpiredMessage, FrameworkInfo, FrameworkRegisteredMessage, FrameworkReregisteredMessage, FrameworkToExecutorMessage, KillTaskMessage, LaunchTasksMessage, Offer, ReconcileTasksMessage, RegisterExecutorMessage, ReregisterExecutorMessage, ResourceRequestMessage, ResourceUsage, ReviveOffersMessage, RunTaskMessage, ShutdownFrameworkMessage, StatusUpdate, StatusUpdateAcknowledgementMessage, Task, UnregisterFrameworkMessage, UpdateFrameworkMessage

• required string value = 1

* Describes a framework. If the user field is set to an empty string Mesos will automagically set it to the current user. Note that the ID is only available after a framework has registered, however, it is included here in order to facilitate scheduler failover (i.e., if it is set then the MesosSchedulerDriver expects the scheduler is performing failover). The amount of time that the master will wait for the scheduler to failover before removing the framework is specified by failover_timeout. If checkpoint is set, framework pid, executor pids and status updates are checkpointed to disk by the slaves. Checkpointing allows a restarted slave to reconnect with old executors and recover status updates, at the cost of disk I/O. The role field is used to group frameworks for allocation decisions, depending on the allocation policy being used. If the hostname field is set to an empty string Mesos will automagically set it to the current hostname.

Used in: Archive.Framework, ExecutorRegisteredMessage, RegisterFrameworkMessage, ReregisterFrameworkMessage, RunTaskMessage

• required string user = 1

• required string name = 2

• optional FrameworkID id = 3

• optional double failover_timeout = 4

• optional bool checkpoint = 5

• optional string role = 6

• optional string hostname = 7

• required FrameworkID framework_id = 1

• required MasterInfo master_info = 2

• required FrameworkID framework_id = 1

• required MasterInfo master_info = 2

• required SlaveID slave_id = 1

• required FrameworkID framework_id = 2

• required ExecutorID executor_id = 3

• required bytes data = 4

• required SlaveID slave_id = 1

• required FrameworkID framework_id = 1

  TODO(bmahler): Include the SlaveID here to improve the Master's ability to respond for non-activated slaves.

• required TaskID task_id = 2

• required FrameworkID framework_id = 1

• optional OfferID offer_id = 2

  Deprecated.

• repeated TaskInfo tasks = 3

• required Filters filters = 5

• repeated OfferID offer_ids = 6

Represents a "learned" event, that is, when a particular action has been agreed upon (reached consensus).

• required Action action = 1

• required SlaveID slave_id = 1

* Describes a master. This will probably have more fields in the future which might be used, for example, to link a framework webui to a master webui.

Used in: FrameworkRegisteredMessage, FrameworkReregisteredMessage

• required string id = 1

• required uint32 ip = 2

• required uint32 port = 3

• optional string pid = 4

• optional string hostname = 5

The metadata of a replica. It has to be persisted on the disk. We store the current status of the replica as well as the implicit promise that a replica has made. This message is intended to replace the old Promise message to support catch-up.

Used in: Record

• required Metadata.Status status = 1

• required uint64 promised = 2

Used in: Metadata, RecoverResponse

• VOTING = 1

  Normal voting member in Paxos group.

• RECOVERING = 2

  In the process of catching up.

• STARTING = 3

  The log has been initialized.

• EMPTY = 4

  The log is empty and is not initialized.

* Describes some resources available on a slave. An offer only contains resources from a single slave.

Used in: ResourceOffersMessage

• required OfferID id = 1

• required FrameworkID framework_id = 2

• required SlaveID slave_id = 3

• required string hostname = 4

• repeated Resource resources = 5

• repeated Attribute attributes = 7

• repeated ExecutorID executor_ids = 6

* A unique ID assigned to an offer.

Used in: LaunchTasksMessage, Offer, RescindResourceOfferMessage

• required string value = 1

* A generic (key, value) pair used in various places for parameters.

Used in: Parameters

• required string key = 1

• required string value = 2

* Collection of Parameter.

Used in: ProjdUpdateResourcesMessage

• repeated Parameter parameter = 1

• required string project = 1

• optional Parameters parameters = 1

Represents a "promise" that a replica has made. A promise is *implicitly* valid for _all_ future actions that get performed on the replicated log (provided the action comes from the same proposer), until a new promise is made to a proposer with a higher proposal number. Each replica writes every promise it makes as a log record so that it can recover this information after a failure. TODO(benh): Does the promise actually need to be written to stable storage? Can we get away with looking at the last written action and using it's promised value? In this case, what happens if we make a promise but don't receive an action from that coordinator?

Used in: Record

• required uint64 proposal = 1

Represents a "promise" request from a proposer with the specified 'proposal' to a replica. If the proposer is a coordinator, most such requests will occur after a coordinator has failed and a new coordinator is elected. In such a case, the position that the coordinator is asking the replica to promise is implicitly *all* positions that the replica has made no promises (thus the position field is not be used). In other instances, however, a proposer might be explicitly trying to request that a replica promise a specific position in the log (such as when trying to fill holes discovered during a client read), and then the 'position' field will be present.

• required uint64 proposal = 1

• optional uint64 position = 2

Represents a "promise" response from a replica back to a proposer. A replica represents a NACK (because it has promised a proposer with a higher proposal number) by setting the okay field to false. The 'proposal' is either the aforementioned higher proposal number when the response is a NACK, or the corresponding request's proposal number if it is an ACK. The replica either sends back the highest position it has recorded in the log (using the 'position' field) or the specific action (if any) it has at the position requested in PromiseRequest (using the 'action' field).

• required bool okay = 1

• required uint64 proposal = 2

• optional uint64 position = 4

• optional Action action = 3

• required FrameworkID framework_id = 1

• repeated TaskStatus statuses = 2

• required SlaveID slave_id = 1

Represents a log record written to the local filesystem by a replica. A log record may store a promise (DEPRECATED), an action or metadata (defined above).

• required Record.Type type = 1

• optional Promise promise = 2

  DEPRECATED!

• optional Action action = 3

• optional Metadata metadata = 4

Used in: Record

• PROMISE = 1

  DEPRECATED!

• ACTION = 2

• METADATA = 3

Represents a recover request. A recover request is used to initiate the recovery (by broadcasting it).

(message has no fields)

When a replica receives a RecoverRequest, it will reply with its current status, and the begin and the end of its current log.

• required Metadata.Status status = 1

• optional uint64 begin = 2

• optional uint64 end = 3

• required FrameworkID framework_id = 1

• required ExecutorID executor_id = 2

• required FrameworkInfo framework = 1

• required string project = 1

• required SlaveInfo slave = 1

* Describes a request for resources that can be used by a framework to proactively influence the allocator. If 'slave_id' is provided then this request is assumed to only apply to resources on that slave.

Used in: ResourceRequestMessage

• optional SlaveID slave_id = 1

• repeated Resource resources = 2

• required ExecutorID executor_id = 1

• required FrameworkID framework_id = 2

• repeated TaskInfo tasks = 3

• repeated StatusUpdate updates = 4

• required FrameworkInfo framework = 2

• required bool failover = 3

• required SlaveID slave_id = 1

  TODO(bmahler): Deprecate and remove the explicit slave_id as SlaveInfo already includes this information.

• required SlaveInfo slave = 2

• repeated ExecutorInfo executor_infos = 4

• repeated Task tasks = 3

• repeated Archive.Framework completed_frameworks = 5

• required OfferID offer_id = 1

* Describes a resource on a machine. A resource can take on one of three types: scalar (double), a list of finite and discrete ranges (e.g., [1-10, 20-30]), or a set of items. A resource is described using the standard protocol buffer "union" trick. TODO(benh): Add better support for "expected" resources (e.g., cpus, memory, disk, network).

Used in: ExecutorInfo, Offer, Request, SlaveInfo, Task, TaskInfo

• required string name = 1

• required Value.Type type = 2

• optional Value.Scalar scalar = 3

• optional Value.Ranges ranges = 4

• optional Value.Set set = 5

• optional string role = 6

• repeated Offer offers = 1

• repeated string pids = 2

• required FrameworkID framework_id = 1

• repeated Request requests = 2

A snapshot of resource usage statistics.

Used in: ResourceUsage

• required double timestamp = 1

  Snapshot time, in seconds since the Epoch.

• optional double cpus_user_time_secs = 2

  CPU Usage Information: Total CPU time spent in user mode, and kernel mode.

• optional double cpus_system_time_secs = 3

• required double cpus_limit = 4

  Number of CPUs allocated.

• optional uint32 cpus_nr_periods = 7

  cpu.stat on process throttling (for contention issues).

• optional uint32 cpus_nr_throttled = 8

• optional double cpus_throttled_time_secs = 9

• optional uint64 mem_rss_bytes = 5

  Memory Usage Information:

  Resident Set Size.

• optional uint64 mem_limit_bytes = 6

  Amount of memory resources allocated.

• optional uint64 mem_file_bytes = 10

  Broken out memory usage information (files, anonymous, and mmaped files)

• optional uint64 mem_anon_bytes = 11

• optional uint64 mem_mapped_file_bytes = 12

* Describes a snapshot of the resource usage for an executor. TODO(bmahler): Note that we want to be sending this information to the master, and subsequently to the relevant scheduler. So this proto is designed to be easy for the scheduler to use, this is why we provide the slave id, executor info / task info.

• required SlaveID slave_id = 1

• required FrameworkID framework_id = 2

• optional ExecutorID executor_id = 3

  If present, this executor was

• optional string executor_name = 4

  explicitly specified.

• optional TaskID task_id = 5

  If present, the task did not have an executor.

• optional ResourceStatistics statistics = 6

  If missing, the isolation module cannot provide resource usage.

• required FrameworkID framework_id = 1

Describes a role, which are used to group frameworks for allocation decisions, depending on the allocation policy being used. The weight field can be used to indicate forms of priority.

• required string name = 1

• optional double weight = 2

• required FrameworkID framework_id = 1

• required FrameworkInfo framework = 2

• required string pid = 3

• required TaskInfo task = 4

Tells the executor to initiate a shut down by invoking Executor::shutdown.

(message has no fields)

Tells a slave to shut down all executors of the given framework.

• required FrameworkID framework_id = 1

(message has no fields)

* A unique ID assigned to a slave. Currently, a slave gets a new ID whenever it (re)registers with Mesos. Framework writers shouldn't assume any binding between a slave ID and and a hostname.

Used in: ExecutorRegisteredMessage, ExecutorReregisteredMessage, ExecutorToFrameworkMessage, ExitedExecutorMessage, FrameworkToExecutorMessage, HeartbeatMessage, LostSlaveMessage, Offer, ReconnectExecutorMessage, Request, ReregisterSlaveMessage, ResourceUsage, SlaveInfo, SlaveRegisteredMessage, SlaveReregisteredMessage, StatusUpdate, StatusUpdateAcknowledgementMessage, Task, TaskInfo, TaskStatus, UnregisterSlaveMessage

• required string value = 1

* Describes a slave. Note that the 'id' field is only available after a slave is registered with the master, and is made available here to facilitate re-registration. If checkpoint is set, the slave is checkpointing its own information and potentially frameworks' information (if a framework has checkpointing enabled).

Used in: ExecutorRegisteredMessage, ExecutorReregisteredMessage, RegisterSlaveMessage, ReregisterSlaveMessage

• required string hostname = 1

• optional int32 port = 8

• repeated Resource resources = 3

• repeated Attribute attributes = 5

• optional SlaveID id = 6

• optional bool checkpoint = 7

• optional string webui_hostname = 2

  Deprecated!

• optional int32 webui_port = 4

• required SlaveID slave_id = 1

• required SlaveID slave_id = 1

* Status is used to indicate the state of the scheduler and executor driver after function calls.

• DRIVER_NOT_STARTED = 1

• DRIVER_RUNNING = 2

• DRIVER_ABORTED = 3

• DRIVER_STOPPED = 4

Used in: ReregisterExecutorMessage, StatusUpdateMessage, StatusUpdateRecord

• required FrameworkID framework_id = 1

• optional ExecutorID executor_id = 2

• optional SlaveID slave_id = 3

• required TaskStatus status = 4

• required double timestamp = 5

• required bytes uuid = 6

• required SlaveID slave_id = 1

• required FrameworkID framework_id = 2

• required TaskID task_id = 3

• required bytes uuid = 4

NOTE: If 'pid' is present, scheduler driver sends an acknowledgement to the pid.

• required StatusUpdate update = 1

• optional string pid = 2

This message encapsulates how we checkpoint a status update to disk. NOTE: If type == UPDATE, the 'update' field is required. NOTE: If type == ACK, the 'uuid' field is required.

• required StatusUpdateRecord.Type type = 1

• optional StatusUpdate update = 2

• optional bytes uuid = 3

Used in: StatusUpdateRecord

• UPDATE = 0

• ACK = 1

• required string name = 1

• required bool okay = 1

TODO(benh): It would be great if this could just be a TaskInfo wherever it gets used! However, doing so would require adding the framework_id field, the executor_id field, and the state field into TaskInfo though (or send them another way). Also, one performance reason why we don't do that now is because storing whatever data is coupled with a TaskInfo could be large and unnecessary. TODO(bmahler): Add executor_uuid here, and send it to the master. This will allow us to expose executor work directories for tasks in the webui when looking from the master level. Currently only the slave knows which run the task belongs to.

Used in: Archive.Framework, ReregisterSlaveMessage

• required string name = 1

• required TaskID task_id = 2

• required FrameworkID framework_id = 3

• optional ExecutorID executor_id = 4

• required SlaveID slave_id = 5

• required TaskState state = 6

• repeated Resource resources = 7

• repeated TaskStatus statuses = 8

* A framework generated ID to distinguish a task. The ID must remain unique while the task is active. However, a framework can reuse an ID _only_ if a previous task with the same ID has reached a terminal state (e.g., TASK_FINISHED, TASK_LOST, TASK_KILLED, etc.).

Used in: KillTaskMessage, ResourceUsage, StatusUpdateAcknowledgementMessage, Task, TaskInfo, TaskStatus

• required string value = 1

* Describes a task. Passed from the scheduler all the way to an executor (see SchedulerDriver::launchTasks and Executor::launchTask). Either ExecutorInfo or CommandInfo should be set. A different executor can be used to launch this task, and subsequent tasks meant for the same executor can reuse the same ExecutorInfo struct.

Used in: LaunchTasksMessage, ReregisterExecutorMessage, RunTaskMessage

• required string name = 1

• required TaskID task_id = 2

• required SlaveID slave_id = 3

• repeated Resource resources = 4

• optional ExecutorInfo executor = 5

• optional CommandInfo command = 7

• optional bytes data = 6

* Describes possible task states. IMPORTANT: Mesos assumes tasks that enter terminal states (see below) imply the task is no longer running and thus clean up any thing associated with the task (ultimately offering any resources being consumed by that task to another task).

Used in: Task, TaskStatus

• TASK_STAGING = 6

  Initial state. Framework status updates should not use.

• TASK_STARTING = 0

• TASK_RUNNING = 1

• TASK_FINISHED = 2

  TERMINAL.

• TASK_FAILED = 3

  TERMINAL.

• TASK_KILLED = 4

  TERMINAL.

• TASK_LOST = 5

  TERMINAL.

* Describes the current status of a task.

Used in: ReconcileTasksMessage, StatusUpdate, Task

• required TaskID task_id = 1

• required TaskState state = 2

• optional string message = 4

  Possible message explaining state.

• optional bytes data = 3

• optional SlaveID slave_id = 5

• optional double timestamp = 6

• required FrameworkID framework_id = 1

• required SlaveID slave_id = 1

• required FrameworkID framework_id = 1

• required string pid = 2

* Describes an Attribute or Resource "value". A value is described using the standard protocol buffer "union" trick.

• required Value.Type type = 1

• optional Value.Scalar scalar = 2

• optional Value.Ranges ranges = 3

• optional Value.Set set = 4

• optional Value.Text text = 5

Used in: Ranges

• required uint64 begin = 1

• required uint64 end = 2

Used in: Attribute, Resource, Value

• repeated Range range = 1

Used in: Attribute, Resource, Value

• required double value = 1

Used in: Attribute, Resource, Value

• repeated string item = 1

Used in: Attribute, Value

• required string value = 1

Used in: Attribute, Resource, Value

• SCALAR = 0

• RANGES = 1

• SET = 2

• TEXT = 3

Represents a write request for a specific type of action. Note that we deliberately do not include the entire Action as it contains fields that are not relevant to a write request (e.g., promised, performed) and rather than ignore them we exclude them for safety.

• required uint64 proposal = 1

• required uint64 position = 2

• optional bool learned = 3

• required Action.Type type = 4

• optional Action.Nop nop = 5

• optional Action.Append append = 6

• optional Action.Truncate truncate = 7

Represents a write response corresponding to a write request. A replica represents a NACK (because it has promised a proposer with a higher proposal number) by setting the okay field to false. If the proposer is a coordinator, then it has been demoted. The 'position' should always correspond to the position set in the request. The 'proposal' is either the same proposal number set in the request in the case of an ACK, or the higher proposal number this position has been promised to in the case of a NACK.

• required bool okay = 1

• required uint64 proposal = 2

• required uint64 position = 3