Package google.genomics.v2alpha1

Index

PipelinesV2Alpha1

A service for running pipelines consisting of Docker containers.

RunPipeline

rpc RunPipeline(RunPipelineRequest) returns (Operation)

Runs a pipeline.

Note: Before you can use this method, the Genomics Service Agent must have access to your project. This is done automatically when the Cloud Genomics API is first enabled, but if you delete this permission, or if you enabled the Cloud Genomics API before the v2alpha1 API launch, you must disable and re-enable the API to grant the Genomics Service Agent the required permissions. Authorization requires the following Google IAM permission:

  • genomics.operations.create
Authorization Scopes

Requires one of the following OAuth scopes:

  • https://www.googleapis.com/auth/cloud-platform
  • https://www.googleapis.com/auth/genomics

For more information, see the Authentication Overview.

Accelerator

Carries information about an accelerator that can be attached to a VM.

Fields
type

string

The accelerator type string (for example, "nvidia-tesla-k80").

Only NVIDIA GPU accelerators are currently supported. If an NVIDIA GPU is attached, the required runtime libraries will be made available to all containers under /usr/local/nvidia. The driver version to install must be specified using the NVIDIA driver version parameter on the virtual machine specification. Note that attaching a GPU increases the worker VM startup time by a few minutes.

count

int64

How many accelerators of this type to attach.

Action

Specifies a single action that runs a Docker container.

Fields
name

string

An optional name for the container. The container hostname will be set to this name, making it useful for inter-container communication. The name must contain only upper and lowercase alphanumeric characters and hypens and cannot start with a hypen.

image_uri

string

The URI to pull the container image from. Note that all images referenced by actions in the pipeline are pulled before the first action runs. If multiple actions reference the same image, it is only pulled once, ensuring that the same image is used for all actions in a single pipeline.

commands[]

string

If specified, overrides the CMD specified in the container. If the container also has an ENTRYPOINT the values are used as entrypoint arguments. Otherwise, they are used as a command and arguments to run inside the container.

entrypoint

string

If specified, overrides the ENTRYPOINT specified in the container.

environment

map<string, string>

The environment to pass into the container. This environment is merged with any values specified in the Pipeline message. These values overwrite any in the Pipeline message.

In addition to the values passed here, a few other values are automatically injected into the environment. These cannot be hidden or overwritten.

GOOGLE_PIPELINE_FAILED will be set to "1" if the pipeline failed because an action has exited with a non-zero status (and did not have the IGNORE_EXIT_STATUS flag set). This can be used to determine if additional debug or logging actions should execute.

GOOGLE_LAST_EXIT_STATUS will be set to the exit status of the last non-background action that executed. This can be used by workflow engine authors to determine whether an individual action has succeeded or failed.

pid_namespace

string

An optional identifier for a PID namespace to run the action inside. Multiple actions should use the same string to share a namespace. If unspecified, a separate isolated namespace is used.

flags[]

Flag

The set of flags to apply to this action.

port_mappings

map<int32, int32>

A map of containers to host port mappings for this container. If the container already specifies exposed ports, use the PUBLISH_EXPOSED_PORTS flag instead.

The host port number must be less than 65536. If it is zero, an unused random port is assigned. To determine the resulting port number, consult the ContainerStartedEvent in the operation metadata.

mounts[]

Mount

A list of mounts to make available to the action.

In addition to the values specified here, every action has a special virtual disk mounted under /google that contains log files and other operational components.

  • /google/logs

    All logs written during the pipeline execution.

  • /google/logs/output

    The combined standard output and standard error of all actions run as part of the pipeline execution.

  • /google/logs/action/*/stdout

    The complete contents of each individual action's standard output.

  • /google/logs/action/*/stderr

    The complete contents of each individual action's standard error output.

labels

map<string, string>

Labels to associate with the action. This field is provided to assist workflow engine authors in identifying actions (for example, to indicate what sort of action they perform, such as localization or debugging). They are returned in the operation metadata, but are otherwise ignored.

credentials

Secret

If the specified image is hosted on a private registry other than Google Container Registry, the credentials required to pull the image must be specified here as an encrypted secret.

The secret must decrypt to a JSON-encoded dictionary containing both username and password keys.

timeout

Duration

The maximum amount of time to give the action to complete. If the action fails to complete before the timeout, it will be terminated and the exit status will be non-zero. The pipeline will continue or terminate based on the rules defined by the ALWAYS_RUN and IGNORE_EXIT_STATUS flags.

Flag

Flags that control the execution of this action.

Enums
FLAG_UNSPECIFIED Unspecified flag.
IGNORE_EXIT_STATUS Normally, a non-zero exit status causes the pipeline to fail. This flag allows execution of other actions to continue instead.
RUN_IN_BACKGROUND This flag allows an action to continue running in the background while executing subsequent actions. This is useful to provide services to other actions (or to provide debugging support tools like SSH servers).
ALWAYS_RUN By default, after an action fails, no further actions are run. This flag indicates that this action must be run even if the pipeline has already failed. This is useful for actions that copy output files off of the VM or for debugging.
ENABLE_FUSE

Enable access to the FUSE device for this action. Filesystems can then be mounted into disks shared with other actions. The other actions do not need the ENABLE_FUSE flag to access the mounted filesystem.

This has the effect of causing the container to be executed with CAP_SYS_ADMIN and exposes /dev/fuse to the container, so use it only for containers you trust.

PUBLISH_EXPOSED_PORTS Exposes all ports specified by EXPOSE statements in the container. To discover the host side port numbers, consult the ACTION_STARTED event in the operation metadata.
DISABLE_IMAGE_PREFETCH

All container images are typically downloaded before any actions are executed. This helps prevent typos in URIs or issues like lack of disk space from wasting large amounts of compute resources.

If set, this flag prevents the worker from downloading the image until just before the action is executed.

DISABLE_STANDARD_ERROR_CAPTURE A small portion of the container's standard error stream is typically captured and returned inside the ContainerStoppedEvent. Setting this flag disables this functionality.

ContainerKilledEvent

An event generated when a container is forcibly terminated by the worker. Currently, this only occurs when the container outlives the timeout specified by the user.

Fields
action_id

int32

The numeric ID of the action that started the container.

ContainerStartedEvent

An event generated when a container starts.

Fields
action_id

int32

The numeric ID of the action that started this container.

port_mappings

map<int32, int32>

The container-to-host port mappings installed for this container. This set will contain any ports exposed using the PUBLISH_EXPOSED_PORTS flag as well as any specified in the Action definition.

ip_address

string

The public IP address that can be used to connect to the container. This field is only populated when at least one port mapping is present. If the instance was created with a private address, this field will be empty even if port mappings exist.

ContainerStoppedEvent

An event generated when a container exits.

Fields
action_id

int32

The numeric ID of the action that started this container.

exit_status

int32

The exit status of the container.

stderr

string

The tail end of any content written to standard error by the container. If the content emits large amounts of debugging noise or contains sensitive information, you can prevent the content from being printed by setting the DISABLE_STANDARD_ERROR_CAPTURE flag.

Note that only a small amount of the end of the stream is captured here. The entire stream is stored in the /google/logs directory mounted into each action, and can be copied off the machine as described elsewhere.

DelayedEvent

An event generated whenever a resource limitation or transient error delays execution of a pipeline that was otherwise ready to run.

Fields
cause

string

A textual description of the cause of the delay. The string can change without notice because it is often generated by another service (such as Compute Engine).

metrics[]

string

If the delay was caused by a resource shortage, this field lists the Compute Engine metrics that are preventing this operation from running (for example, CPUS or INSTANCES). If the particular metric is not known, a single UNKNOWN metric will be present.

Disk

Carries information about a disk that can be attached to a VM.

See https://cloud.google.com/compute/docs/disks/performance for more information about disk type, size, and performance considerations.

Fields
name

string

A user-supplied name for the disk. Used when mounting the disk into actions. The name must contain only upper and lowercase alphanumeric characters and hypens and cannot start with a hypen.

size_gb

int32

The size, in GB, of the disk to attach. If the size is not specified, a default is chosen to ensure reasonable I/O performance.

If the disk type is specified as local-ssd, multiple local drives are automatically combined to provide the requested size. Note, however, that each physical SSD is 375GB in size, and no more than 8 drives can be attached to a single instance.

type

string

The Compute Engine disk type. If unspecified, pd-standard is used.

source_image

string

An optional image to put on the disk before attaching it to the VM.

Event

Carries information about events that occur during pipeline execution.

Fields
timestamp

Timestamp

The time at which the event occurred.

description

string

A human-readable description of the event. Note that these strings can change at any time without notice. Any application logic must use the information in the details field.

details

Any

Machine-readable details about the event.

FailedEvent

An event generated when the execution of a pipeline has failed. Note that other events can continue to occur after this event.

Fields
code

Code

The Google standard error code that best describes this failure.

cause

string

The human-readable description of the cause of the failure.

Metadata

Carries information about the pipeline execution that is returned in the long running operation's metadata field.

Fields
pipeline

Pipeline

The pipeline this operation represents.

labels

map<string, string>

The user-defined labels associated with this operation.

events[]

Event

The list of events that have happened so far during the execution of this operation.

create_time

Timestamp

The time at which the operation was created by the API.

start_time

Timestamp

The first time at which resources were allocated to execute the pipeline.

end_time

Timestamp

The time at which execution was completed and resources were cleaned up.

Mount

Carries information about a particular disk mount inside a container.

Fields
disk

string

The name of the disk to mount, as specified in the resources section.

path

string

The path to mount the disk inside the container.

read_only

bool

If true, the disk is mounted read-only inside the container.

Network

VM networking options.

Fields
name

string

The network name to attach the VM's network interface to. The value will be prefixed with global/networks/ unless it contains a /, in which case it is assumed to be a fully specified network resource URL.

If unspecified, the global default network is used.

use_private_address

bool

If set to true, do not attach a public IP address to the VM. Note that without a public IP address, additional configuration is required to allow the VM to access Google services.

See https://cloud.google.com/vpc/docs/configure-private-google-access for more information.

subnetwork

string

If the specified network is configured for custom subnet creation, the name of the subnetwork to attach the instance to must be specified here.

The value is prefixed with regions/*/subnetworks/ unless it contains a /, in which case it is assumed to be a fully specified subnetwork resource URL.

If the * character appears in the value, it is replaced with the region that the virtual machine has been allocated in.

Pipeline

Specifies a series of actions to execute, expressed as Docker containers.

Fields
actions[]

Action

The list of actions to execute, in the order they are specified.

resources

Resources

The resources required for execution.

environment

map<string, string>

The environment to pass into every action. Each action can also specify additional environment variables but cannot delete an entry from this map (though they can overwrite it with a different value).

timeout

Duration

The maximum amount of time to give the pipeline to complete. This includes the time spent waiting for a worker to be allocated. If the pipeline fails to complete before the timeout, it will be cancelled and the error code will be set to DEADLINE_EXCEEDED.

If unspecified, it will default to 7 days.

PullStartedEvent

An event generated when the worker starts pulling an image.

Fields
image_uri

string

The URI of the image that was pulled.

PullStoppedEvent

An event generated when the worker stops pulling an image.

Fields
image_uri

string

The URI of the image that was pulled.

Resources

The system resources for the pipeline run.

At least one zone or region must be specified or the pipeline run will fail.

Fields
project_id

string

The project ID to allocate resources in.

regions[]

string

The list of regions allowed for VM allocation. If set, the zones field must not be set.

zones[]

string

The list of zones allowed for VM allocation. If set, the regions field must not be set.

virtual_machine

VirtualMachine

The virtual machine specification.

RunPipelineRequest

The arguments to the RunPipeline method. The requesting user must have the iam.serviceAccounts.actAs permission for the Cloud Genomics service account or the request will fail.

Fields
pipeline

Pipeline

The description of the pipeline to run.

labels

map<string, string>

User-defined labels to associate with the returned operation. These labels are not propagated to any Google Cloud Platform resources used by the operation, and can be modified at any time.

To associate labels with resources created while executing the operation, see the appropriate resource message (for example, VirtualMachine).

RunPipelineResponse

The response to the RunPipeline method, returned in the operation's result field on success.

Secret

Holds encrypted information that is only decrypted and stored in RAM by the worker VM when running the pipeline.

Fields
key_name

string

The name of the Cloud KMS key that will be used to decrypt the secret value. The VM service account must have the required permissions and authentication scopes to invoke the decrypt method on the specified key.

cipher_text

string

The value of the cipherText response from the encrypt method. This field is intentionally unaudited.

ServiceAccount

Carries information about a Google Cloud service account.

Fields
email

string

Email address of the service account. If not specified, the default Compute Engine service account for the project will be used.

scopes[]

string

List of scopes to be enabled for this service account on the VM, in addition to the Cloud Genomics API scope.

UnexpectedExitStatusEvent

An event generated when the execution of a container results in a non-zero exit status that was not otherwise ignored. Execution will continue, but only actions that are flagged as ALWAYS_RUN will be executed. Other actions will be skipped.

Fields
action_id

int32

The numeric ID of the action that started the container.

exit_status

int32

The exit status of the container.

VirtualMachine

Carries information about a Compute Engine VM resource.

Fields
machine_type

string

The machine type of the virtual machine to create. Must be the short name of a standard machine type (such as "n1-standard-1") or a custom machine type (such as "custom-1-4096", where "1" indicates the number of vCPUs and "4096" indicates the memory in MB). See Creating an instance with a custom machine type for more specifications on creating a custom machine type.

preemptible

bool

If true, allocate a preemptible VM.

labels

map<string, string>

Optional set of labels to apply to the VM and any attached disk resources. These labels must adhere to the name and value restrictions on VM labels imposed by Compute Engine.

Labels applied at creation time to the VM. Applied on a best-effort basis to attached disk resources shortly after VM creation.

disks[]

Disk

The list of disks to create and attach to the VM.

network

Network

The VM network configuration.

accelerators[]

Accelerator

The list of accelerators to attach to the VM.

service_account

ServiceAccount

The service account to install on the VM. This account does not need any permissions other than those required by the pipeline.

boot_disk_size_gb

int32

The size of the boot disk, in GB. The boot disk must be large enough to accommodate all of the Docker images from each action in the pipeline at the same time. If not specified, a small but reasonable default value is used.

cpu_platform

string

The CPU platform to request. An instance based on a newer platform can be allocated, but never one with fewer capabilities. The value of this parameter must be a valid Compute Engine CPU platform name (such as "Intel Skylake"). This parameter is only useful for carefully optimized work loads where the CPU platform has a significant impact.

For more information about the effect of this parameter, see https://cloud.google.com/compute/docs/instances/specify-min-cpu-platform.

boot_image

string

The host operating system image to use.

Currently, only Container-Optimized OS images can be used.

The default value is projects/cos-cloud/global/images/family/cos-stable, which selects the latest stable release of Container-Optimized OS.

This option is provided to allow testing against the beta release of the operating system to ensure that the new version does not interact negatively with production pipelines.

To test a pipeline against the beta release of Container-Optimized OS, use the value projects/cos-cloud/global/images/family/cos-beta.

nvidia_driver_version

string

The NVIDIA driver version to use when attaching an NVIDIA GPU accelerator. The version specified here must be compatible with the GPU libraries contained in the container being executed, and must be one of the drivers hosted in the nvidia-drivers-us-public bucket on Google Cloud Storage.

enable_stackdriver_monitoring

bool

Whether Stackdriver monitoring should be enabled on the VM.

WorkerAssignedEvent

An event generated after a worker VM has been assigned to run the pipeline.

Fields
zone

string

The zone the worker is running in.

instance

string

The worker's instance name.

WorkerReleasedEvent

An event generated when the worker VM that was assigned to the pipeline has been released (deleted).

Fields
zone

string

The zone the worker was running in.

instance

string

The worker's instance name.

¿Te ha resultado útil esta página? Enviar comentarios:

Enviar comentarios sobre...