gcloud compute tpus queued-resources create

gcloud compute tpus queued-resources create - create a Queued Resource

gcloud compute tpus queued-resources create (QUEUED_RESOURCE : --zone=ZONE) --runtime-version=RUNTIME_VERSION (--accelerator-type=ACCELERATOR_TYPE     | --topology=TOPOLOGY --type=TYPE) (--node-id=NODE_ID     | [--node-count=NODE_COUNT : --node-prefix=NODE_PREFIX]) [--async] [--data-disk=[mode=MODE],[source=SOURCE]] [--description=DESCRIPTION] [--guaranteed] [--internal-ips] [--labels=[KEY=VALUE,…]] [--metadata=[KEY=VALUE,…]] [--metadata-from-file=[KEY=VALUE,…]] [--network=NETWORK; default="default"] [--range=RANGE] [--reserved] [--scopes=[SCOPES,…]] [--service-account=SERVICE_ACCOUNT] [--shielded-secure-boot] [--spot] [--subnetwork=SUBNETWORK] [--tags=[TAGS,…]] [--valid-after-duration=VALID_AFTER_DURATION] [--valid-after-time=VALID_AFTER_TIME] [--valid-until-duration=VALID_UNTIL_DURATION] [--valid-until-time=VALID_UNTIL_TIME] [--reservation-host-folder=RESERVATION_HOST_FOLDER --reservation-host-organization=RESERVATION_HOST_ORGANIZATION --reservation-host-project=RESERVATION_HOST_PROJECT] [GCLOUD_WIDE_FLAG …]

gcloud compute tpus queued-resources create my-queued-resource --accelerator-type=v4-8 --runtime-version=v2-alpha-tpuv4 --node-id=my-node-001 --zone=us-central2-b --project=my-project

To create a Queued Resource with multiple nodes, run:

gcloud compute tpus queued-resources create my-queued-resource --accelerator-type=v4-8 --runtime-version=v2-alpha-tpuv4 --node-count=2 --zone=us-central2-b --project=my-project

Queued resource resource - The Queued Resource you want to create. The arguments in this group can be used to specify the attributes of this resource. (NOTE) Some attributes are not given arguments in this group but can be set in other ways.

To set the project attribute:

• provide the argument queued_resource on the command line with a fully specified name;
• provide the argument --project on the command line;
• set the property core/project.

This must be specified.

QUEUED_RESOURCE

ID of the queued_resource or fully qualified identifier for the queued_resource.

To set the queued_resource attribute:

• provide the argument queued_resource on the command line.

This positional argument must be specified if any of the other arguments in this group are specified.

--zone=ZONE

The compute/zone of the Cloud TPU.

If not specified, will use default compute/zone. To set the zone attribute:

• provide the argument queued_resource on the command line with a fully specified name;
• provide the argument --zone on the command line;
• set the property compute/zone.

--runtime-version=RUNTIME_VERSION

Runtime version for the TPU, such as 2.3.

Specify a chip-based or core-based TPU node.

Exactly one of these must be specified:

Core-based TPU configuration

--accelerator-type=ACCELERATOR_TYPE

Accelerator type for the TPU.

Chip-based TPU configuration

--topology=TOPOLOGY

Chip topology for TPU.

This flag argument must be specified if any of the other arguments in this group are specified.

--type=TYPE

Type of TPU. TYPE must be one of: v2, v3, v4.

This flag argument must be specified if any of the other arguments in this group are specified.

Specify TPU node(s) with either a single node id or a node count and an optional node prefix

Exactly one of these must be specified:

Single node creation

--node-id=NODE_ID

Unqualified node identifier used to identify the node in the project once provisioned.

To request a resource with multiple nodes, in place of --node-id, use --node-count to specify the number of nodes and optionally use --node-prefix to specify the prefix for each node.

Multinode creation

--node-count=NODE_COUNT

The number of nodes in a multislice provision, also used to generate the qualified name for nodes in the provision. Value must be greater than 1.

This flag argument must be specified if any of the other arguments in this group are specified.

--node-prefix=NODE_PREFIX

Node prefix used to generate the qualified name of each node the multislice node provision. If not supplied, the queued resource id will be used as the prefix.

Must also specify --node-count.

--async

Return immediately, without waiting for the operation in progress to complete.

--data-disk=[mode=MODE],[source=SOURCE]

Additional data disks for the TPU VM.

This flag must be repeated to provide multiple data disks. For example:

gcloud compute tpus queued-resources create --data-disk source=projects/my-project/zones/us-central1-c/disks/my-disk,mode=read-only

The following keys are allowed:

source

Specifies the full path to an existing disk. Required. The disk must be in the same zone.

mode

Specifies the mode in which to attach this disk. Valid options are 'read-write', 'read-only'. If not specified, the default is 'read-write'.

--description=DESCRIPTION

Text description of the TPU.

--guaranteed

If provided, the Node requested here will only be scheduled at the 'guaranteed' tier.

--internal-ips

Indicates that the IP addresses for the node should be internal. The default is that external IP addresses will be associated with the TPU workers.

--labels=[KEY=VALUE,…]

Resource labels to represent user-provided metadata. See https://cloud.google.com/compute/docs/labeling-resources for details.

--metadata=[KEY=VALUE,…]

List of comma-separated metadata key-value pairs for the Cloud TPU VM node. Example: --metadata='key1=value1,key2=value2'

--metadata-from-file=[KEY=VALUE,…]

Same as --metadata except the value for the entry will be read from a local file. Example: --metadata-from-file='key1=value1.txt'

--network=NETWORK; default="default"

Network that this TPU will be a part of.

--range=RANGE

CIDR range for the TPU.

The IP range that the TPU will select an IP address from. Must be in CIDR notation and a /29 range, for example 192.168.0.0/29. Errors will occur if the CIDR range has already been used for a currently existing TPU, the CIDR range conflicts with any networks in the user's provided network, or the provided network is peered with another network that is using that CIDR range.

--reserved

Specifies the request should be scheduled on reserved capacity.

If --reservation-host-project, --reservation-host-folder, or --reservation-host-organization are present then this flag has no effect.

--scopes=[SCOPES,…]

List of comma-separated scopes to be made available for the service account.

--service-account=SERVICE_ACCOUNT

Email address of the service account. If empty, default Google Compute Engine service account will be used.

--shielded-secure-boot

Specifies that the TPU instances are created with secure boot enabled. This implicitly makes them Shielded VM instances.

--spot

If provided, the Node requested here will be created as Spot VMs.

--subnetwork=SUBNETWORK

Subnetwork that this TPU will be a part of.

--tags=[TAGS,…]

Tags to apply to the TPU Node. Tags are used to identify valid sources or targets for network firewalls. See https://cloud.google.com/vpc/docs/add-remove-network-tags for more details.

--valid-after-duration=VALID_AFTER_DURATION

A duration before which the TPU must not be provisioned, relative to the current time. See $ gcloud topic datetimes for information on duration formats.

--valid-after-time=VALID_AFTER_TIME

An absolute time before which the TPU must not be provisioned. See $ gcloud topic datetimes for information on duration formats.

--valid-until-duration=VALID_UNTIL_DURATION

A duration after which the TPU must not be provisioned, relative to the current time. See $ gcloud topic datetimes for information on duration formats.

--valid-until-time=VALID_UNTIL_TIME

An absolute time after which resources must not be created. See $ gcloud topic datetimes for information on duration formats.

--reservation-host-folder=RESERVATION_HOST_FOLDER

The folder hosting the reservation that the TPU should use. Only one reservation host entity may be specified.

--reservation-host-organization=RESERVATION_HOST_ORGANIZATION

The organization hosting the reservation that the TPU should use. Only one reservation host entity may be specified.

--reservation-host-project=RESERVATION_HOST_PROJECT

The project hosting the reservation that the TPU should use. Only one reservation host entity may be specified.

Run $ gcloud help for details.