create

Description

Creates a named MANAGED or CUSTOM pool. MANAGED uses the service-provided python-3.11 runtime and requires containerImage to be omitted. CUSTOM requires a digest-pinned containerImage. Project-wide shape capacity limits include named pools of both types and reserved default pools. Type, image, shape and name are immutable. The CP atomically persists CREATING and a Work Request, then its workflow prepares an immutable pool-owned template and applies the pool configuration. Each MANAGED pool has a distinct template based on the service runtime configuration, isolated from other pools of the same shape. ACTIVE means configuration was applied and verified; it does not require ready members, a successful image boot or available cluster capacity. Zero capacity pauses idle prewarming. An ACTIVE named pool can still create Sandboxes through cold fallback when its applied template is available. Retry identity and configuration are persisted atomically. Within the 24-hour retry-token validity period, the same token and input identify the original pool id, including after name reuse. A retry never creates or selects a replacement pool. A retry after the original pool is deleted returns 409; changed input with the same token conflicts. Reserved default names cannot be reused for either type.

Usage

oci generative-ai sandbox-warm-pool create [OPTIONS]

Required Parameters

--desired-capacity [integer]

Desired number of idle members. Zero pauses idle prewarming. Once the pool is ACTIVE, named-pool Sandbox creation can still cold-start from its available applied template when no eligible warm member is available. Existing customer Sandboxes continue. A positive target does not guarantee ready capacity. Existing service capacity validation still applies.

--name [text]

Unique immutable name within the Project.

--project-id [text]

Owning Generative AI Project OCID. Resolve parent compartment for authorization.

--shape [text]

Compute shape for each idle Sandbox member.

Accepted values are:

LARGE, MEDIUM, SMALL
--type [text]

MANAGED uses the service-provided python-3.11 runtime. CUSTOM uses a customer-provided containerImage. Both types support named pool creation.

Accepted values are:

CUSTOM, MANAGED

Optional Parameters

--container-image [text]

Required for CUSTOM and must be omitted for MANAGED. A CUSTOM image must be a digest-pinned same-region OCIR reference, using sha256 with exactly 64 lowercase hexadecimal characters. The CP workflow prepares the template through the data plane, reusing current BYOC image and Project identity validation. Successful configuration does not guarantee image pull or runtime readiness.

--from-json [text]

Provide input to this command as a JSON document from a file using the file://path-to/file syntax.

The --generate-full-command-json-input option can be used to generate a sample json file to be used with this command option. The key names are pre-populated and match the command option names (converted to camelCase format, e.g. compartment-id –> compartmentId), while the values of the keys need to be populated by the user before using the sample file as an input to this command. For any command option that accepts multiple values, the value of the key can be a JSON array.

Options can still be provided on the command line. If an option exists in both the JSON document and the command line then the command line specified value will be used.

For examples on usage of this option, please see our “using CLI with advanced JSON options” link: https://docs.cloud.oracle.com/iaas/Content/API/SDKDocs/cliusing.htm#AdvancedJSONOptions

--max-wait-seconds [integer]

The maximum time to wait for the work request to reach the state defined by --wait-for-state. Defaults to 1200 seconds.

--wait-for-state [text]

This operation asynchronously creates, modifies or deletes a resource and uses a work request to track the progress of the operation. Specify this option to perform the action and then wait until the work request reaches a certain state. Multiple states can be specified, returning on the first state. For example, --wait-for-state ACCEPTED --wait-for-state CANCELED would return on whichever lifecycle state is reached first. If timeout is reached, a return code of 2 is returned. For any other error, a return code of 1 is returned.

Accepted values are:

ACCEPTED, CANCELED, CANCELING, FAILED, IN_PROGRESS, SUCCEEDED, WAITING
--wait-interval-seconds [integer]

Check every --wait-interval-seconds to see whether the work request has reached the state defined by --wait-for-state. Defaults to 30 seconds.

Example using required parameter

Copy the following CLI commands into a file named example.sh. Run the command by typing “bash example.sh” and replacing the example parameters with your own.

Please note this sample will only work in the POSIX-compliant bash-like shell. You need to set up the OCI configuration and appropriate security policies before trying the examples.

    export desired_capacity=<substitute-value-of-desired_capacity> # https://docs.cloud.oracle.com/en-us/iaas/tools/oci-cli/latest/oci_cli_docs/cmdref/generative-ai/sandbox-warm-pool/create.html#cmdoption-desired-capacity
    export name=<substitute-value-of-name> # https://docs.cloud.oracle.com/en-us/iaas/tools/oci-cli/latest/oci_cli_docs/cmdref/generative-ai/sandbox-warm-pool/create.html#cmdoption-name
    export project_id=<substitute-value-of-project_id> # https://docs.cloud.oracle.com/en-us/iaas/tools/oci-cli/latest/oci_cli_docs/cmdref/generative-ai/sandbox-warm-pool/create.html#cmdoption-project-id
    export shape=<substitute-value-of-shape> # https://docs.cloud.oracle.com/en-us/iaas/tools/oci-cli/latest/oci_cli_docs/cmdref/generative-ai/sandbox-warm-pool/create.html#cmdoption-shape
    export type=<substitute-value-of-type> # https://docs.cloud.oracle.com/en-us/iaas/tools/oci-cli/latest/oci_cli_docs/cmdref/generative-ai/sandbox-warm-pool/create.html#cmdoption-type

    oci generative-ai sandbox-warm-pool create --desired-capacity $desired_capacity --name $name --project-id $project_id --shape $shape --type $type