Managing compute pools

A compute pool is a collection of one or more virtual machine (VM) nodes on which Snowflake runs your Snowpark Container Services jobs and services.

For more information about compute pools, see Snowpark Container Services: Working with compute pools.

This topic shows how to do the following tasks with services:

For common operations, such as listing or dropping, Snowflake CLI uses snow object commands as described in Managing Snowflake objects.

Create a compute pool

To create a compute pool named “pool_1” composed of two CPUs with 4 GB of memory, enter a spcs pool create command similar to the following:

snow spcs compute-pool create "pool_1" --min-nodes 2 --max-nodes 2 --family "CPU_X64_XS"
Copy

For more information about instance families, see the SQL CREATE COMPUTE POOL command.

Create a compute pool from a project definition

You can create a compute pool from a snowflake.yml project definition file and then executing the snow spcs compute-pool deploy command.

The following shows a sample snowflake.yml project definition file:

definition_version: 2
entities:
  my_compute_pool:
    type: compute-pool
    identifier:
      name: my_compute_pool
    min_nodes: 1
    max_nodes: 2
    instance_family: CPU_X64_XS
    auto_resume: true
    initially_suspended: true
    auto_suspend_seconds: 60
    comment: "My compute pool"
    tags:
      - name: my_tag
        value: tag_value
Copy

The following table describes the properties of a compute pool project definition.

Compute pool project definition properties

Property

Definition

type

required, string

Must be compute-pool.

identifier

optional, string

Snowflake identifier for the entity. The value can have the following forms:

  • String identifier text

    identifier: my-compute-pool
    Copy

    Both unquoted and quoted identifiers are supported. To use quoted identifiers, include the surrounding quotes in the YAML value (for example, "My Compute Pool").

  • Object

    identifier:
  name: my-compute-pool
  schema: my-schema # optional
  database: my-db # optional
    Copy

    Note

    An error occurs if you specify a schema or database and use a fully qualified name in the name property (such as mydb.schema1.my-app).

instance_family

required, string

Name of the instance family. For a list of available instance families, see the CREATE COMPUTE POOL INSTANCE_FAMILY parameter.

min_nodes

optional, string

Minimum number of nodes for the compute pool. This value must be greater than 0.

Default: 1

max_nodes

optional, int

Maximum number of nodes for the compute pool.

auto_resume

optional, boolean

Whether to automatically resume a compute pool when a service or job is submitted to it.

Default: True

initially_suspended

optional, boolean

Whether the compute pool is created initially in the suspended state. If true, Snowflake doesn’t provision any nodes requested for the compute pool at the compute pool creation time.

Default: False

auto_suspend_seconds

optional, int

Number of seconds of inactivity after which you want Snowflake to automatically suspend the compute pool.

Default: 3600

comment

optional, string

Comments to associate with the compute pool.

tags

optional, Tag sequence

Tag names and values for the compute pool. For more information, see Tag quotas for objects and columns

To create and deploy the compute pool to a stage, do the following:

  1. Change your current directory to the directory containing the project definition file.

  2. Run a snow spcs compute-pool deploy command similar to the following:

    snow spcs compute-pool deploy
    Copy
    +---------------------------------------------------------------------+
| key    | value                                                      |
|--------+------------------------------------------------------------|
| status | Compute pool MY_COMPUTE_POOL successfully created.         |
+---------------------------------------------------------------------+

Suspend and resume a compute pool

Note

The current role must have OPERATE privilege on the compute pool to suspend or resume it.

To suspend a compute pool, enter a command similar to the following:

snow spcs compute-pool suspend tutorial_compute_pool
Copy
+-------------------------------------------+
| key    | value                            |
|--------+----------------------------------|
| status | Statement executed successfully. |
+-------------------------------------------+

To resume a suspended compute pool, enter a command similar to the following:

snow spcs compute-pool resume tutorial_compute_pool
Copy
+-------------------------------------------+
| key    | value                            |
|--------+----------------------------------|
| status | Statement executed successfully. |
+-------------------------------------------+

Set and unset a compute pool’s properties or parameters

Note

The current role must have MODIFY privilege on the compute pool to set properties.

To set a property or parameter, enter a command similar to the following:

snow spcs compute-pool set tutorial_compute_pool --min-nodes 2 --max-nodes 4
Copy
+-------------------------------------------+
| key    | value                            |
|--------+----------------------------------|
| status | Statement executed successfully. |
+-------------------------------------------+

To reset a property or parameter to its default value, enter a command similar to the following:

snow spcs compute-pool unset tutorial_compute_pool --auto-resume
Copy
+-------------------------------------------+
| key    | value                            |
|--------+----------------------------------|
| status | Statement executed successfully. |
+-------------------------------------------+

Stop all services in a compute pool

Stopping a compute pool deletes all of the services running on the compute pool; however, it does not stop the compute pool itself.

To stop a compute pool, enter a spcs compute-pool stop-all command similar to the following:

snow spcs compute-pool stop-all "pool_1"
Copy
Language: English