Product Documentation

Appearance

CREATE CLUSTER ​

Create a new compute cluster or replace an existing cluster.

CREATE [ OR REPLACE ] [ COMPUTE ] CLUSTER [ IF NOT EXISTS ] name 
WITH (
[ NODE_COUNT number, ]
[ DEFAULT_CLUSTER [ TRUE | FALSE ], ]
[ HARDWARE_INSTANCE name, ]
[ WLM_PROFILE name,]
[ AUTO_SUSPEND number | NULL,]
[ AUTO_RESUME TRUE | FALSE,]
[ MAX_SPILL_PCT number | NULL,]
[ INITIALLY_SUSPENDED ]
)
OR REPLACE
Create a new compute cluster, or replace an existing compute cluster if one exists with the same name within the same instance.
COMPUTE
Optional keyword; in the current release, all clusters are compute clusters. They do the work of executing queries and other operations such as bulk loads.
name
Unique name for a cluster within an instance. Use double quotes for the cluster name if it contains special characters.
NODE_COUNT
The number of worker nodes assigned to the cluster.
DEFAULT_CLUSTER
Whether this cluster is declared the default cluster for the instance. Only one cluster within a given instance may be the default cluster.
HARDWARE_INSTANCE
Hardware instance type: small-v1, small-v2, large-v1, or large-v2. The cluster nodes must have the same hardware instance type.
WLM_PROFILE
WLM (workload management) profile associated with the cluster. A default profile is assigned when a cluster is created.
AUTO_SUSPEND
The number of seconds of user inactivity, after which the cluster is automatically suspended.
AUTO_RESUME
Whether to automatically resume the cluster when a new user query is submitted against the cluster (and it is suspended).
MAX_SPILL_PCT
Maximum disk space allocated for spilling as a percentage of total disk space.
INITIALLY_SUSPENDED
Put the cluster in suspended state on creation. The user must resume the cluster when it is ready for use.

Examples ​

The following example creates a compute cluster called premdb-queries with:

  • One node
  • The default WLM profile
  • The small hardware instance type (small-v1)
  • Automatic suspension after 300 seconds
  • Automatic resumption when a new query is run against the cluster
CREATE CLUSTER "premdb-queries" WITH (
NODE_COUNT 1,
WLM_PROFILE 'default',
HARDWARE_INSTANCE 'small-v1',
AUTO_SUSPEND 300,
AUTO_RESUME TRUE
);
CREATE COMPUTE CLUSTER

The following example creates a compute cluster called April12_LargeV1_2Nodes with:

  • 2 nodes
  • The flex WLM profile
  • A large hardware instance (large-v1)
  • Automatic suspension after 30 seconds
  • No automatic resumption when a new query is run against the cluster
  • An initial suspended state (not running on creation)
CREATE CLUSTER "April12_LargeV1_2Nodes" WITH (
  NODE_COUNT 2
, WLM_PROFILE 'flex'
, HARDWARE_INSTANCE 'large-v1'
, AUTO_SUSPEND 30
, AUTO_RESUME FALSE
, INITIALLY_SUSPENDED
);
CREATE COMPUTE CLUSTER

Parent topic:SQL Commands