Compute Cluster Management
Overview
This feature allows users to start, stop, cancel jobs, modify property configurations, and update cluster descriptions for a specified compute cluster. With these operations, users can flexibly manage compute resources to meet various business needs.
Syntax
Object Properties Description
Parameter Details
1. name
Specifies the name of the compute cluster.
2. objectProperties
Compute cluster property configuration. The specific fields and descriptions are as follows:
| Field Name | Description | Value Range | Default |
|---|---|---|---|
VCLUSTER_SIZE | Compute cluster specification. Supports specifications from 1 CRU to 256 CRU, with progressively stronger compute power. (Integration clusters additionally support 0.25 CRU and 0.5 CRU small specifications.) | Number: 1-256, in CRU (Compute Resource Unit). | 1 |
MIN_VCLUSTER_SIZE | Applicable only to General Purpose clusters. Minimum specification when the compute cluster scales down; must be less than or equal to MAX_VCLUSTER_SIZE. Cannot be used together with VCLUSTER_SIZE. | Number: 1-256, in CRU. | None |
MAX_VCLUSTER_SIZE | Applicable only to General Purpose clusters. Maximum specification when the compute cluster scales up; must be greater than or equal to MIN_VCLUSTER_SIZE. Cannot be used together with VCLUSTER_SIZE. | Number: 1-256, in CRU. | None |
MIN_REPLICAS | Minimum number of compute cluster instances. Only applicable to Analytics Purpose clusters. | 1-10 | 1 |
MAX_REPLICAS | Maximum number of compute cluster instances. Only applicable to Analytics Purpose clusters. | 1-10 | 1 |
AUTO_SUSPEND_IN_SECOND | Idle duration before the cluster auto-suspends. Unit: seconds. -1 means never auto-suspend. | -1 or integer >= 0 | 600 |
AUTO_RESUME | Whether to auto-resume. | TRUE | FALSE | TRUE |
MAX_CONCURRENCY | Maximum concurrency per compute instance. Only applicable to Analytics Purpose clusters. | 1-32 | 8 |
QUERY_RUNTIME_LIMIT_IN_SECOND | Maximum duration a job submitted to this cluster can execute. Unit: seconds. | Integer > 0 | 86400 |
PRELOAD_TABLES | Periodically or on-trigger pull specified table data into cluster local SSD cache. Only applicable to Analytics Purpose clusters. | schema_name.table_name, multiple table names separated by commas, supports wildcards | null |
QUERY_RESOURCE_LIMIT_RATIO | Single job resource ratio threshold: the CPU/memory resources used by a single query task must not exceed the specified proportion of total cluster resources. | 0.0 ~ 1.0 | 1.0 |
3. Modifying the Scaling Range of General Purpose Clusters
VCLUSTER_SIZE, MIN_VCLUSTER_SIZE, and MAX_VCLUSTER_SIZE cannot be set simultaneously.
Examples
-
Start the compute cluster named
sample_vc: -
Stop the compute cluster named
sample_vc: -
Force stop the compute cluster named
sample_vc(immediately terminate all running jobs): -
Cancel all jobs in the compute cluster named
sample_vc: -
Change the compute cluster named
sample_vcspecification to XSMALL (1 CRU): -
Change the maximum concurrency of the compute cluster named
sample_vcto 4: -
Modify the scaling range of the General Purpose cluster
sample_vc(minimum 1 CRU, maximum 4 CRU): -
Modify the cluster auto-suspend time to 300 seconds:
-
Disable cluster auto-suspend (never auto-suspend):
-
Modify the description of the compute cluster named
sample_vc:
