@Rocky give me documentation about multiwarehousing in starrocks

StarRocks' Multi-warehouse feature (also referred to as Compute Groups) is a core capability of its shared-data architecture (v3.0+). It allows you to create multiple independent compute clusters that share the same underlying storage (e.g., S3, GCS, or HDFS), enabling physical resource isolation and independent scaling for different workloads. Key Concepts * Compute-Storage Separation: In a shared-data cluster, data is stored in remote object storage. Compute Nodes (CNs) are stateless and handle query execution. * Resource Isolation: You can group CN nodes into separate warehouses. This ensures that a heavy ingestion job or a complex ad-hoc query in one warehouse does not impact the performance of production BI reports running in another. * Elasticity: Warehouses can be scaled up, scaled down, or even suspended/resumed independently without the need for data migration. Key Documentation Details Based on recent releases (StarRocks 3.3 and later), the multi-warehouse feature includes the following capabilities: 1. SQL Operations You can manage warehouses using standard SQL commands: * Create a Warehouse:

CREATE WAREHOUSE <warehouse_name>;

* Switch Warehouse: To specify which warehouse the current session should use:

sql
    SET warehouse = 'warehouse_name';

* View Warehouses:

SHOW WAREHOUSES;

* Drop a Warehouse:

DROP WAREHOUSE <warehouse_name>;

2. Benefits for Workload Management * Isolated Ingestion: Use a dedicated warehouse for data loading to prevent "write" workloads from competing for CPU/memory with "read" queries. * Multi-tenancy: Assign different departments or business units to their own warehouses to simplify resource accounting and ensure SLAs. * Data Sharing: Since all warehouses access the same data in object storage, there is no need to copy or move data between clusters. Requirements & Availability * Version: Explicit multi-warehouse management is generally available for shared-data clusters starting from StarRocks 3.3. * Architecture: It requires the shared-data (cloud-native) architecture where FEs manage metadata and CNs provide compute resources. It is not available in the classic shared-nothing (BE-based) architecture. For specific implementation details in your environment, you may also refer to cloud-managed versions (like Alibaba Cloud EMR Serverless or CelerData), which provide UI-based management for these warehouses. References • data_source/catalog/hive_catalog.md • sql-reference/sql-statements/Resource/CREATE_RESOURCE.md • sql-reference/sql-statements/Catalog/SET_CATALOG.md • data_source/catalog/unified_catalog.md • sql-reference/sql-statements/backup_restore/CREATE_REPOSITORY.md

@Rocky no there should be separate page for warehouse compute groups in starrocks documentation and give me that url

In the official StarRocks documentation, the "Multi-warehouse" feature (also known as Compute Groups) is primarily documented under the Administration and SQL Reference sections, specifically for clusters using the shared-data architecture (v3.0+). While there isn't a single "multiwarehousing" landing page in older versions, the feature is fully detailed starting with StarRocks v3.3. You can find the dedicated management page here: Official Documentation URL * Warehouse Management: https://docs.starrocks.io/docs/administration/management/warehouse/ Key Sections for Multi-warehouse If you are navigating the documentation manually, look for these specific areas: 1. Core Concept & Architecture: Found under Shared-data Clusters, which explains the separation of storage and compute required for this feature. 2. SQL Reference: * CREATE WAREHOUSE: Syntax for creating a new compute group. * SET WAREHOUSE: How to switch the current session to a specific warehouse. * SHOW WAREHOUSES: Monitoring your existing warehouses. 3. Observability: * `warehouse_queries`: An Information Schema view to see queries running on specific warehouses. * `warehouse_metrics`: Metrics for monitoring performance per compute group. Version Note Explicit warehouse management commands (

CREATE WAREHOUSE

, etc.) were introduced as GA in StarRocks 3.3. If you are on an earlier 3.x version, you may still achieve resource isolation via Resource Groups, but the physical isolation provided by independent compute nodes (CNs) is the hallmark of the Multi-warehouse feature. References • administration/management/BE_configuration.md • administration/management/FE_configuration.md

@Rocky everything related to warehouses is deleted in documentation why?

Documentation for Multi-warehouse (also known as Compute Groups) in StarRocks is not deleted; however, its location and terminology have been updated in recent versions (v3.3+). The multi-warehouse feature is exclusively available for the Shared-data architecture (where storage and compute are decoupled). If you are looking at documentation for a Shared-nothing (classic) cluster, you will not find these settings because that architecture relies on Resource Groups for isolation instead. 1. Where to find the Documentation The core documentation for managing warehouses is now part of the Shared-data cluster management section: * Concepts & Architecture: Found under Shared-data Clusters. * Information Schema: You can monitor warehouse activity via: * `warehouse_queries`: View current queries per warehouse. * `warehouse_metrics`: View performance metrics per warehouse. * Utility Functions: `current_warehouse()` returns the warehouse used by your session. 2. Why it might seem "missing" * Architecture Specific: The explicit SQL commands like

CREATE WAREHOUSE

only work if your cluster is deployed in Shared-data mode. In a standard Shared-nothing deployment,

SHOW COMPUTE NODES

will simply show

default_warehouse

for all nodes, and you cannot create more. * Terminology Shift: In newer docs, StarRocks often uses the term Compute Groups to describe the physical isolation of Compute Nodes (CNs). * Resource Groups vs. Warehouses: For software-level isolation (CPU/Memory limits within a node), StarRocks uses Resource Groups. For physical isolation (separate clusters sharing data), it uses Warehouses. 3. Key SQL Commands (Shared-data Only) If you are on a Shared-data cluster (v3.3+), you can use the following commands:

sql
-- Create a new warehouse (group of CN nodes)
CREATE WAREHOUSE <warehouse_name>;

-- Switch your session to a specific warehouse
SET warehouse = '<warehouse_name>';

-- Check which warehouse you are currently using
SELECT current_warehouse();

-- List all warehouses in the cluster
SHOW WAREHOUSES;

If you are trying to navigate the website and the links are broken, it is likely due to the documentation restructuring for the v3.3 and v3.4 releases. Always ensure you have selected the correct version in the documentation sidebar (e.g., Stable-3.3 or Latest-4.1). References • sql-reference/sql-statements/cluster-management/resource_group/CREATE_RESOURCE_GROUP.md • sql-reference/sql-statements/Resource/CREATE_RESOURCE.md • sql-reference/sql-statements/backup_restore/CREATE_REPOSITORY.md