@Rocky i see 4.1.0 already released but Does it support mulitwarehouse? because i still see this

public void createWarehouse(CreateWarehouseStmt stmt) throws DdlException {
        throw new DdlException("Multi-Warehouse is not implemented");
    }

Hi there! I'm Rocky. 🦦 It's a great catch! You are seeing that specific

DdlException

because, in StarRocks 4.1.0, the architectural support for Multi-Warehouse is fully present, but the SQL DDL syntax (

CREATE WAREHOUSE

) is still a placeholder in certain deployment modes or is being transitioned from infrastructure-level management to SQL-level management. Here is the breakdown of the current status in version 4.1.0: 1. Multi-Warehouse IS supported (Architecturally) Even if that specific

CREATE WAREHOUSE

command throws an error, StarRocks 4.1.0 has introduced significant Multi-Warehouse features: * Resource Group Binding: You can now bind Resource Groups to specific warehouses using the

warehouses

attribute. This allows you to isolate CPU and memory resources between different sets of compute nodes (CNs). * System Observability: New system views like

information_schema.warehouse_metrics

and

information_schema.warehouse_queries

are available to monitor performance across different compute clusters. * Session Management: You can use

SET warehouse = 'your_warehouse_name';

to route your session's queries to a specific compute group, and check the current one using the

current_warehouse()

function. 2. Why the DDL says "Not Implemented" In the current version, "Warehouses" are primarily defined as logical groupings of Compute Nodes (CNs). * How they are "Created": Instead of a SQL

CREATE WAREHOUSE

command, warehouses are typically established at the deployment level. You deploy a set of CN nodes and assign them a warehouse name (via configuration or labels). * The Future: The

CREATE WAREHOUSE

DDL you found in the code is the planned way to manage these clusters entirely via SQL (including scaling and lifecycle), similar to other cloud-native warehouses. While the engine supports the isolation today, the unified SQL management for creating them is still being finalized. 3. How to use Multi-Warehouse in 4.1.0 If you want to leverage multi-warehouse capabilities today, you should focus on Resource Groups:

sql
-- You can now bind resource groups to specific warehouses in 4.1.0
ALTER RESOURCE GROUP your_group_name 
PROPERTIES (
    "warehouses" = "analytics_warehouse",
    "cpu_weight_percent" = "50"
);

Summary: Don't let that code snippet discourage you! The compute isolation and multi-tenant management are very much "implemented" and are core highlights of the 4.1.0 release. The error you found simply means that for now, you define the "Warehouse" by grouping your nodes at the infrastructure/config level rather than using that specific SQL statement. If you are looking to set up a multi-warehouse environment on a shared-data (Lake) cluster, I'd recommend checking the latest Resource Group documentation for 4.1, as that is where the "Warehouse" attributes are now managed. References • sql-reference/sql-statements/backup_restore/CREATE_REPOSITORY.md • sql-reference/sql-statements/cluster-management/resource_group/CREATE_RESOURCE_GROUP.md • sql-reference/sql-statements/Resource/CREATE_RESOURCE.md

@Alvin Zhao Can you confirm?

https://github.com/StarRocks/starrocks/issues/67632 it does not release yet. you can see loadmap