Resource Management

Managing a BLUE cluster involves two core aspects: pools and Workers. Pools serve as organized resource compartments, enabling strategic allocation of Workers based on specific compute power needs. Workers, operating within these pools, execute tasks and process data. By adjusting the number of Workers in each pool, users can optimize resource usage, ensuring efficient compute power and task execution while effectively utilizing the cluster’s capabilities.

clusteradmin

Only a clusteradmin can manage Resource Pools.

Pools

Pools offer the ability to effectively manage available resources for various purposes within your BLUE cluster. By default, your cluster includes a single default pool that encompasses all the Workers in the cluster. You have the flexibility to create additional pools to further divide the resources based on your specific business needs, priorities, and concurrency preferences. This allocation of resources allows you to have better control over your business priorities and optimize parallelism, resulting in improved resource utilization and overall system efficiency.

Creating a New Pool

When creating new pools, Workers must be assigned to them. The number of Workers allocated to the new pool will be automatically deducted from your default pool.

1. In the sidebar, go to Settings and select the Resource Pool tab.

2. To create a new pool, select the Create New Pool button, provide a pool name, and assign Workers. Pool names are rendered as identifiers, which means they may not include whitespace characters.

   See full list of identifier rules.

Readjusting Existing Pools

To readjust an existing pool, the pool must be in either an idle or suspended state.

1. In the sidebar, go to Settings and select the Resource Pool tab.

2. Select a pool you wish to readjust.

   The pool you selected is now highlighted.

3. To add or reduce Workers, in the pool’s settings panel, select the number to the right of the Worker Count, and use the + or - key.

   When adding workers to a pool, ensure that there are enough available workers in your default pool to be reduced from.

4. To make this pool your default pool, go to the three-dot menu located in the top right corner of the pool panel, and choose Make Default.

5. To use this pool for ANALYZE TABLE go to the three-dot menu located in the top right corner of the pool panel, and choose Use For Statistics.

6. To rename pool, you may either:

   • Go to the three-dot menu located in the top right corner of the pool panel and choose Rename

   • Select the pool name in the pool’s settings panel and rename pool

7. To delete pool, go to the three-dot menu located in the top right corner of the pool panel and choose Delete.

Managing Pools within a Session

You can connect to a specific pool using third-party tools. Additionally, you have the ability to list all of your pools and shift between them within your current session as needed.

Syntax

The DESCRIBE [RESOURCE] POOLS is a CPU based SQL command that lists all of your pools.

This command requires CONNECT permission.

DESCRIBE [RESOURCE] POOLS
DESC [RESOURCE] POOLS

The USE [RESOURCE] POOL command lets you shift between pools within a session.

This command requires CONNECT permission.

USE [RESOURCE] POOL <pool_name>

Parameter	Description
pool_name	Specifies the name of a specific pool you wish to shift to within the current session

Using the Editor

You may also shift between pools within a session using the Editor.

In the left-hand side of the ribbon, select a pool from the Pool drop-down menu.

Examples

Listing all existing pools:

DESCRIBE RESOURCE POOLS;

Shifting between pools:

USE POOL bi_pool;

Using Third-Party Tools

Parameter	Description
pool	Specifies the name of a specific pool to connect to

Examples

Connecting to a specified pool:

sudo java -jar jdbc-console-0.0.88-43.jar --host=myhost.isqream.com --access-token=******* --pool=bi_pool

Connecting to a default pool:

sudo java -jar jdbc-console-0.0.88-43.jar --host=myhost.isqream.com --access-token=*******

Performance and Concurrency Preferences

Each pool in the system is associated with a parallelism policy that determines whether it prioritizes performance or high concurrency. By configuring different pools with distinct parallelism modes, you can optimize resource usage based on the specific needs of various organizational departments or work groups.

1. In the sidebar, go to Settings and select the Resource Pool tab.

2. Select a pool you wish to set.

   The pool you selected is now highlighted.

3. Under Parallelism Policy, select one of the policies:

   • No parallelism

   • Partial parallelism (50% of workers)

   • Maximum parallelism

Parallelism Policies

Policy	Description
No parallelism	A single query can be executed using only one Worker. It means that the query will run sequentially on a single worker, which may result in slower execution time compared to parallel execution.
Partial parallelism (50% of workers)	Utilizes 50% of the available Workers to execute a query. The query is divided among the selected Workers, enabling faster execution compared to the no parallelism policy, but not utilizing the full capacity of the worker pool.
Maximum parallelism	Enables the execution of a single query using multiple Workers. This allows the query to be divided among all available Workers, significantly reducing the execution time. It fully utilizes the Worker pool and provides the highest level of performance.