Help Center/ GaussDB(for MySQL)/ User Guide/ Parameter Management/ GaussDB(for MySQL) High-Performance Parameter Template
Updated on 2024-09-24 GMT+08:00
View PDF
Share

GaussDB(for MySQL) High-Performance Parameter Template

To improve database performance, GaussDB(for MySQL) provides a high-performance parameter template. You can select this template when buying an instance.

This section explains the parameter settings in the high-performance parameter template and how the template enhances performance.

Introduction

The high-performance parameter template is a set of optimized configuration parameters that aim to enhance the performance and reliability of database servers. This template can be adjusted based on different application scenarios and hardware configurations to achieve optimal performance and reliability.

The parameters in the high-performance parameter template are as follows.

Table 1 Parameter description

Parameter

Description

Value in a High-Performance Parameter Template

Value in the Default Parameter Template

innodb_flush_log_at_trx_commit

If this parameter is set to 0, logs are not flushed to disks when transactions are committed. Instead, they are only flushed once per second or when the log buffer (innodb_log_buffer_size) is full. This provides low durability but high performance.

0

1

rds_plan_cache

If this parameter is set to ON, an execution plan is cached. The next time the same query is executed, the cached execution plan can be reused, which improves the database's query performance.

ON

OFF

Application Scenarios and Potential Risks

Generally, the high-performance parameter template can improve database performance. However, it should be adjusted based on specific application scenarios and hardware configurations.

While the template is designed to enhance performance and reliability of database servers, it does come with some risks during database usage.

  • Setting innodb_flush_log_at_trx_commit to 0 can improve low-concurrency write performance, but in extreme cases, it may result in data loss of up to one second.
  • Setting rds_plan_cache to ON caches the execution plan of the PREPARE statement, which reduces the need for the optimizer to generate an execution plan and improves query performance. However, it may not be effective in all read/write scenarios.

Constraints

The kernel version of your GaussDB(for MySQL) instance must be 2.0.51.240300 or later.

Usage

You can select the high-performance parameter template when buying an instance.

Figure 1 Selecting the high-performance parameter template

Performance Comparison

Test environment:

  • GaussDB(for MySQL) instance specifications: Dedicated, 8 vCPUs | 32 GB
  • Kernel version: 2.0.51.240300
Sysbench test process:
  • Test scenarios: read-only, write-only, and read/write
  • Data volume: 250 tables x 25,000 rows of data for read-only or write-only and 25 tables x 250,000 rows of data for read/write
  • Performance metric: queries per second (QPS) with 1, 2, 4, 8, 16, 32, 64, 128, and 256 concurrent requests, that is, the number of SQL statements executed by the database per second

Test results:

  • Read-only model

  • Write-only model

  • Read/Write model

Conclusion: The preceding sysbench test results show that the high-performance parameter template significantly improves the database performance.

Parent Topic: Parameter Management