In April, YashanDB officially released the long-term support version YashanDB V23.2 LTS, marking that YashanDB's complete product system such as stand-alone active/standby, shared clusters, and distributed real-time data warehouses has fully entered the long-term support stage for large-scale use.
YashanDB V23.2 LTS has undergone rigorous and comprehensive testing, involving millions of test cases, hundreds of long-term stability and stress model tests, continuous operation for more than 30*24 hours, and reliability verification through hundreds of fault scenarios and millions of continuous extreme brute force faults. As a long-term support version for enterprise-level users, YashanDB V23.2 meets the stringent requirements of database systems in various business scenarios, especially core production scenarios, and is a milestone version for supporting core systems in key industries such as finance, energy, and government affairs.
In the new version, YashanDB enhances many enterprise-level capabilities and features, and greatly improves the performance, compatibility, ease of use, and availability of the product.
High performance: The performance of the two-node TPCC in the shared cluster reaches 312W
The performance of the shared cluster dual-node TPCC reaches 312W
23.2 As a milestone release, LTS shared clusters have large-scale commercial capabilities and provide transparent multi-write, high-availability, and high-performance database capabilities for high-end core business scenarios.
Based on the following environment configuration,
According to the actual measurement, in the client and server extension deployment mode, the TPCC performance of the two nodes with 1000 warehouses and 700 concurrent operation for 10 minutes reaches 312W, and based on this environment, the shared cluster product runs stably and passes the 7*24 hours long stability test.
The networking environment is as follows:
Client-server
CPU 96 Core/256G Mem/100Gb Nic/ SATA 500GB*1
Database servers
CPU 96 Core/512G Mem/100Gb Nic/NVMe 3.84TB*2
WDS Storage Server
CPU 96 Core/256G Mem/100Gb Nic/4 x NVMe 3.84TB
Network switches
100Gb/s switches
The storage configuration is as follows:
Device type
Distributed block storage products
Device model
华瑞指数云SDS2.0 WDS V3
Detailed parameters
HDD: 4 x NVMe 3.84TB
Port: 100Gb Nic
Cache capacity: 256GB cache
quantity
Servers: 3
Number of hard disks per drive: 4 (14TB capacity)
The incremental synchronization performance is improved by 8 times
For data warehouses, incremental data synchronization and warehousing capability is a very critical capability. 23.2 LTS optimizes this part of the capability to greatly improve the performance of incremental data synchronization and storage, and for the lineitem table model, the single-task synchronization performance is increased from 1 MB/s to 8 MB/s.
By optimizing and reducing the memory overhead of data synchronization tasks, concurrent data synchronization is improved, and the expansion ratio is close to 1.
The performance of synchronization tasks is improved by optimizing the synchronization task pipeline.
Write the incremental synchronous data directly to the steady-state data, so that the data storage has a high compression ratio, and at the same time, the data storage can obtain high query access performance.
Stronger compatibility: Oracle compatibility has been greatly improved
Based on V23.1, V23.2 LTS absorbs the actual requirements of multiple business scenarios such as banking, securities, energy, and government affairs, continuously improving the Oracle compatibility of the product, and allowing more business scenarios to use V23.2 LTS for smooth application migration.
The compatibility of syntax, semantics, and advanced features has been fully enhanced
The first is the data type aspect. V23.2 LTS has developed the float(n) type to provide users with higher accuracy floating-point number types. The syntax, numeric range, and significant digits of this data type are fully compatible with Oracle.
This is followed by the built-in premium package. V23.2 LTS comes with a series of software packages that provide PL functions and stored procedures for functions that are difficult to complete using SQL statements or PL statements, so as to facilitate complex data management and system O&M tasks. V23.2 LTS adds support for the following premium packages:
Premium Package
function
DBMS_SQL
Parse and execute dynamic SQL statements
DBMS_LOB
Create, read, and modify LOBs
DBMS_ROWID
EXTRACT THE INFORMATION FROM YOUR ROWID
Finally, in terms of procedural language (PL), V23.2 LTS adds Oracle-compatible batch capabilities, including BULK_COLLECT and FORALL, to improve the efficiency of batch statement execution in stored procedures.
Shared cluster kernel enhancements
V23.2 LTS shared clusters support DBLink, second-level partitioning, and automatic collection of statistics, as well as the creation of custom local temporary tablespaces and local swap tablespaces, providing instance-level temporary data access services, reducing information exchange between instances and improving performance in corresponding business scenarios.
External tables are supported, making data query more convenient
V23.2 LTS adds support for external tables and is syntactically compatible with Oracle 19C. When you create a read-only external table in YashanDB, you can specify a specific CSV file on the storage system or a DIRECTORY file of the data directory object, so that you can access the external table as if it were a normal table, instead of importing the external data into the database and processing it. When large-scale data processing and analysis are involved, external tables can be queried directly on the storage system, avoiding the overhead of data transmission, significantly reducing data storage costs, and improving query performance.
More complete high-availability capabilities
Backup and recovery capabilities have been enhanced
Data backup and recovery is the last line of defense for high data availability and is critical to the robust operation of your systems. The lack of an effective backup and redundancy mechanism can lead to the permanent loss and corruption of important data. V23.2 LTS shared clusters support yasrman backup and recovery tools to facilitate backup and backup management, as well as Point-In-Time Recovery (PITR) recovery and remote backup.
Data disaster recovery is enhanced
In addition, when the YFS metadata is corrupted, the V23.2 LTS shared cluster can obtain the backup data from the fast recovery area and perform automatic recovery. At the same time, in the case of multiple failgroups, it supports the ability of multiple copies of data, and users can verify and recover data through the multi-copy mechanism, enhancing the reliability of the system from many aspects.
Lower cost of ownership
Two replicas are deployed, reducing data storage costs by one-third
Data storage is the most expensive component of a data warehouse system, especially when dealing with big data. In order to reduce the cost of data storage in the high-availability mode of services, the DN group of V23.2 LTS introduces a high-availability deployment mode of one primary and one standby, so that the data storage cost can be reduced by one-third while the service obtains high availability capabilities, and the data storage cost is reduced by one-third. In addition, the deployment mode of one primary and one standby DN group also supports automatic failover and high availability to meet RTO
Better ease of use
V23.2 LTS has made significant improvements in ease of use, making O&M and data compression simpler and more efficient.
Easy O&M and efficient management
Chapters related to shared clusters (Global Cache Load Profile, Global Cache Efficiency Percentages, etc.) are added to the AWR report to provide statistics related to various clusters during operation, so that users can better understand the performance of shared clusters. At the same time, V23.2 optimizes cluster-related parameters, and some parameters adopt a self-tuning strategy to reduce user costs and implement self-management of background service threads in shared clusters.
Adaptive encoding provides efficient data compression
The system automatically detects and selects the best encoding type to achieve the best encoding compression effect, so that the user is not perceptible and greatly reduces the threshold for users.
One-click collection of operating system and database information is supported
In order to locate problems and optimize the system faster, YashanDB V23.2 provides one-click diagnostic information, including basic operating system information, operating system and database logs, database configuration information, and database views.
A more complete tool system
In this iteration, YashanDB has launched new development, migration, and O&M tools to implement full lifecycle management from installation and deployment, one-click migration, intelligent O&M, and application development, helping users simplify learning and O&M costs.
Supports migration of MySQL, DM8 and other databases
The Yashan Migration Platform (YMP) for V23.2 LTS adds offline migration links from MySQL and DM8 databases to YashanDB, as well as statistical verification, full verification, failure retries, batch rewrites, migration fault tolerance, and pause and resume.
Distributed visual O&M capabilities are added
V23.2 LTS implements distributed visual O&M capabilities, and uses YashanDB Cloud Manager (YCM) to manage YashanDB distributed databases, and currently supports visualization functions such as inspection management, monitoring and alarming, backup and restoration, and slow SQL analysis for stand-alone, shared clusters, and distributed full-product forms, implementing convenient O&M efficiency.
Developer Tools is new
V23.2 LTS is also equipped with the new developer tool YDC V1.0 (YashanDB Developer Center, YDC), which allows users to visually manage database objects such as tables, views, and functions in the database through YDC. It has a wealth of functions and tools, including database connection management, SQL editing and running, intelligent IDE, graphical object management, execution records, logs, result set display, and PL/SQL DUBUG.
Strengthen the ability to apply open source products
The sqlalchemy dialect package is supported, and the YashanDB database can be accessed through the sqlalchemy framework.
The Prometheus monitoring system is supported, and you can use the yashandb-exporter component to customize the collection of yashandb monitoring items.
You can use DataX, a data synchronization tool, to migrate data to a YashanDB database.
YashanDB geospatial data can be visualized through the GeoServer geospatial system.
YashanDB V23.2 Personal Edition (download.yashandb.com) has been released here, please visit the official website to learn more about the new and optimized features.