1

8.1.3.200 cluster rollout

The 8.1.3.200 kernel has been rolled out on a large scale and is used as the primary version.

• The GUC parameter ddl_lock_timeout is added to specify the DDL lock timeout interval.

• Added the views PGXC_WAIT_DETAIL and PGXC_LOCKWAIT_DETAIL.

• Columns including wait_on_pid and query_id are added.

• Optimized some other issues.

Commercial use

8.1.3

1

DAS connections to DWS databases

You can connect to a database through DAS.

DAS is a service on HUAWEI CLOUD that allows users to manage DB instances on a web console where users can perform basic SQL operations, advanced database management, and intelligent O&M, making work easy, secure, and intelligent.

Commercial use

Using DAS to Connect to a Cluster

2

Workload management for logical clusters

You can manage the workloads of logical clusters.

Commercial use

Managing Workloads (for Logical Clusters)

3

Connections from the Windows gsql clients to clusters

The database connection command line tool gsql supports Windows.

Commercial use

Using the Windows gsql Client to Connect to a Cluster

4

Cluster connections from the Python libraries psycopg2 and PyGreSQL

In GaussDB(DWS), you connect to a cluster from the Python libraries psycopg2 and PyGreSQL.

Commercial use

Using the Python Library psycopg2 to Connect to a Cluster

Using the Python Library PyGreSQL to Connect to a Cluster

5

8.1.3 kernel function enhancement: 8.1.3.110 cluster version

If a number is divided by zero, NULL will be returned. The compatibility option enable_division_by_zero_mysql is added to the behavior_compat_options parameter.

Commercial use

behavior_compat_options

1

Cluster versions 8.1.1.500 and 8.1.3.100

• 8.1.1.500: Improved features and fixed known issues.
• 8.1.3.100: Improved hint bit logging and MySQL compatibility.

Commercial use

8.1.3 Version Changes

1

Specifications Change

You can change your node specifications on the management console. For example, you can change to higher specifications before service peak hours to quickly improve the computing capability of the cluster, and change to lower specifications afterwards to reduce costs.

Commercial use

Modifying Specifications

2

Snapshot type display

You can check whether a snapshot is an incremental or full snapshot on the management console.

Commercial use

Viewing Snapshot Information

3

GUI improvement

The cluster details page is improved. The product specifications display on the pages for creating a data warehouse cluster and purchasing a discount package is improved.

Commercial use

Viewing Cluster Details

Creating a Cluster

1

Enhanced DMS

1. The cluster dashboard page is displayed. You can view the cluster status, real-time resource consumption, top SQL statements, cluster resource consumption, and key database metrics.
2. Query monitoring is changed to real-time query. It displays real-time information about all the queries and sessions running in the cluster.
3. Session monitoring is changed to historical query. It displays all the historical query information in the current cluster.
4. The DDL audit function is added. DDL audit is used to check whether DDL metadata is standard, helping users detect potential problems in table definition.
5. The SQL probe function is added. You can upload and verify SQL probes, execute probe tasks in one click, and periodically execute probe tasks.
6. The Workload Diagnosis Report (WDR) function collects and analyzes database performance data. You can create workload snapshots to record cluster workload data in a specified period.

Commercial use

Database Monitoring Overview

2

8.1.3 database kernel

8.1.3 database functions are supported.

Commercial use

8.1.3 Version Changes

3

Cluster scale-in

Cluster scale-in is supported.

Commercial use

Scaling In a Cluster

4

Concurrency configuration for scale-out

During scale-out, you can configure the number of concurrent redistribution tasks.

Commercial use

Cluster Scale-out

5

Support for 3.x MRS data source connections

You can create a data source connection for a user-defined MRS 3.0.x cluster.

Commercial use

Creating an MRS Data Source Connection

6

Logical cluster scale-out online

Logical clusters can be scaled out online.

Commercial use

Scaling Out Logical Clusters

7

Intelligent O&M

Intelligent O&M helps GaussDB(DWS) users with O&M tasks. With this feature, you can specify the proper time window and number of tasks to execute based on the cluster workload. Besides, Intelligent O&M can adjust task execution policies according to service changes in a timely manner to reduce the impact on services. Periodic tasks and one-off tasks are supported, and you can configure the time window as required.

Only clusters of 8.1.3 and later versions support this function.

Commercial use

Intelligent O&M

8

Offline redistribution scheduling

Cluster redistribution can be scheduled. Tables are redistributed only within the time window you configured.

Only clusters of 8.1.3 and later versions support this function.

Commercial use

Redistributing Data

9

Cluster capacity display

You can check the snapshot capacity, monitoring statistics, and cold data capacity.

Commercial use

Viewing Basic Cluster Information

10

Interconnection with Log Tank Service (LTS)

You can view the collected cluster logs or dump logs on LTS.

Commercial use

Log Management Overview

11

Resource pool queue congestion alarm

Alarms can be generated for resource pool queue congestions.

Commercial use

Alarm Management

1

Disk capacity expansion

You can expand the disk capacity of cloud data warehouse clusters on the console.

Commercial use

Disk Capacity Expansion

2

The hybrid data warehouse supports distributed deployment.

The hybrid data warehouse supports distributed deployment, and provides the same capabilities as the cloud data warehouse does. The capabilities of a standalone data warehouse remain unchanged.

Commercial use

Data Warehouse Types

1

Real-time data warehouse

• The data warehouse provides high-concurrency, high-performance, and low-latency transaction processing capabilities based on large-scale data query and analysis capabilities. It is suitable for hybrid transaction/analytical processing (HTAP). A database can be used for both production and analysis. Currently, HA features are not supported. It can be deployed only in standalone mode, which is cost-effective and can be used for lightweight, non-core services, such as learning and testing.
• Changed the data warehouse name from "real-time data warehouse" in OBT to "stream data warehouse".

Commercial use

Data Warehouse Types

2

Switchback

In a cluster in Unbalanced state, there are too many primary instances on part of the nodes, making the nodes heavily loaded. In this case, the cluster can still run properly, but its performance is not as good as that in the balanced state. In this case, you can perform a primary/standby switchback for the cluster during off-peak hours to improve performance.

Commercial use

Performing a Primary/Standby Switchback

3

Stream data warehouses can be bound to Elastic Load Balance (ELB).

Stream data warehouses can be bound to Elastic Load Balance (ELB) to avoid single-CN failures.

Open beta testing

Associating and Disassociating ELB

4

Extended the range of GUC parameters supported by the console.

New GUC parameters can be modified on the console.

Commercial use

Modifying Database Parameters

1

Database monitoring

Fine-grained database monitoring on nodes, performance, databases, sessions, queries, instances, and load metrics.

Commercial use

Database Monitoring Overview

2

Phase-based scaling

Added the phase-based scale-out function. The phases include node addition, cluster scale-out, and database redistribution.

• Note addition: This function is more suited for large-scale scale-out. Nodes can be added in batches in advance.
• Cluster scaling: The added nodes can be used for scale-out.
• Data redistribution: Data in existing nodes is evenly allocated to the new nodes after you scale out a cluster. Data redistribution is a time-consuming yet crucial task that accelerates service response.

Commercial use

Node Management

Cluster Scale-out

Redistributing Data

3

Clusters can be bound to Elastic Load Balance (ELB).

If the internal IP address or EIP of a CN is used to connect to a cluster, the failure of this CN will lead to cluster connection failure. If a private domain name is used for connection, connection failures can be avoided by polling. However, private domain names cannot be used for public network access, and requests cannot be forwarded in the case of a CN failure. Therefore, ELB is used to avoid single CN failures.

Commercial use

Associating and Disassociating ELB

4

More GUC parameters can be configured on the console.

More GUC parameters can be modified on the console.

Commercial use

Modifying Database Parameters

5

Alarm

Added the alarm management function to allow users to view alarms triggered by DMS on the console.

Commercial use

Alarm Management

1

8.1.1.100 database kernel

• Query bands in resource pool queues
• Slow node detection
• Fine-grained backup and restoration
• Full build optimization
• Supported primary keys in column storage.
• Supported single-column CU update in column-based storage, greatly improving update performance.
• Supported reused free space on data pages and optimized the performance of the AUTOVACUUM index.
• Provided a tool to automatically deploy GDS. GDS monitors traffic.
• Added the feature of storing cold data in internal OBS tables.
• Added table-level authorization of ANALYZE.
• Supported SM4 CBC.
• Supported WDRs.
• Added the features that the base table can be modified directly and related views support automatic update and cascade update.
• Enhanced the log audit function. Transactions and O&M operations are audited. DDL operations are audited only once. Internal database operations can be filtered.
• Enhanced the data masking function. Data masking rules can be customized.
• After the password expires, a user can log in to the database only after changing the password.
• Improved SQL performance: Enlarged hash tables to reduce conflicts, changed local redistribution to batch copy, and optimized row2vec.
• Disabled SMP adaptation by default. Related parameters can be adjusted based on the cluster flavor for optimal performance.
• Improved MySQL compatibility: Supported UPSERT, distinguished between NULL and empty strings, and supported limit offset, count, and concat functions.
• Returning NULL when TO_DATE and TO_TIMESTAMP process empty strings
• Supported INSERT OVERWRITE during online scale-out, which does not block services. When online scale-out is performed, no error is reported for adding the DROP column during partition exchange.
• Adapted to DLI foreign table modification.
• Hot loading of MRS data sources
• Added the percentile_cont and percentile_disc functions.

Commercial use

8.1.1

2

CN management

The number of manageable CNs increased from 5 to 20. CNs can be added or deleted on the console.

Commercial use

CNs

3

Online scale-out

During cluster scale-out, data can be continuously imported to the database.

Commercial use

Cluster Scale-out

4

Protection for mission-critical operations

Protected mission-critical operations. A credential for identity verification is required if users want to perform a mission-critical operation on the management console. For account security, you are advised to enable operation protection. The setting will take effect for both the account and its users.

Commercial use

Protection for Mission-Critical Operations

5

Dual-cluster DR

A homogeneous GaussDB(DWS) DR cluster is deployed in another AZ. If the production cluster fails to provide read and write services due to natural disasters in the specified region or cluster internal faults, the DR cluster becomes the production cluster to ensure service continuity.

Commercial use

DR Overview

1

8.1.0 kernel version

Made the following changes:

• Optimized cost estimation.
• Enhanced ANALYZE for temporary tables and single-transaction operations.
• Improved performance of the vectorized executor.
• Alleviated the problem that partition pruning can be performed only for constants but not the variables that need to be calculated.
• Added schema-level full backup and table- and schema-level fine-grained restoration.
• Supported third-party user-defined data conversion of GDS.
• Optimized RTO.
• Reconstructed CM for feature enhancement and interface modification.
• Supported data export in ORC format using OBS foreign tables. Foreign tables can read Parquet data.
• Supported XML function parsing.
• Supported automatic job migration when a CN fails.
• Supported smooth upgrade of C functions.
• Synchronized the time zone with the IANA Time Zone Database (tz).
• Supported Teradata DSQL compatibility of gsql.

Commercial use

8.1.0

2

Logical cluster

A physical cluster can be divided into logical clusters that use the node-group mechanism.

Users can associate a logical cluster to create tables. Tables in a database can be allocated to different physical nodes by the logical cluster, and a logical cluster can contain tables of multiple databases.

Commercial use

Logical Cluster Management

1

Optimized flavor page of second-generation clusters

Cloud or standard data warehouses can be selected during cluster creation.

• Standard: User-defined storage class or capacity is not supported.
• Cloud: The storage class and capacity of EVS disks can be set. CDR storage of EVS disks is billed independently in the second generation cluster.

Commercial use

Creating a Cluster

2

Optimized discount package page

Supported monthly and yearly discount packages.

Commercial use

Purchasing a Discount Package

1

Added the second generation of the EVS storage type. Users can set the storage type and storage capacity.

When creating a cluster, you can select Second Generation as required. This flavor is used for EVS disks. You can set the storage type and storage capacity, and storage resources are billed independently.

Commercial use

Creating a Cluster

1

gs_dump, gs_dumpall

• gs_dump: Exports a single database or its objects.
• gs_dumpall: Exports all databases in a cluster or common global objects each database.

Commercial use

Using gs_dump and gs_dumpall to Export Data

2

Workload management

Workload management enables concurrent and independent execution of two types of jobs.

Commercial use

Managing Cluster Workloads

1

Fine-grained permission control

The IAM service provides fine-grained access control.

On IAM, the administrator (that is, a user in the admin user group) can create a custom policy containing various required permissions. After a policy is granted to a user group, users in the group can obtain all permissions defined by the policy.

Commercial use

Fine-Grained Policies

2

Canceling read-only status on the console

The read-only status of a cluster can be canceled on the management console. Only v8.0.1.100 or later support this feature.

Commercial use

Canceling Read-only Status

1

Added the request parameter Enterprise Project.

Added the Enterprise Project parameter to the API.

Commercial use

Creating a Cluster

1

Increase in automatic snapshot frequency

By default, automated snapshots are token three times each day, starting from 00:00:00 (UTC). You can set the snapshot execution period during cluster creation or modify the existing automated snapshot policy.

Commercial use

Configuring an Automated Snapshot Policy

2

Support for manual snapshot creation

In the cluster list, you can click More in the Operation column of a cluster to select the Create Snapshot operation to manually create a snapshot.

Commercial use

Manually Creating Snapshots

1

Transparent encryption and decryption

In GaussDB(DWS), you can enable database encryption during cluster creation to protect static data. After you enable encryption, data of the cluster and its snapshots is encrypted.

Commercial use

Encrypting GaussDB(DWS) Databases