DCE 5.0 v0.19.0-20240630 Enterprise Released¶

This page explains the latest features, improvements and bug fixes, included in DCE 5.0 v0.19.0-20240630 Enterprise. DCE 5.0 modularizes each product module, and the versions included in this release are the default versions in the DCE 5.0 installer. If you need to independently upgrade or replace each module, please contact Daocloud's support team to download the proper versions and upgrade. Please also refer to DCE 5.0 Enterprise Deployment Requirements.

Included Modules¶

Features and Improvements¶

Detailed feature descriptions are reflected in the Release Notes of each version in the table above. Here are some brief descriptions of new features and improvements.

Installer v0.19.0¶

• Updated the default K8s version to v1.29.5
• Added support for configuring ubuntu_kernel_unattended_upgrades_disabled: true under the kubeanConfig parameter in clusterconfig.yaml to disable Ubuntu kernel automatic updates

Workbench v0.28.0¶

• Added Pipeline DAG optimization, supporting switching between new and old versions. Note: Saving pipelines across new and old versions is prohibited as it may cause compatibility issues
• Improved auto-completion for the jenkinsfile editor

Container Management v0.29.1¶

• Added support for the K3s Kubernetes distribution for connected clusters
• Added support for GPU card status monitoring, viewable via XDI metrics on the Insight platform

Kubean v0.16.3¶

• Updated the default K8s version to v1.29.5
• Added support to disable Ubuntu kernel automatic updates

Network v0.15.0¶

• Added Spidernet with a proper backend Spiderpool version of v0.9.3
• Fixed an error in comparing pointer values in the ListMultusCniConfig API, and corrected the mistake of writing sriov as ipvlan in the judgment

HwameiStor v0.14.7¶

• Improved affinity issues and added an option to skip affinity
• Improved Cleaned up the data set manager code

Virtual Machine v0.10.0¶

• Added support for real-time updates of virtual machine memory and CPU
• Added support for real-time migration of virtual machines to specified nodes

Multicloud Management v0.20.0¶

• Added support for selecting existing deployment strategies when creating multicloud resources
• Improved the display of multicloud resource distribution across various clusters

Container Registry v0.19.0¶

• Fixed an issue where workspaces could not be deleted due to a prompt requiring unbinding of registry spaces, even though the registry space was not bound to the workspace

Insight¶

Insight Server v0.28.0

• Added support for displaying calls to MySQL and PostgreSQL databases in the service topology map
• Fixed an issue where Elasticsearch-generated indexes did not have clusterid when logs were reported to Kafka cache

Insight Agent v0.28.0

• Added support for reporting logs to Elasticsearch v8.0
• Fixed the path exposure issue for fluent-bit serviceMonitor metrics

Microservices Engine v0.38.1¶

• Added API testing capabilities into the gateway API documentation
• Added custom data for gateway logs

Service Mesh v0.27.0¶

• Added Istio resource analysis to help identify resource configuration anomalies and improve user experience

Middleware¶

• Improved PostgreSQL support to v15.5
• Added support for configuring fixed IP addresses when creating Redis instances

AI Lab v0.6.0¶

• Added support for using vLLM as an inference engine when creating inference services, exposing native vLLM capabilities
• Added support for creating Notebook of type VisualStudio Code

Global Management v0.28.0¶

• Added a prompt to install insight-agent in billing configuration under operations management
• Added username and userLdpaFilter fields in LDAP configuration

Cloud Edge Collaboration v0.12.0¶

• Added support for creating device models
• Added support for dynamically updating download links for edge installation packages based on online or offline environments

Upgrade Notes¶

• Container Management: Before upgrading to v0.29.0, if the node is in MIG mode, the following actions are required:

  • Upgrade with Service Breaking:
    • Stop all GPU applications
    • Uninstall gpu-operator and nvidia-vgpu
    • Reinstall gpu-operator after the upgrade, ensuring the gpu-operator version is greater than v23.9.0+1
  • Zero-downtime Upgrade:
    • Manually modify the node label gpu.node.kpanda.io/nvidia-gpu-mode: mig
    • Upgrade the kpanda version
    • Upgrade the gpu-operator version to >= v23.9.0+1

• Workbench: In v0.28.2, after enabling the pipeline webhook, it cannot be disabled. A separate upgrade to v0.28.3 is required.

Compatibility¶

The versions supported by Global Management v0.28.0 are as follows:

• Workbench v0.17.3 and above
• Container Management v0.18.1 and above
• Multicloud Management v0.9.1 and above
• Container Registry v0.8.0 and above
• Insight v0.17.2 and above
• Microservice Engine v0.22.0 and above
• Service Mesh v0.16.2 and above
• Middleware RabbitMQ v0.11.0 and above
• Middleware MySQL v0.9.0 and above
• Middleware ElasticSearch v0.8.0 and above
• Middleware Redis v0.8.0 and above
• Middleware MinIO v0.6.0 and above
• Middleware Kafka v0.6.0 and above
• Middleware PostgreSQL v0.2.0 and above
• Middleware MongoDB v0.1.0 and above
• Middleware RocketMQ v0.2.0 and above

Known Issues¶

Known issues for installer v0.19.0, see Installation Troubleshooting.

Installation¶

• DCE 5.0 Enterprise Installation
• DCE 5.0 Community Installation

Download¶

• Go to download DCE 5.0 offline package

Comments