Release Notes

This chapter describes the component versions and major releases of ZStack Cloud Foundation (ZCF).

ZStack Cloud Foundation 1.2.0

On July 24, 2026, ZStack Cloud Foundation (ZCF) 1.2.0 was officially released.

Components and Versions

ZCF 1.2.0 includes the following components and versions:

Highlights

ZStack Cloud

• Management Network Supports IPv6: Supports platform deployment and O&M in IPv6 network environments, meeting more network evolution requirements.
• Enhanced Storage Performance and Volume Cache: Optimizes the ZBS volume I/O path to improve VM disk read and write performance. Supports host-local I/O cache to reduce access overhead in high-frequency I/O scenarios.
• Optimized Resource Alarms and Recovery Mechanism: Refines resource alarm monitoring granularity and recovery logic, improving resource exception tracking and handling efficiency.
• Improved Windows Runtime Stability: Optimizes Windows nested virtualization to reduce blue screen risks in scenarios such as Windows Server 2025 and improve runtime stability for newer Windows workloads on the platform.

ZStack ZStone

• Enhanced Replica Policies for High-Performance Storage Pools: Supports creating two-replica high-performance storage pools and switching between two and three replicas, allowing users to balance data reliability and storage utilization more flexibly.
• Read-Only Access for License Exceptions: Allows users to view resources and runtime status when a license is abnormal or expired, so they can continue troubleshooting before authorization is restored.

Unified Portal

• Access Control Overview and Role Authorization: Supports global account system maintenance and centralized management of accounts, roles, and permissions, making it easier to adjust permission boundaries as organizations and services change.

Cloud Federation

• Expanded Platform Connection Scope: Supports connecting ZSphere virtualization platform and ZNS Network Services. Administrators can centrally view runtime status and detailed monitoring data in Observability, reducing cross-system checking and troubleshooting and improving O&M efficiency.
• Security Service Provisioning and Maintenance: Supports one-stop provisioning, deployment, and management of Anheng security services, providing security capabilities such as Log Audit, Bastion, and Host Security EDR.

Observability

• Component Health: Centrally displays the status of core services, functional modules, and external dependencies, helping administrators quickly understand platform health and identify runtime risks.
• Alarm Center: Provides unified management of alarm messages, alarm rules, and notification configurations, improving risk management and troubleshooting efficiency.

Network Services

• Enhanced Network Access Capabilities: Supports unified maintenance of Layer 4 load balancing and NAT rules, helping administrators manage service entries, traffic distribution, and north-south access paths more clearly.

Installation and Deployment

• Initial Installation: Supports installation progress tracking and allows installation to resume from the interruption point after a failure, helping administrators quickly locate installation exceptions, reduce repeated steps, and improve initial delivery efficiency.
• Upgrade: Provides end-to-end visualized upgrade capabilities, including Bundle management, environment backup, upgrade precheck, and upgrade execution, improving upgrade reliability and operational efficiency.

New Features and Enhancements

This section describes the new and enhanced features in ZCF 1.2.0 by component and capability area.

ZStack Cloud

In ZCF 1.2.0, ZStack Cloud includes the following major updates:

VM Instances

Improved Windows VM Runtime Stability

Windows nested virtualization is optimized. Nested virtualization can be enabled or disabled globally or for an individual VM instance. This switch is disabled by default to improve compatibility and runtime stability for newer workloads such as Windows Server 2025 and reduce risks such as blue screens. Users can enable it as needed.

Cloud Networking

Management Network Supports IPv6

IPv6 addresses can be configured for the Management Network, allowing platform deployment and O&M in IPv6 network environments and meeting more network evolution requirements.

Cloud Storage

Optimized ZBS Volume Read/Write Performance

Volume multi-queue can be configured globally or at the cluster, VM instance, or volume level. A dedicated I/O thread can also be enabled for an individual volume on ZBS Primary Storage as needed. These enhancements optimize the volume I/O path and performance and improve storage access efficiency in high-I/O scenarios.

More Storage Combinations in a Single Cluster

A single cluster can attach Ceph + SharedBlock + ZBS (CBD protocol) Primary Storage, or Ceph + Vhost (XINFINI) Primary Storage, further improving flexibility in resource planning and service deployment.

Vhost (XINFINI) Supports VM Scheduling Policies

Clusters attached to Vhost (XINFINI) Primary Storage can be configured with VM scheduling policies, further improving resource scheduling flexibility in Vhost (XINFINI) storage scenarios.

Host-Local I/O Cache

Host-local I/O cache uses local high-performance disks such as SSDs or NVMe drives on hosts as cache media and organizes them into cache pools, providing cache acceleration for volumes on ZBS (CBD protocol), Ceph, or Local Storage. Volume data remains persistently stored on the corresponding primary storage. This feature applies to service scenarios that are sensitive to I/O performance and access latency, such as databases and high-performance computing. Administrators can centrally manage cache pool capacity, health status, and lifecycle at the host level and enable local I/O cache for volumes as needed, improving storage access efficiency in frequent read/write scenarios.

Unified Access to ZBS Primary Storage

ZBS is connected to the cloud platform as a unified Primary Storage type. The former CBD Primary Storage capabilities are merged into ZBS Primary Storage and continue to be used through the CBD protocol. ZBS Primary Storage also adds a Vhost protocol path. Users can connect storage resources through the unified ZBS Primary Storage entry and select the storage protocol that best fits the service workload when creating VM instances or volumes.

With this capability, ZBS Primary Storage can cover VM instance and volume scenarios over different I/O paths, helping users obtain more flexible performance and compatibility options in a unified storage resource pool. The platform also provides a connectivity view between ZBS Primary Storage and hosts, helping administrators view protocol link status and quickly locate storage access issues.

Monitoring and Alarms

Dashboard Displays Resource Utilization Statistics

The cloud platform dashboard adds resource utilization statistics, helping users view real-time usage of key resources at a glance. Currently, VM memory utilization (internal monitoring) and VM memory utilization (external monitoring) are supported. To view internal monitoring data, install GuestTools on the VM instance first.

Optimized Resource Alarm Recovery Notifications

The resource alarm recovery mechanism is optimized. Users can set recovery conditions based on duration and the percentage of normal monitoring points. The system collects monitoring data at fixed intervals and calculates the percentage of monitoring points that do not meet the alarm condition within the specified duration. When this percentage reaches the specified threshold, the resource is considered recovered from the alarm state. This method effectively avoids misleading results caused by metric fluctuations and further improves the accuracy of fault recovery assessment.

Users can also select Alarm Messages > Set Silence Period to suppress alarm and recovery notifications triggered by an individual resource for a specified period. This operation does not affect other resources monitored by the same alarm, improving O&M flexibility while avoiding monitoring blind spots.

Refined Resource Alarm Monitoring Granularity

The alarm details page now displays the alarm trigger status of each resource monitored by the alarm, providing a more granular view for monitoring and troubleshooting.

Installation and Deployment

Alibaba Cloud Linux 4 ISO Supports GreatDB

The x86 edition of Alibaba Cloud Linux 4 ISO supports GreatDB. The ARM edition of Alibaba Cloud Linux 4 ISO supports GreatDB and MySQL.

Experience Optimization

Bulk Project Member Management

Project members can be added in batches and assigned roles in a single operation, reducing permission configuration costs and improving team collaboration efficiency.

Intelligent Deduplication of VM Console Tabs

Multiple console tabs for a single VM instance can be intelligently identified and deduplicated, ensuring that at most one console page is displayed for one VM instance and keeping the tab view simple and easy to use.

ZStack ZStone

In ZCF 1.2.0, ZStack ZStone includes the following major updates:

Storage Resources

High-Performance Storage Pools Support Two Replicas

When creating a high-performance storage pool, users can choose two replicas or three replicas based on actual service requirements. Three replicas are selected by default. Compared with three replicas, two replicas reduce data redundancy overhead and improve storage utilization.

Switch Between Two and Three Replicas

For an existing high-performance storage pool in the Healthy state, users can select Modify Data Protection Policy on the storage pool list page to change the replica count from two to three, or from three to two.

Changing the replica count triggers data migration in the storage pool. During migration, the storage pool enters the Recovering state and returns to Healthy after migration is complete. Data migration may affect service performance. Perform this operation outside of peak service hours.

After replica switching is complete, users can still expand or shrink the high-performance storage pool based on the current replica policy. The platform checks whether the remaining data disks and logical failure domains meet the requirements of the current replica count, preventing operations from degrading the storage pool.

Hardware Resources

Delete Empty High-Performance Storage Clusters

A Delete action is added on the Hardware Resources > Clusters > High-Performance Storage page. When a high-performance storage cluster contains no servers, users can delete the cluster directly in the UI.

If servers remain in the cluster, the platform disables the delete action and prompts users to remove the servers first. Deleting a cluster clears the configuration information of the high-performance storage cluster and cannot be undone. Before deletion, users must confirm the operation risks and complete login password verification.

This feature can be used to clean up high-performance storage clusters that were not fully initialized or whose servers have all been removed, making it easier for users to replan and initialize the high-performance storage environment.

License Management

Read-Only Access When the License Is Abnormal or Expired

When the license status is abnormal or expired, users can still access the ZStack ZStone UI, browse feature pages, and view resource lists, resource details, runtime status, and other information.

After the license becomes abnormal or expires, the platform enters read-only mode. Only query and browsing operations are supported. Operations that change platform configuration or resource status, such as creating, modifying, or deleting resources and handling alarms, are not supported. When users initiate such operations, the platform rejects the request and prompts them to check the license status.

By retaining read-only access, users can continue viewing cluster and storage resource runtime status before the license is restored, and can troubleshoot issues and handle authorization in time.

Unified Portal

In ZCF 1.2.0, Unified Portal includes the following major updates:

Unified Authentication Configuration Management

Unified authentication configuration management is enhanced. Administrators can enable unified authentication for connected platforms in Cloud Federation > SSO Configuration. After configuration is complete, accounts in Unified Portal can access components through the unified identity authentication entry, reducing duplicate account maintenance and repeated sign-ins.

Access Control and Role Authorization Management

Global access control and role authorization management are added. After signing in through the unified identity authentication entry, administrators can centrally maintain accounts, roles, and permissions in Global Management > Access Control. Roles and permissions can be inherited by user group, project, or organization, making it easier to adjust permission boundaries as organizations and services change.

Cloud Federation

In ZCF 1.2.0, Cloud Federation includes the following major updates:

Connect ZSphere Virtualization Platform

ZSphere virtualization platform connection is added. Administrators can add ZSphere and verify the connection in Cloud Federation > Platform Management. After connection, ZCF can display O&M data such as assets, monitoring, logs, and alarms from ZSphere, helping administrators view virtualization platform runtime status and locate issues in a unified view.

Connect ZNS Network Services

ZNS Network Services connection is added. Administrators can add ZNS and verify the connection in Cloud Federation > Platform Management. After connection, administrators can enable asset collection as needed. ZCF can display basic network assets such as Transport Nodes, Edge Nodes, Edge Clusters, Segments, Tier-0 Gateways, and Tier-1 Gateways, providing a data foundation for network inventory and subsequent observability analysis.

Security Service Provisioning and Maintenance

Based on the existing Anheng Tianchi connection, security service provisioning and maintenance are added, covering Log Audit, Bastion, and Host Security EDR. On the Security Services page in the default zone, administrators can select security capabilities, confirm specifications, and initiate provisioning. After provisioning, administrators can enter the corresponding security service and perform maintenance operations such as upgrade, renewal, or unsubscription, adding audit, O&M access, and host protection capabilities for VM instances.

Observability

In ZCF 1.2.0, Observability includes the following major updates:

Observable Data

Platform Component Health

A platform component health view is added. Administrators can view the status of ZCF core services, monitoring collection, alarm evaluation, notification pipelines, capacity risks, and basic dependencies in Operations > Monitoring Center > Platform Health, so they can quickly assess current platform health and identify risks that require attention.

Custom Dashboards

Custom dashboards are supported. Users can flexibly select components and adjust layouts to design monitoring views that better match actual service scenarios. Custom dashboards can be cloned, imported, and exported for quick reuse and lower repeated configuration costs.

Expanded Data Sources

Assets, monitoring, logs, and alarm data from ZSphere virtualization platform, as well as basic network assets from ZNS Network Services, can now be connected. Administrators can view related data in Dashboards and Reports and Operations, reducing entry switching during resource inventory, status checking, and troubleshooting.

Alarm Center

Resource Alarm Rules

For resource monitoring, built-in alarm rules are provided to cover core resources and common risk scenarios, helping administrators promptly identify capacity, performance, and service workload exceptions. Custom alarm rules are also supported. Administrators can flexibly select monitoring items and adjust alarm severity, trigger conditions, and alarm scope to meet risk identification needs in different service scenarios and O&M policies.

Event Alarm Integration

Event alarm integration can be enabled to report event alarms from connected components to ZCF Alarm Center for centralized viewing and handling.

Alarm Message Push

Alarm messages can be pushed to notification channels such as DingTalk, WeCom, email, Webhook, and SMS, adapting to different message receiving habits and reducing the risk of missing important alarms.

Administrators can use system templates or create custom templates to define notification formats and content. They can also use notification policies to filter alarms by severity, source, or resource type before pushing them to channels, improving message relevance and helping O&M personnel focus on information related to their responsibilities.

Alarm Handling and Tracking

The Alarm Messages page supports viewing alarms by status, such as Active, Recovered, and Historical, and records affected objects, trigger conditions, status timelines, and notification delivery results. This helps administrators confirm alarm handling progress and message delivery status.

Network Services

In ZCF 1.2.0, Network Services include the following major updates:

Layer 4 Load Balancing

Layer 4 load balancing is added. In ZNS Network Center, administrators can configure virtual services, backend pools, backend members, and health checks for Tier-1 Gateways, providing a unified access entry and traffic distribution for TCP/UDP services. The system can identify backend member health status and check VIP and NAT address usage, helping administrators reduce forwarding risks in entry configuration, backend maintenance, and fault location.

NAT Rule Management

NAT rule management is added, covering SNAT, DNAT, and Reflexive NAT. In ZNS Network Center, administrators can maintain private network outbound access, external access to internal services, and one-to-one address mapping rules by route gateway, and view rule status and recent operation information. The system supports address and port usage checks, helping administrators reduce rule conflicts and access path configuration errors.

Installation and Deployment

In ZCF 1.2.0, Installation and Deployment include the following major updates:

Installation Progress Tracking and Failed-Step Retry

The installation and deployment wizard enhances progress display and failure handling. During installation, users can view the execution status of each stage. If installation fails, users can rerun installation from the interruption point based on the failure information, reducing the time spent repeating the full installation process.

Upgrade Package Management and Visualized Upgrade

The Lifecycle Management module is added, providing package management and visualized upgrade capabilities. Administrators can upload and manage ZCF Bundles and create upgrade plans. The system advances the upgrade process according to the plan, connecting upgrade package selection, environment backup, risk precheck, upgrade execution, and historical traceability. This helps administrators assess the impact before an upgrade, continuously track status during the upgrade, and quickly locate recovery context when exceptions occur.

Resolved Issues

This section describes the issues resolved in ZCF 1.2.0 by component and capability area.

ZStack Cloud

In ZCF 1.2.0, ZStack Cloud includes the following resolved issues:

Security

OpenSSH, Python, Curl, and HAProxy security vulnerabilities are fixed, and protection for sensitive information and user sign-in security is further improved.

Jira: ZSTAC-85902/ZSTAC-81047/ZSTAC-85009/ZSTAC-85743/ZSTAC-85787/ZSTAC-84651/ZSTAC-72939