v3.11.x
v3.11.12
date: 2025-10-20
The highlights of this v3.11.12 update are:
-
Comprehensive support for online deployment of mainstream Linux server distributions on both x86 and arm architectures. The detailed list is as follows. Due to insufficient testing time, if you encounter problems, please provide feedback via issues.
- openEuler: 22.03 LTS SP3, 22.03 LTS SP4, 24.03 LTS SP2
- CentOS: 7.9, 8 stream, 9 stream, 10 stream
- Debian: 11, 12, 13
- Ubuntu Server: 20.04 LTS, 22.04 LTS, 24.04 LTS, 25.04
- AnolisOS: 8.8, 8.10
- OpenCloudOS: 8.8, 8.10, 9.2, 9.4
- Rocky Linux: 8.x, 9.x, 10.x
- Alma Linux: 8.x, 9.x, 10.x
-
Built-in private cloud improvements:
- VMs support deployment and expansion of LVM disks, supporting online expansion of LVM disks
- Automatically capture VM screenshots when VM PANIC occurs, facilitating subsequent troubleshooting
- Host host service supports cgroup v2 API
-
Cloud management main improvements:
- Upgraded AWS SDK to 2.0, supporting AWS CUR 2.0 bills
- Support host automatic snapshot policies
Feature Improvements
- 【Host】Automatically initialize VM sshd configuration, allowing root password login by default
- 【Host】Support operating systems using NetworkManager for network management, including CentOS Stream 10, Rocky Linux 10.x, Alma Linux 10.x
- 【Host】Host service on hosts supports cgroup v2 interface
- 【Host】Host list resources sorted by utilization or allocation rate
- 【Host】VMs created from host backups support modifying disk types
- 【Host】Do not validate image boot method when creating VMs from ISO images
- 【Host】Execute VM NIC deletion script when deleting VMs to prevent VM port accumulation issues
- 【Cloud Management】Optimized loading speed of public cloud image lists
- 【Cloud Management】MongoDB displays domain addresses
- 【Cloud Management】VMware disks support unmounting
- 【Cloud Management】Support Tencent Cloud SQL Server
- 【Bare Metal】Optimized grub configuration file delivered via PXE, detecting as many boot disks as possible
- 【Image】System images support managing public cloud system images
- 【Image】Support detecting images that support both UEFI and BIOS boot
- 【Monitoring】Added Tencent Cloud disk utilization monitoring
- 【Frontend】Some resources add list status statistics
- 【Deployment】Disable apimap service by default to prevent slow queries due to excessive database data
- 【Billing】Support AWS CUR 2.0 bills (Enterprise Edition)
- 【Billing】Multi-dimensional analysis charts dynamically calculate whether horizontal axis labels should be displayed at an angle to accommodate more displayable information (Enterprise Edition)
- 【Work Order】Optimized work order query and notification approver performance (Enterprise Edition)
- 【Deployment】Support offline deployment of OpenEuler 22.03 LTS SP4 (Enterprise Edition)
Bug Fixes
- 【Host】Fixed issue where missing or incorrect fields in old version VM description files caused host live migration failures
- 【Host】Fixed issue where default routes for VPC networks with multiple NICs were not correctly set
- 【Host】Fixed issue where host service failed to start on machines without numa nodes
- 【Cloud Management】Fixed issue where Huawei Cloud could only synchronize 1000 instances
- 【Cloud Management】Fixed issue where esxi only unregistered but did not completely delete resources when deleting instances
- 【Cloud Management】Fixed resource misdeletion issue caused by conflict between local expiration release function and resource synchronization
- 【Cloud Management】Fixed DNS resolution domain validation rules
- 【Cloud Management】Fixed pay-as-you-go machine expiration release parameter issue
- 【Bare Metal】Fixed issue where arm64 bare metal failed to create megaraid logical disks
- 【Scheduling】Fixed scheduling failure caused by NIC and SKU information cache not being refreshed in time
- 【Authentication】Automatically clean up large amounts of historical passwords generated due to operator BUG
- 【Notification】Fixed issue where Enterprise WeChat token expiration caused message sending failures
- 【Monitoring】Fixed monitoring aggregation parameter issues by domain and project
- 【Monitoring】Fixed issue where alert resource count in monitoring overview was sometimes empty
- 【Image】Fixed upload progress display error for system image uploads, changed to get from request
- 【Image】Fixed issue where some images' UEFI boot was detected as BIOS boot
- 【Frontend】Fixed issue where monitoring alert icons in navigation bar were not permission-checked
- 【Frontend】Fixed inconsistent favicon.ico display issue
- 【Frontend】Fixed issue where some documentation links jumped to wrong addresses
- 【Frontend】Disable webconsole referer check by default to avoid webconsole being unusable after upgrade
- 【Deployment】Fixed OpenEuler LVM system disk initrd boot failure issue
- 【Deployment】Avoid repeated password modifications after enabling -sync-user, preventing accumulation of large amounts of historical password modification records
- 【Billing】Fixed Google Cloud rerun bill API error issue (Enterprise Edition)
- 【Billing】Fixed issue where switching bills in the upper right corner of some billing pages could not be clicked (Enterprise Edition)
- 【Billing】Fixed Alibaba Cloud International bill billing type exception (Enterprise Edition)
- 【Work Order】Fixed issue where work orders got stuck due to shopping cart status changes (Enterprise Edition)
- 【Authorization】Fixed issue where popup was not closed after authorization completion (Enterprise Edition)
New Features
- 【Host】Support host automatic snapshot policies
- 【Host】Automatically save screenshots when VM operating system kernel PANIC occurs
- 【Host】Support deployment and expansion of LVM disks, supporting online expansion of LVM disks
- 【Host】SSH key pairs support ED25519/ECDSA format public keys
CHANGELOG
For specific changes, please refer to:
v3.11.11
date: 2025-07-20
The highlights of this v3.11.11 update are:
- Frontend resource list performance optimization, large page data rendered in batches, reducing page wait time
- Improved lockman lock mechanism, improving lock stability in high concurrency scenarios
- When adding NICs or changing IP addresses to running built-in virtualization VMs, automatically configure NICs via QGA, taking effect in real time
- Bare metal supports automatically joining VLAN subnets, achieving VLAN-based bare metal network isolation
- User query speed optimization after granting organizational structure tree permissions, accelerating more than 10 times (Enterprise Edition)
Feature Improvements
- 【Frontend】List optimization, large page data rendered in batches, reducing page wait time
- 【Frontend】Browser page title displays current menu name, convenient for distinguishing when browser opens multiple tabs
- 【Frontend】Dashboard resource jump buttons add permission checks
- 【Platform】Improved lockman lock mechanism, improving lock stability in high concurrency scenarios
- 【Monitoring】Optimized display format for multiple alert data
- 【Monitoring】Monitoring queries support in and not in filters, allowing batch queries of monitoring data for multiple resources
- 【Monitoring】Support cloning monitoring panels
- 【Cloud Account】Cloud account synchronization supports image types
- 【User】Add secondary confirmation when deleting system users to improve security
- 【Host】When creating VMs, less commonly used configurations for disks and NICs are moved to "Advanced" menu, can be expanded when needed
- 【Host】WebRDP adds password verification
- 【Host】Details add manual/automatic renewal display
- 【Host】When adding NICs or changing IP addresses to running built-in virtualization VMs, automatically configure NICs via QGA, taking effect in real time
- 【Host】Support Tencent Cloud extreme speed and general SSD cloud disks
- 【Host】VMs support TencentOS operating system
- 【Host】Support using VM labels for new disks when adjusting public cloud VM configurations
- 【Host】Support adding disks when Alibaba Cloud VMs are powered on
- 【Host】Supplement AWS disk specifications, disk capacity adds maximum/minimum limits
- 【Host】Support synchronizing Kingsoft Cloud RDS resources
- 【Message Notification】Support DingTalk robot signature
- 【Storage】Block storage list resources add utilization display
- 【MongoDB】Display domain addresses
- 【Deployment】Newly deployed clusters' onecloud namespace etcd enables TLS by default
- 【Deployment】Support Ubuntu 20.04 x86_64 and openEuler 24.03 LTS SP1 x86_64
- 【Authentication】User query speed optimization after granting organizational structure tree permissions, accelerating more than 10 times (Enterprise Edition)
- 【Bastion Host】Bastion host-VM adds information display and deletion operations (Enterprise Edition)
- 【Bastion Host】VMs support joining bastion host and removing from bastion host operations (Enterprise Edition)
- 【Cloud User】Cloud users support "Enable/Disable Console Login" (Enterprise Edition)
- 【Report】Report types add category distinction (Enterprise Edition)
- 【Work Order】Host work orders add information display; host configuration adjustment work order tables support adjusting column width (Enterprise Edition)
- 【Billing】Cost-project list supports modifying resource labels (Enterprise Edition)
- 【Billing】Optimized GCP cost pulling to avoid full table query of cost data (Enterprise Edition)
- 【Billing】Bill summary adds cloud subscription level (Enterprise Edition)
Bug Fixes
- 【Frontend】Fixed issue where some detail pages opened with content area too small
- 【Platform】mysql database utf8mb4 fields do not support storing emoji characters
- 【Authentication】Optimized issue where too many system service password updates caused authentication timeout
- 【Authentication】Fixed issue where "Other" permission was incorrectly backfilled when modifying permissions
- 【Authentication】Fixed issue where users were not properly obtained when creating passwordless login users
- 【Scheduling】Fixed uneven VM allocation issue caused by simultaneously using anti-affinity groups and scheduling labels
- 【Documentation】Fixed some documentation links pointing to wrong addresses
- 【Host】Optimized zstack monitoring data pulling, corrected novnc access path (by zero)
- 【Host】Fixed PVE VM IP synchronization issue
- 【Host】Fixed AWS system reinstallation timeout failure issue
- 【Host】Fixed Google Cloud elastic public IP synchronization exception issue
- 【Host】Shutdown without charge operation only validates in shutdown state
- 【Host】Fixed VMware disk category recognition error after system reinstallation
- 【Host】Fixed AWS VM creation disk DeviceName parameter exception issue
- 【Host】Fixed issue where attached disks were not properly deleted when deleting public cloud machines
- 【Host】windows && GPU-VGA requires images to support UEFI
- 【Host】Fixed parameter error when forcing shutdown
- 【Host】Fixed deployment failure issue when creating ceph VMs
- 【Host】Fixed issue where setting qemu extra params failed
- 【Host】Fixed failure issue when converting esxi to kvm lvm disk VMs
- 【Host】Fixed issue where userdata was not decoded during deployment
- 【Host】Fixed inaccurate memory utilization calculation when hosts enable huge pages and simultaneously set mem_cmt_bound
- 【Deployment】After migrating deb sources to object storage, ubuntu update failures occurred
- 【Billing】Multi-dimensional analysis charts, when horizontal axis fields are too many, labels displayed at an angle may still overlap, adaptively hide some labels (Enterprise Edition)
- 【Billing】Fixed multi-dimensional analysis page, simultaneous requests for reports and default multi-dimensional analysis templates causing data confusion; fixed multi-dimensional analysis error reporting for empty data (Enterprise Edition)
- 【Work Order】Fixed host batch configuration adjustment, work order displaying incorrect disk information issue (Enterprise Edition)
- 【Billing】Bill resources add UI policy restrictions (Enterprise Edition)
- 【Billing】Fixed filtering parameters when exporting selected cloud accounts in cost-cloud account (Enterprise Edition)
- 【Authorization】Fixed issue where function selection page did not default to selected authorized functions after authorization completion (Enterprise Edition)
- 【Authorization】Fixed issue where server identification codes would change after public cloud configuration adjustment (Enterprise Edition)
- 【Authorization】Fixed issue where dictionary settings would be reset after changing function selection (Enterprise Edition)
New Features
- 【Host】Support public cloud resource conversion of billing modes, already supports Alibaba Cloud VMs and disks, Tencent Cloud VM prepaid and pay-as-you-go mutual conversion
- 【Host】Local virtualization disks support "Set Auto Reset on Shutdown" feature
- 【Host】Bare metal supports automatically joining VLAN subnets, achieving VLAN-based bare metal network isolation
- 【webconsole】Add page source check, prohibit direct opening via links, improving website security
- 【webconsole】Support setting referer whitelist, allowing pages in referer whitelist to directly open webconsole console
CHANGELOG
For specific changes, please refer to:
v3.11.10
date: 2025-04-04
The highlights of this v3.11.10 update are:
- Systematic optimization of service performance, including VM scheduling time when resources are abundant, monitoring service query time, and frontend page rendering time.
- Monitoring panel and monitoring query experience have significant improvements, supporting folding monitoring panels, exporting and sorting monitoring data, optimizing rendering speed when there are many monitoring metrics, etc.
- Comprehensive security upgrade, all services disable TLS v1.0 and 1.1 by default, disable flawed encryption algorithms such as 3DES, SHA1, close pprofiling interfaces, frontend removes webpack source map information, etc.
- Built-in virtualization continues to improve user experience, such as automatically caching disk images to ceph storage, improving ceph disk usage experience, local disk cloning speed increased by more than 10 times, etc.
Feature Improvements
- 【Frontend】Page title automatically displays current page information
- 【Frontend】List loading optimization, large page data supports batch rendering, switching data does not clear current data and switches directly
- 【Scheduling】Optimized VM scheduling time for large-scale host nodes
- 【Scheduling】Detect subnet IP shortage causing scheduling failures as early as possible
- 【VM】host-image service optimization, using nbd protocol to remotely mount local disks, local disk cloning speed increased by more than 10 times
- 【VM】kvm VMs support sata ahci disk type
- 【VM】When creating VMs, disks use xfs default setting ftype=1 to support docker's overlayfs (issue#22211)
- 【Host】Optimized time for hosts to report passthrough devices, avoiding host service startup timeout due to too many passthrough devices
- 【Host】Expanded host table sys_info field to use longtext, avoiding hostagent startup failures on some models
- 【Host】Host list displays CPU, memory, and disk allocation and utilization rates
- 【Host】Provide independent API for setting host and storage oversubscription ratios, convenient for permission control
- 【Bare Metal】Adjusted redfish client connection timeout to 60s, avoiding timeout issues when managing bare metal
- 【Network】Allow setting IPv6 network prefix when creating VPCs
- 【Network】Creating VMs supports manually setting IPv6 addresses, also allows manual specification when changing VM IPv6 addresses
- 【DNS】Allow different accounts' dns zones to have the same name
- 【Monitoring】Optimized time for querying monitoring metric items
- 【Monitoring】Support sorting metrics for result aggregation functions
- 【Monitoring】telegraf component adds more monitoring metrics
- 【Monitoring】No longer batch send when alert count is too high
- 【Monitoring】Optimized time for querying alert history
- 【Monitoring】Support cloning monitoring alert policies
- 【Monitoring】Added default alerts for host NIC bond devices and system.load1_pcore
- 【Monitoring】Added monitoring metric queries for NIC packet send/receive counts and send/receive bytes
- 【Monitoring】Display optimization, support modifying monitoring panel names, folding panels, cloning panels, adjusting panel order, displaying curve data, supporting data pagination, supporting IN queries, etc.
- 【Monitoring】Added zstack memory monitoring, corrected novnc access address (contributed by open source user zero)
- 【Authentication】Optimized time for password authentication
- 【Deployment】Platform deployment etcd enables TLS authentication by default
- 【Cloud Management】s3 supports specifying signature algorithm version as v2 or v4, backup storage also supports setting s3 signature algorithm version
- 【Cloud Management】Support Alibaba Cloud VM prepaid to pay-as-you-go conversion
- 【Cloud Management】Support VMware VMs automatically setting UEFI boot items based on templates
- 【Cloud Management】Support operating managed Cloudpods VM migration
- 【Cloud Management】Allow Alibaba Cloud to add disks when adjusting configuration while powered on
- 【Cloud Management】Support public cloud users enabling/disabling console login (Enterprise Edition)
- 【Work Order】Work orders support setting power on, power off, and delete related parameters (Enterprise Edition)
- 【Work Order】Host configuration adjustment work orders add more detailed configuration information (Enterprise Edition)
- 【Bastion Host】Support adding, updating, and deleting Jumpserver host information, maintaining full lifecycle synchronization with cloud management hosts (Enterprise Edition)
- 【Bastion Host】Support adding hosts to organizational structure (Enterprise Edition)
- 【Monitoring】Monitoring dashboard adds resource information display, CPU, memory, storage utilization and allocation rates (Enterprise Edition)
- 【Frontend】List export optimization, support custom column order (Enterprise Edition)
- 【Inspection】Automatic inspection supports chrony as time synchronization service (Enterprise Edition)
Bug Fixes
- 【VM】Fixed state inconsistency issue caused by synchronizing state when mounting and unmounting VM disks
- 【VM】Fixed issue where sriov NICs could not access virtio NICs on the same network segment as current host
- 【VM】Fixed CPU and memory hot-plug failure issue caused by old version VMs not setting default memory devices
- 【VM】Fixed issue where deployed VMs could not start on some CPUs
- 【VM】Fixed issue where lvm disks were occupied by other machines after host restart
- 【VM】Fixed deployment failure issue caused by incorrect path when deploying lvm disk VMs
- 【VM】Fixed conflict issue when multiple disks simultaneously execute automatic snapshot policies
- 【VM】Fixed incorrect progress reporting issue when deleting multiple VM snapshots
- 【VM】Unable to delete disk images after combining multiple disk images into host images
- 【Monitoring】Fixed issue where filtering and aggregation based on partial labels did not take effect
- 【Monitoring】Fixed issue where monitoring query statements were not completely copied after cloning monitoring panels
- 【Deployment】Avoid issue where router id might be set to 0 when deploying high availability keepalived
- 【Deployment】Fixed occasional issue where etcd generated large amounts of etcdctl zombie processes
- 【Container】Fixed possible authentication failure issue when synchronizing system K3s clusters
- 【webconsole】Fixed high CPU usage issue with webconsole service ssh client
- 【Backend】Fixed issue where cronjobs did not execute as planned due to incorrect timezone settings
New Features
- 【VM】VM disks support auto-reset on shutdown feature
- 【VM】Image service supports automatically caching system images to ceph storage
- 【Authentication】Support forcing global MFA switch
- 【Cloud Management】Support synchronizing Kingsoft Cloud RDS resources
- 【Cloud Management】Support batch creating public cloud users (Enterprise Edition)
- 【Billing】Support Azure RI resource pulling (Enterprise Edition)
- 【Billing】Support Volcano Cloud bills (Enterprise Edition)
- 【Frontend】Login page supports custom small screen/full screen style settings (Enterprise Edition)
CHANGELOG
For specific changes, please refer to:
v3.11.9
date: 2024-12-24
The highlights of this v3.11.9 update are:
- Functional optimization of monitoring services, including chart panel interactions, adding more host metrics, and sorting of result aggregation functions.
- Fixed issue where scheduler repeatedly calculated resources in a short time when concurrently creating VMs, improving scheduling accuracy and availability during concurrent creation.
- When creating VMs, support mounting recognition and initialization of cross-disk LVM logical volumes.
In addition, previous versions automatically set dns_domain to a domain name cloud.onecloud.io that we thought would not be used as an example when deploying clusters. However, this domain name began accepting resolution on October 31, 2024, causing some users using 8.8.8.8 DNS service who did not modify dns_domain configuration to encounter DNS resolution exceptions. Therefore, starting from 3.11.9, dns_domain is no longer set by default. Please pay attention to this configuration if you encounter similar issues.
Feature Optimizations
- 【VM】VMs no longer clean spool directory when saving images
- 【VM】VM creation deployment supports lvm spanning the first cloud disk partition
- 【Monitoring】Support outputting monitoring data to kafka
- 【Monitoring】Support more monitoring metrics, including system, processes, bond, etc.
- 【Monitoring】Add alert policies for bond NIC devices
- 【Monitoring】Monitoring panel function improvements, add min and max aggregation functions, support sorting by aggregation function results
- 【Monitoring】Optimized monitoring curve display style and chart interactions, support modifying dashboard names, allow adjusting panel order, more precise custom time
- 【Bare Metal】Support updating root-disk-matcher parameter
- 【Bare Metal】Increase redfish api client timeout to avoid management failure due to long API access time
- 【Disk】Image automatic conversion efficiency optimization, allowing concurrency
- 【Host】Optimized shared storage resource reporting, when mounting multiple hosts, select one master host for resource reporting
- 【Host】Resource statistics for CPU, memory, storage distinguish between virtual and physical dimensions
- 【Cloud Management】Synchronizing Tencent Cloud balance includes credit limit
- 【Cloud Management】Update azure cloud user creation api
- 【Cloud Management】When clearing recycle bin VMs, add secondary confirmation if running VMs exist
- 【Permission】Extended permissions, single function may involve multiple permissions, extended so each permission can affect function usage
- 【Operation Log】Support viewing operation logs for scheduling failures
- 【In-Site Message】List display optimization, excessive content displayed in popup boxes
- 【Billing】Optimized logic for pulling bills from GCP BIGQUERY, minimize full table scan frequency, reduce high query costs (Enterprise Edition)
- 【Billing】Exported billing tables support Euro currency (Enterprise Edition)
- 【Permission】Shopping cart function improvements, support permission control, work orders, resource names changed to more descriptive names (Enterprise Edition)
- 【Monitoring Dashboard】Resource overview display optimization, add more resource metrics (Enterprise Edition)
Bug Fixes
- 【VM】Cancel configuration of cloud.onecloud.io as default dns domain
- 【Scheduler】Avoid uneven VM allocation issue caused by scheduling strategy assigning to the same host
- 【VM】Issue where state was not synchronized after mounting and unmounting disks
- 【VM】Fixed China Telecom Cloud instance billing type synchronization error
- 【VM】VM instance azure SKU and image selection add adaptation rules
- 【VM】Fixed libvirt import parsing xml failure issue, will also parse live xml
- 【VM】Fixed Cloudpods VM migration issue via cloud management
- 【Storage】Fixed snapshot deletion failure issue caused by source disk already being deleted
- 【Storage】Fixed issue where block storage configuration updates did not take effect
- 【Scheduling】Fixed issue where resources in pending usage were repeatedly calculated during resource statistics
- 【Monitoring】Fixed issue where alert recovery included other resources
- 【Monitoring】Fixed unclear display issue for multi-metric alert email content
- 【Monitoring】Issue where charts were not copied after cloning monitoring panels
- 【Monitoring】Fixed issue where new cloud accounts did not pull monitoring
- 【Bare Metal】Fixed issue where default gateway was not set during deployment
- 【Cloud Management】Extended RDS maintain_time field length
- 【Billing】Allocation project selection supports remote filtering (Enterprise Edition)
- 【Billing】Multi-dimensional analysis second dimension aggregated by labels, table header should display label key (Enterprise Edition)
CHANGELOG
For specific changes, please refer to:
v3.11.8 (including v3.11.7)
date: 2024-10-14
The highlights of this 3.11.8 update are:
- ocboot defaults to using k3s as the base for running platform services, replacing the old v1.15.12 k8s version. The new deployment method will use buildah and container images for deployment, cleaner and lighter, without needing to install ansible and other python environments on hosts. At the same time, ocboot will also be compatible with old k8s deployed clusters, adding nodes and upgrading services are not affected.
- Automatically deploy QGA (qemu-guest-agent) when deploying VMs. Functions dependent on QGA (such as modifying host passwords when powering on, changing host IPs, etc.) will be more reliable. More functions will be improved and extended based on QGA in the future.
- Frontend has overall performance optimization for resource lists, faster response speed, smoother experience.
Feature Optimizations
- 【Frontend】Resource list loading speed optimization
- 【Frontend】Alert resources add IP search
- 【Multi-Cloud Management】When managing Cloudpods cloud accounts, automatically set endpoint_type type
- 【k8s Management】Cluster and namespace dropdown lists support search
- 【Monitoring】Add inodes utilization monitoring
- 【Deployment】ocboot deployment tool supports running with sudo permissions
- 【Deployment】ocboot deployment control nodes will not restart servers
- 【Work Order】Work order approval process speed improvement (Enterprise Edition)
- 【Frontend】External data as separate first-level menu (Enterprise Edition)
- 【Frontend】mfa authentication tool recommends using NingDun token
- 【Storage】Optimized lvm slvm snapshot, image, and backup related functions
- 【Storage】lvm slvm storage supports encrypted disks (Enterprise Edition)
Bug Fixes
- 【Frontend】When adding VMware cloud accounts, selecting availability zones used usable=true filter, causing availability zone list to be empty
- 【host-deployer】When manually deploying network configuration, did not add 169.254.169.254 route, causing inability to web ssh login to cloud hosts
- 【host-deployer】Fixed host-deployer service authentication failure issue in cmp deployment mode
- 【host-deployer】Fixed issue where uefi image boot method detection was incorrect
- 【VM】When deploying VMs, add udev rules for automatic online of memory and CPU hot expansion, avoiding hot-expanded CPU and memory not being automatically enabled
- 【VM】Fixed issue where image cache was deleted during live migration causing inability to live migrate
- 【VM】Adjusted concurrent VM deployment count, add configuration for configurable concurrency
- 【VM】Fixed occasional port conflict deployment failure issue when concurrently creating VMs
- 【VM】Fixed rbd VM image save failure issue
- 【VM】Allow creating uefi boot VMs when image boot method cannot be detected
- 【Storage】Fixed issue where image cache in use was mistakenly deleted during automatic image cache cleanup
- 【Bare Metal】Occasionally deployment failures occurred
- 【Bare Metal】Fixed failure issue when bmc NIC appeared with 0.0.0.0 ip when detecting ipmi
- 【Bare Metal】Fixed validate-ipmi error when frontend batch managing bare metal
- 【Frontend】Fixed menu permission unchecked issue
- 【cloudshell】Fixed issue where processes executing interactive mode would not exit causing process leaks
- 【lbagent】lbagent startup should enable path mtu probe
- 【Monitoring】Fixed OR query condition error issue
- 【Monitoring】Fixed AWS, GCP CPU monitoring data pulling issue
- 【Command Line Tool】Fixed climc host-ssh login host error issue
- 【Multi-Cloud Authentication】Fixed multi-cloud permission synchronization exception issue
- 【Deployment】Fixed issue where ingress could not be accessed when deploying via helm chart
- 【Deployment】Fixed issue where VictoriaMetrics CSI definition did not take effect when deploying via helm chart
- 【Deployment】Fixed issue where imageRepository would refresh during ocboot upgrade
- 【Billing】Avoid negative growth rate caused by negative last month cost (Enterprise Edition)
- 【Dictionary】Dictionary default values use original values (Enterprise Edition)
New Features
- 【Monitoring】Monitoring queries and monitoring panels support result aggregation functions
- 【Monitoring】Container hosts add monitoring metrics
- 【Multi-Cloud Management】Cloud accounts support cloning
- 【Monitoring Panel】Utilization tiles support jumping to current resources
- 【VM】Support automatically deploying QGA (qemu-guest-agent) when creating VMs
- 【VM】Scheduling supports allocating CPU pinning and NUMA memory, supports reallocation when VMs restart
- 【Frontend】VM NICs support setting auxiliary IPs
- 【Billing】Bill reports hide bill entries with payable amount of 0 (Enterprise Edition)
- 【Frontend】Monitoring dashboard supports setting custom addresses (Enterprise Edition)
- 【Billing】Support batch rerun bills (Enterprise Edition)
- 【Frontend】Login page supports style settings (Enterprise Edition)
CHANGELOG
For specific changes, please refer to:
v3.11.6
date: 2024-08-15
Feature Optimizations
- 【Platform】Frontend and backend optimized VM and host list query speed
- 【Multi-Cloud Monitoring】Avoid generating large amounts of operation logs
- 【Multi-Cloud】Optimized Tencent Cloud offline availability zone resource synchronization
- 【Network】Enable host NIC gso feature by default using ethtool
- 【Bare Metal】Add default redfish api compatibility check for bare metal
- 【Multi-Cloud Management】Cloud users support displaying cloud subscription information
- 【Platform】Add uos system icon
- 【VM】Cloudpods platform supports selecting iso images to create VMs
- 【Bare Metal】New bare metal adds management port MAC address validation
Bug Fixes
- 【Frontend】When adding VMware cloud accounts, availability zone list was empty
- 【VM】When manually deploying network configuration, did not add 169.254.169.254 route, causing inability to web ssh login to cloud hosts
- 【VM】Fixed issue where setting VM disk cache mode caused VM startup failure
- 【Bare Metal】Fixed deployment failure issue when installing bare metal
- 【Bare Metal】Fixed management failure issue when detecting ipmi NIC appeared with 0.0.0.0
- 【Bare Metal】Fixed error when deleting bare metal NICs
- 【Bare Metal】Installation failure issue when mixing megaraid and pcie disks
- 【Network】Simplified cleanup of unused security groups
- 【Storage】Fixed lvm storage encrypted disk and snapshot creation failure issue
- 【Network】Unknown status issue when synchronizing private cloud kvm VM security groups
- 【Notification】Avoid panic error caused by not setting phone number
- 【Monitoring】Fixed duplicate message records caused by multiple alert metrics
- 【Host】Avoid generating too many qga ping logs
- 【Multi-Cloud】Fixed VMware iso VM creation failure issue
- 【Billing】Avoid negative growth rate caused by negative last month cost (Enterprise Edition)
- 【Security Group】Public cloud security group rules inbound and outbound source restrictions cannot be empty
- 【Multi-Cloud Management】Fixed cloud user operation unclickable issue, removed health status check
New Features
- 【Monitoring】Add network pps, disk inodes_usage related monitoring metrics
- 【Network】Support setting dhcp_relay attribute when creating subnets, convenient for bare metal network settings
CHANGELOG
For specific changes, please refer to:
v3.11.5
date: 2024-07-04
Feature Optimizations
- 【VM】Clean image cache by default
- 【VM】Adjust configuration to adapt disk_type=swap data disk display
- 【Network】Hide ipv6 settings by default when creating subnets
- 【Network】VM NIC bandwidth adjustment optimization, upper limit to 100Gbps
- 【Host】When converting hosts, added disk mount paths can only be subdirectories of /opt/cloud/workspace
- 【Multi-Cloud】Support Tencent Cloud new version natgateway API
- 【Remote Control】VM vnc and ssh login and other remote operations, disable watermark by default
- 【Multi-Cloud】New scheduled tasks, add default time
- 【Work Order】Add operation log records (Enterprise Edition)
Bug Fixes
- 【VM】Fixed issue where deleting VM snapshots did not delete
- 【VM】Fixed issue where resource label keys were duplicated when setting VM list items
- 【Monitoring】Fixed network monitoring data unit display issue
- 【Installation Deployment】Fixed issue where CloudShell might not open on first installation
- 【Installation Deployment】Fixed issue where CloudShell opened via docker compose deployment did not use current login permissions
- 【Installation Deployment】operator compatible with mysql 8.0
- 【Installation Deployment】Fixed issue where kube-apiserver occasional restart caused subsequent installation steps to fail
- 【Monitoring】Fixed issue where role or alert robot selection box was not set as required when creating alert policies with notification type as role or alert robot
- 【Authentication & Security】Fixed issue where connection test did not validate form content validity when creating SMS notification channels, added phone number validation
New Features
- 【VM】Support VMware migration (v2v migration) to Cloudpods private cloud
- 【Network】eip supports querying associated resources
- 【Host】Automatically clean image cache, by default when image cache exceeds 12% storage space, automatically clean unused images cached for more than 30 days
- 【Bare Metal】When creating bare metal, add system disk matcher (command line support)
- 【Bare Metal】Support setting lan channels, root id and other configuration information for various brand models of bare metal (command line support)
- 【Authentication & Security】User list export supports simultaneously exporting projects and groups users joined
- 【Network】waf details support displaying domain information, origin address, origin method and other information
CHANGELOG
For specific changes, please refer to:
v3.11.4
date: 2024-06-06
Feature Optimizations
- 【Billing】Cost prediction function optimization (Enterprise Edition)
- 【Multi-Cloud】Enable Volcano Engine, Kingsoft Cloud and Oracle Cloud platform switches by default
- 【Network】Tencent Cloud supports WAF policies
- 【Network】Azure application services display more attribute information
- 【Host】Support automatic detection of bare metal passthrough devices
- 【Host】VM export disk and memory units unified to GB
- 【Monitoring】Allow setting alert policies containing multiple trigger conditions
- 【Multi-Cloud】Support ZStack shared storage
Bug Fixes
- 【System】Fixed one-click inspection failure issue in multi-IP environment (Enterprise Edition)
- 【Monitoring】Fixed issue where multiple alert metric relationship errors caused inability to alert
- 【Host】Fixed SLVM storage related issues
- 【Host】Fixed occasional hostname update failure issue
- 【Host】Fixed storage registration failure issue when hosts register
- 【Host】Fixed issue where active access to physical gateway was blocked after enabling IPv6 security groups
New Features
- 【Host】New community images, support one-click import
CHANGELOG
For specific changes, please refer to:
v3.11.3
date: 2024-4-28
Feature Optimizations
- 【Host】Support deleting prepaid hosts from Alibaba Cloud, Huawei Cloud and other public clouds
- 【Host】Support filtering VMware platform images by availability zone
- 【Host】Optimized local disk snapshot deletion logic, synchronously delete snapshots on hosts when API deletes snapshots
- 【Host】Custom device types remove automatic detection logic, need to specify detection host
- 【Host】GPU device detection only filters manufacturers as NVIDIA/AMD for DISPLAY class code
- 【Host】Support host reset data disks, support local disk cold migration
- 【Multi-Cloud】Optimized cloud account synchronization speed, support concurrent synchronization
- 【Multi-Cloud】Kingsoft Cloud hosts, networks, labels and other resource synchronization support
- 【Multi-Cloud】cloudid refactoring, cloud and local user information changed from many-to-one to one-to-one model
- 【Multi-Cloud】Automatically clean invalid image cache information, optimize managed cloud image loading speed
- 【Authentication】Update each service's project information synchronization in real time as needed
- 【Console】Allow hiding dashboard, cloudshell and other function entries via UI policies
Bug Fixes
- 【System】Fixed bastion host not displaying belonging project issue (Enterprise Edition)
- 【Host】Fixed resource display information error issue when creating resources based on shopping cart (Enterprise Edition)
- 【Host】Fixed ARM VM deployment failure issue using lightweight VMs
- 【Host】Fixed monitoring agent online installation failure issue for Debian/Ubuntu and other operating systems
- 【Host】Fixed Ubuntu/Debian netplan bond NIC network configuration invalidation issue
- 【Host】Fixed userdata script not executing automatically due to default umask setting issue
- 【Load Balancer】Fixed error when adding load balancer listener http rules reporting domain not found
- 【Network】Fixed IPv6 subnet search failure issue
- 【Network】Fixed vpcagent ovn synchronization command timeout signal killed issue
- 【Multi-Cloud】Fixed issue where public cloud sub-account notification could not select email
- 【Authentication】Fixed incomplete permission information display issue
- 【Multi-Cloud】Fixed Cloudpods region synchronization loss issue
- 【Multi-Cloud】Fixed Tencent Cloud API interface changes causing synchronization exception issue
New Features
- 【Billing】Billing module supports Kingsoft Cloud bills (Enterprise Edition)
- 【Billing】Support bill "closing" operation, bills for closed months cannot be updated (Enterprise Edition)
- 【Billing】Support integrating Azure (international and domestic) object storage bucket bills (Enterprise Edition)
- 【Billing】Support custom external data billing (Enterprise Edition)
- 【Host】Support disk reset interface
- 【Host】LVM cloud disks support snapshots, LVM cloud disks support expansion
- 【Multi-Cloud】Support integrating Oracle Cloud
- 【Multi-Cloud】Support Volcano Engine Cloud SSO passwordless login
CHANGELOG
For specific changes, please refer to:
v3.11.2 (including v3.11.1)
date: 2024-3-20
Feature Optimizations
- 【Report】Support ignoring project lists with cost of 0 (Enterprise Edition)
- 【Network】Support synchronizing load balancer security group information
- 【Network】Support synchronizing Huawei Cloud public NAT gateway
- 【Multi-Cloud】Support integrating Azure international storage bucket bills
- 【Host】Support file upload and download via webconsole
- 【Host】Support bidirectional synchronization of VMware host remarks
Bug Fixes
- 【System】Fixed host mount leak issue caused by incorrect host directory mount method for telegraf containers
- 【Host】Fixed error when adjusting configuration for hosts created from host images
- 【Host】Fixed issue where VMs could not find bridge when starting
- 【System】Fixed occasional unresponsive remark modification issue
- 【Host】Fixed GPU display duplication issue
New Features
- 【Billing】Support cloud account closing function (Enterprise Edition)
CHANGELOG
For specific changes, please refer to:
v3.11.0
Version 3.11 started development in April 2023, taking about 10 months, with more than 4000 PRs.
- Open Source Summer 2023
Thanks to four students from Open Source Summer 2023 for contributing the following features to Cloudpods v3.11:
- VM Emergency Mode: Allows booting VMs into a temporary in-memory Linux system. When VM kernel files are corrupted, rootfs file system failures occur, etc., and users cannot normally start VMs, they can restart VMs to enter emergency mode, view VM file systems, locate problems, backup data, and repair problems when possible. @IRONICBo
- Support Volcano Engine: Support Volcano Engine resource management, and Enterprise Edition also supports managing and analyzing Volcano Engine bills. @LyndonKong
- Support deploying Ubuntu 22.03 (already backported to 3.10 and released) @niconical
- Support changing KVM VM NIC configuration via QGA (already backported to 3.10 and released) @chentianxiang
- Host Improvements
Host functionality in v3.11 has many improvements, focusing on two directions: First, private cloud hosts more fully utilize hardware performance; Second, improve cloud host usability.
To fully utilize hardware performance, Cloudpods already supports many features, including PCI device passthrough (including NICs, GPUs, NVME devices, etc.), NIC SR-IOV and ovs offload, enabling host huge pages, binding virtual CPU cores to physical cores, etc. In v3.11, Cloudpods implemented the long-awaited VM NUMA-aware memory allocation strategy. After enabling, it tries to allocate a VM's memory within the same NUMA node's memory space, and the VM's vCPUs are also allocated to the same NUMA node as much as possible, avoiding performance degradation caused by cross-NUMA allocation of VM CPUs and memory.
Another important feature is support for using LVM to manage shared block storage (SAN storage). Shared block storage uses LVM management, VMs directly mount logical volumes (LV) allocated from shared block storage, eliminating file system overhead introduced by using files as virtual disks. Cloudpods supports clvm and slvm two shared LVM management methods. Simply put, clvm mode uses cluster-mode clvm service to manage LVM volumes on shared network storage, clvm automatically provides distributed high availability for lvm management service. But the disadvantage is clvm configuration is complex. slvm does not require configuring complex cluster clvm service, each host node only needs to install single-machine lvm service, platform selects one node as shared storage's lvm configuration node, platform automatically refreshes lvm metadata on nodes that need to mount after allocating LV on that node, obtaining allocated LV. This method avoids the complexity of deploying distributed clvm. But after the configuration node becomes unavailable, manual switching of configuration node is needed. But this issue can be gradually solved in later iterations.
In addition, Cloudpods also has preliminary support for vGPU, including Nvidia and AMD's virtualized vGPU.
In improving host usability, the following work was done: 1) Supported emergency mode for built-in private cloud KVM VMs, allowing users to diagnose and repair VMs that cannot start through a temporary operating system. 2) Support logging into Windows VMs via WebRDP protocol. The advantage of using RDP protocol is that it can automatically passwordlessly log into Windows VMs and supports clipboard, better usability. 3) Full support for userdata, allowing users to set scripts to execute after host initial startup when applying for hosts, making host application process more convenient. 4) Support migrating managed Cloudpods cluster KVM VMs to built-in private cloud, users can use this function to consolidate multiple Cloudpods private clouds.
Finally, in host networking, IPv6 support is an important function introduced in v3.11, supporting IPv6 address management and assigning IPv6 addresses to hosts.
Another important network improvement is VMware adopting a new Layer 2 network model, each VMware network port group corresponds to a Cloudpods Layer 2 network. This way VMware users can intuitively see VMware vCenter port groups on the platform, easier to understand and configure corresponding Layer 2 network IP subnets. But users who used VMware before v3.11 will be affected by this change after upgrading to 3.11. After upgrade, need to reconfigure VMware IP subnets, change IP subnet's Layer 2 network to the Layer 2 network corresponding to its VMware port group.
- Multi-Cloud Iteration
Multi-cloud functionality continues to iterate in v3.11. Besides adding Volcano Engine and Kingsoft Cloud public cloud platforms, the most important update is refactoring the configuration information management model.
We once invented a multi-cloud scenario configuration information management model, using one-to-many mapping relationships, i.e., platform maintains one configuration data, this configuration corresponds to multiple configurations on various cloud platforms. The most common is security group configuration. Platform maintains one security group configuration, this security group corresponds to security groups with the same configuration on various cloud platforms, various accounts, various regions. For example, if a user defines a security group allowing access to TCP port 22, then security groups allowing access to TCP 22 under various VPCs in various regions under various cloud accounts on various cloud platforms for this user are all associated with this security group, platform maintains mapping relationships between cloud security groups and local security groups through security group cache data structures. At design time, we expected this configuration management method to improve usability of using multi-cloud configurations, i.e., users don't need to care whether this configuration is maintained on cloud, just select local security group, platform will automatically synchronize local security group to cloud, or find a corresponding security group on cloud. But in actual use, users were confused by this configuration maintenance method, local and cloud configurations easily became inconsistent, and security group changes caused user host network failures multiple times. Although previous versions made many improvements and restrictions, avoiding some problems, overall effect was not good. Therefore, starting from v3.11, we abandoned this one-to-many configuration maintenance model, changed all configuration information maintenance to one-to-one mapping model. Affected configuration information includes: security groups, DNS records, automatic snapshot policies, load balancer SSL certificates, load balancer access control.
- FinOps
2023 was a year of rapid FinOps development. Cloudpods Enterprise Edition's FinOps functionality has significant improvements, mainly in the following aspects:
Cost Allocation: v3.11 Enterprise Edition did a lot of work in cost allocation, supporting complex cost allocation strategies. Including allocating bill costs to local projects based on cloud projects or labels, allocating prepaid (including CPP) resource costs to months within the period, and supporting allocating shared costs to specified projects according to equal allocation or proportional allocation strategies.
Cost Analysis: v3.11 Enterprise Edition cost analysis capabilities are stronger, on one hand functionality is more powerful, can filter, aggregate and generate reports on costs based on bill labels, project labels, resource labels and other multi-dimensional labels. On the other hand, query performance is faster, through joint queries of configuration data stored in MySQL and transaction bill data stored in ClickHouse, making queries both fast and accurate.
Cost Prediction: Based on multiple linear regression model to achieve prediction of fine-grained dimension cost data such as project+account+resource type+billing method, support users inputting regression model parameters to adjust prediction data, support multi-dimensional analysis queries on prediction data.
Billing Strategy: Support rate sets, can set different rates for different projects, domains or organizational structures, support setting billing time granularity of year, month, day, hour, second, support fixed price, pay-as-you-go and tiered pricing methods.
Also worth mentioning is support for container billing. Platform automatically collects resource usage of pods in configured container clusters, including CPU, memory and PVC space quotas allocated to pods, for billing.
- Installation and Deployment
Cloudpods v3.11 adapted to more operating systems. Besides CentOS 7, Debian 10, 11, Ubuntu 22.04 already supported in previous versions, also adapted to a series of operating systems corresponding to CentOS 8, including: AnolisOS 8.8, OpenCloudOS 8.8, RockyLinux 8.9, AlmaLinux 8.9, etc. Also adapted openEuler 22.03 LTS SP3 to address domestic innovation scenarios. All above systems can use ocboot for online deployment.
At the same time, for commercial version offline deployment scenarios, with CentOS 7 EOL in June 2024, v3.11 no longer uses CentOS 7.9 as the main offline deployment operating system, but adopts openEuler 22.03 LTS SP3.
- Others
Starting from v3.11, default monitoring storage changed to VictoriaMetrics, further improving monitoring data query speed.
New Features (New Features)
- 【Host】KVM VM NUMA memory allocation
- 【Host】Support network configuration management IPv6 addresses, assign IPv6 addresses to hosts
- 【Host】KVM supports vGPU (NVIDIA/AMD vGPU)
- 【Host】KVM supports VM emergency mode
- 【Host】Support centralized block storage (SAN) allocating disks via CLVM/SLVM
- 【Host】Support migrating managed Cloudpods cluster KVM VMs to local private cloud
- 【Host】Support NIC traffic limits
- 【Host】Support logging into Windows VMs via WebRDP protocol
- 【Multi-Cloud】Support Volcano Engine integration
- 【Multi-Cloud】Integrate AWS container EKS
- 【Multi-Cloud】Support Alibaba Cloud resource directory, manage multiple accounts under Alibaba Cloud resource groups
- 【Authentication】Organizational structure support (Enterprise Edition)
- 【Billing】Support container billing (Enterprise Edition)
- 【Billing】Rates support applying to different projects/tenants, also support different granularity settings (Enterprise Edition)
- 【Billing】Support cost allocation, allocate/equalize specified project costs to other projects (Enterprise Edition)
- 【Billing】Support resource label lists, can uniformly maintain cloud labels (or projects) not supported or missing label information through this list (Enterprise Edition)
- 【Cost Optimization】Support abnormal consumption function, generate exception lists based on project daily consumption or monthly cumulative consumption thresholds (Enterprise Edition)
- 【System】Web console supports Japanese
- 【System】Support integrating bastion host, automatically add JumpServer bastion host when creating hosts (Enterprise Edition)
- 【System】Support shopping cart function, can simultaneously batch create resources of different specifications (Enterprise Edition)
- 【Deployment】Support Euler/SUSE system installation and deployment
Feature Enhancements (Enhancements)
- 【Host】Support userdata, new hosts also support specifying initialization execution scripts via userdata
- 【Host】KVM host backups support storage via object storage
- 【Host】VMware adopts new Layer 2 network model, each Layer 2 network corresponds to one network port group
- 【Host】Disk speed limit optimization, support speed limiting for single disk images
- 【Host】Keys support sharing
- 【Multi-Cloud】Refactored cloud and local resource mapping relationship model, changed security groups, DNS records, snapshot policy configuration relationships from many-to-one to one-to-one
- 【Multi-Cloud】Alibaba Cloud/Huawei/AWS support object storage setting cross-domain rules and policies
- 【Authentication】Optimized user/project deletion process, no need to check relationships
- 【Authentication】Optimized message subscription sending, support silence period settings
- 【Billing】Support setting bill label mapping to specified projects (Enterprise Edition)
- 【Billing】Support cost analysis based on project labels, build organizational structure based on projects (Enterprise Edition)
- 【Billing】Add prediction function, support entering related data to calibrate prediction data (Enterprise Edition)
- 【Billing】Multi-dimensional analysis supports saving as reports, can save frequently used query conditions as reports for one-click viewing (Enterprise Edition)
- 【Billing】Cost query performance optimization, combining MySQL and ClickHouse, achieving query speed acceleration (Enterprise Edition)
- 【System】Monitoring replaced with VictoriaMetrics
- 【System】Dashboard supports cost KPI tiles, such as committed cost ratio, label coverage, etc. (Enterprise Edition)
- 【System】Support using ID/IP search to simultaneously query multiple data
CHANGELOG
For specific changes, please refer to:
warning
- Starting from 3.11, Huawei Cloud will support organizational account functionality. After upgrade completion, resources under Huawei Cloud accounts will be completely cleaned once, then account resources will be re-synchronized. Therefore, VM password information under Huawei Cloud accounts will be lost, and ID information of all Huawei Cloud resources will change.