CubeCOS image service
Image import overviewโ
The CubeCOS Import feature provides a simple, seamless virtual machine (VM) migration solution. It ensures that both official cloud images and legacy workloads can be integrated into the Bigstack CubeCOS platform.
To improve user adoption and lower the barrier to entry from other virtualization platforms, CubeCOS offers two optimized migration paths based on your deployment scenario:
- Cloud Image: Use this path to quickly deploy standardized cluster services, such as Load Balancers or pre-configured front-end images. This method utilizes QCOW2 images, the standard format used in OpenStack environments, to ensure high performance and compatibility during the import process.
- Virtual Disk Image: Designed for migrations, this path allows you to move existing workloads from external platformsโincluding VMware (VMDK) and Microsoft Hyper-V (VHDX) into the CubeCOS ecosystem.
As the CubeCOS platform is integrated with the Skyline (OpenStack) architecture, it is natively compatible with Cloud Images (such as QCOW2). For virtual disk images (such as VMDK or VHDX), CubeCOS has introduced a new conversion toolkit. During the import process, the toolkit automatically converts the format and makes any necessary compatibility adjustments, transforming the virtual disks into a format recognisable by the CubeCOS platform.
Compare different image typesโ
The following section provides a technical deep dive into the two core image types supported by CubeCOS. Understanding these differences is essential for administrators to determine the optimal migration strategy, ensuring that guest operating systems are provisioned with the correct drivers, storage backends, and performance optimizations required for a stable KVM environment.
Cloud imageโ
The QCOW2 (QEMU Copy-On-Write version 2) format is the cornerstone of the Skyline (OpenStack) workflow.
- Design Goal: Serves as the standard virtual disk format for QEMU and KVM (Linux Kernel-based Virtual Machine) environments and is the preferred native format for OpenStack (Nova/Glance).
- Scalability: Optimized for spawning multiple instances quickly from a single source.
- Copy-On-Write (COW): This core feature ensures that underlying data blocks are only copied when data is written, effectively saving storage space.
- Thin Provisioning: Focuses on COW mechanisms to provide high-efficiency thin provisioning, keeping file sizes small while supporting encryption.
- Snapshot Mechanism: Snapshots are built directly into the QCOW2 file format itself, allowing multiple snapshot history records to be stored within a single file.
- Openness: It is a fully open-source and open-standard format maintained by the QEMU community.
Virtual disk imageโ
This category represents images imported via the From another hypervisor workflows.
VMDKโ
VMDK is the proprietary standard for the VMware ecosystem.
- Legacy Ecosystems: Specifically designed for VMware (vSphere/ESXi) environments
- Snapshot Mechanism: Typically managed using a base VMDK file paired with multiple Delta files (difference files).
- Copy-on-Write (COW): Supported, though the implementation and nomenclature differ from KVM standards.
- Openness: While much of the format specification is public, it remains primarily driven and maintained by VMware.
VHDXโ
VHDX is the successor to the VHD format, designed by Microsoft for modern virtualization needs.
- Microsoft Ecosystems: Designed for Microsoftโs Hyper-V virtualization platform, including Windows Server and Windows 10/11.
- Features:
- Large Capacity: Supports virtual disks up to 64 TB, significantly exceeding the 2 TB limit of the original VHD format.
- Data Protection: Features a built-in logging mechanism to prevent disk corruption during sudden power failures.
- Performance Optimization: Improves alignment and block sizes to enhance I/O performance.
- Snapshot Mechanism: Managed through a base VHDX file and corresponding differential VHDX files.
- Openness: The format specification is published, but the core ecosystem and tools remain centered around Microsoft products.
Import imagesโ
Visibility typesโ
When you import an image, Visibility determines how the resource is shared across the Bigstack CubeCOS multi-tenant architecture.
- Public: Allow users in all projects to view the image and use it to create virtual machines.
- Private: Limit image visibility and usage to members of the owning project to ensure resource isolation between tenants.
Visibility logic matrixโ
Depending on the user role and import path, resource ownership is assigned as follows:
| Import Path | Role | Project | Visibility | Result |
|---|---|---|---|---|
| From Bigstack | Administrator | Admin | public | The image will be put in Admin project. All users in any project can use the image. |
| Administrator | Any project that selected | private | The image will be put in specific projects. Only the users belongs to the projects can use the image. | |
| User | Project that the user belongs to | public | The image will be put in the project that the user belongs to. All users in any project can use the image. | |
| User | Project that the user belongs to | private | The image will be put in the project that the user belongs to. Only the users belongs to the projects can use the image. | |
| From another hypervisor (only private) | Administrator | Any project that selected | private | The image will be put in specific projects. Only the users belongs to the projects can use the image. |
| User | Project that the user belongs to | private | The image will be put in the project that the user belongs to. Only the users belongs to the projects can use the image. |
Overview
Upload, manage, and control access to images using centralized storage, versioning, and visibility settings.
Import OS image or virtual machines
Migrate virtual machines from various platforms to Bigstack CubeCOS using optimized UI workflows.
Import via CLI
Upload and import OS images, cloud templates, and virtual machines into CubeCOS via command-line interface. Comprehensive guide for v2v migrations from VMware, p2v migrations, Hyper-V, and other hypervisors to prepare VMs for deployment.
Create instance flavors
Configure a flavor, the virtual hardware profile that defines vCPU count, memory, and root disk size for an instance.
Create SSH keypairs
Create an SSH keypair to access instances securely without a password.
Create an instance
Provision a virtual machine from a predefined resource template, OS image, and configuration parameters.
Create an instance from ISO
Boot a virtual machine from a guest ISO image to manually install an operating system that doesn't have a ready-made cloud image available.
Connect to Instance by noVNC
Connect to a launched Linux or Windows instance using the browser-based noVNC console in the web UI.
Attach Security Groups to Instance
Attach security groups to an instance to control inbound and outbound network access.
Associate Floating IP to Instance
Associate a floating IP address with an instance to make it reachable from external networks.
Resize an instance
Resize an instance by switching it to a different flavor without recreating the virtual machine.
Create instance snapshot
Capture the current disk and configuration state of an instance as a reusable snapshot.
Restore an instance snapshot
Restore an OpenStack volume from a Ceph RBD snapshot, including prerequisites, restore steps, verification, and troubleshooting.
Live migrate instances
Move a running instance from one physical host to another without shutting it down.
Configure AI accelerator (GPU)
Enable direct GPU passthrough so virtual machines can access dedicated GPU hardware for AI and machine learning workloads.
Export an instance
Export an instance to a downloadable image file so it can be archived, transferred, or imported elsewhere.