Shared Storage Container is a storage-focused node type designed for centralized data hosting and sharing. Compared with general-purpose managed containers, it adds storage-oriented features such as NFSv4 export, Gluster Native (FUSE) support, larger disk capacity, optional auto-clustering, and built-in tools for exports and mount points.
Shared Storage Container is designed to serve data to multiple application nodes.
NFS and Gluster Native (FUSE)
NFS is optimized for straightforward, high-performance file access, while Gluster Native adds replicated-storage capabilities for backup and failover scenarios.
NFSv4 Export
Shared Storage Container can export data over NFSv4, offering better performance for large file counts, stronger security, and additional export capabilities.
Enlarged Disk Space
Storage nodes are provisioned with larger disk allowances than ordinary general-purpose containers.
Preinstalled Storage Software
Required NFS, RPC, GlusterFS, and related components are included for storage-oriented use.
Platform Elasticity
Vertical/horizontal scaling, pay-as-you-use pricing, file exports, mount points, and standard platform management tools remain available.
Any platform-managed container can receive NFSv4 mounts, but only the dedicated Shared Storage Container can export data over NFSv4.
Storage Container Creation
Create a Shared Storage Container from the environment topology wizard. The Storage section is available for all engine types, including custom Docker-based environments.
Enable the Storage layer and configure disk size, clustering, public IPs, and other settings.
1Enable the Storage layer
Open the topology wizard and activate the Shared Storage Container option.
2Set the Disk Limit
Choose the amount of disk capacity required for the storage workload.
3Optionally enable Auto-Clustering
Use the Auto-Clustering switch when you need a replicated storage cluster rather than independent storage nodes.
4Add public IPs if required
IPv4 and IPv6 addresses can be attached when storage must be reachable from outside the platform.
RAM and CPU are not the primary resources for a storage server, but they are still consumed during network operations. Pricing mainly depends on actual disk usage and external network traffic rather than only the configured disk limit.
Disk limit: If the maximum disk capacity available to a storage node is not sufficient, request a quota increase from the hosting provider.
Shared Storage Auto-Cluster
When Auto-Clustering is enabled, the platform automatically creates a replicated GlusterFS volume. Files are copied across storage bricks so the cluster can continue serving data even when one or more nodes become unavailable.
Auto-Clustering creates a replicated GlusterFS storage topology.
An existing standalone Shared Storage node cannot be automatically converted into a GlusterFS cluster; migration must be handled separately.
The source states that auto-clustering depends on a compatible Virtuozzo 7 virtualization region.
The storage auto-cluster requires at least 3 nodes.
Auto-Clustering cannot be disabled after cluster creation.
Horizontal scaling is performed in increments of 2 nodes to maintain quorum.
During cluster creation, the GlusterFS volume is mounted under /data and exposed through NFSv4. The cluster is then treated as one storage component rather than a collection of independent nodes.
Client containers mount the Shared Storage cluster as a single NFS server.
If one or more storage nodes fail, the default AutoFS client can switch to a working instance on the next read/write attempt.
Use Case Considerations
GlusterFS-backed Shared Storage is appropriate for many workloads, but the source highlights some cases where it is better or worse suited.
Recommended
Read-heavy shared content such as WordPress or Magento media files and applications that can retry failed storage writes.
Not Recommended
SQL database primary storage and applications with very heavy I/O requirements.
Database warning: Do not use Shared Storage as the main data store for SQL databases. It can still be appropriate for database backup and restore files.
Heavy I/O warning: NFS is not ideal for high-I/O applications, and a node failure during a write operation can create a risk of data corruption.
Storage Container Management
After the storage environment is created, common management operations can be handled directly from the platform dashboard.
Configuration File Manager
Manage files directly through the built-in file manager.
SFTP / SSH Gate
Use third-party tools or SSH when more advanced administration is required.
Mount Points
Attach exported storage to application containers.
Exports
Share selected directories with other platform instances or external servers.
Use the built-in file manager or the SFTP / SSH Gate details for storage administration.
SSH access is available either through the browser or a local SSH client for full server control.
Open a Web SSH session for advanced Shared Storage administration.
External Access and Dependencies
If Shared Storage must serve systems outside the current platform installation, attach a public IP and configure the container as an external NFS server.
Expand the storage node in the dashboard to see its attached public IP addresses.
When an environment is stopped or deleted, the platform checks whether other nodes depend on its mounted storage and displays the affected components before the action is confirmed.
The confirmation window identifies components that depend on the Shared Storage node.
Migration limitation: Live migration is not available for environments that contain Shared Storage Containers. Review dependent components before migration because the storage will be temporarily unavailable.
Shared Storage environments do not support live migration.
Shared Storage summary: Use the dedicated Shared Storage Container for centralized NFSv4 storage, enable Auto-Clustering when replicated GlusterFS storage and failover are required, use Mount Points and Exports to share data, and avoid using this storage as the primary location for SQL databases or high-I/O workloads.
