Dedicated NAS system
HP EliteDesk 800 G3 SFF
- Intel Core i5-7500 processor
- 16 GB DDR4-2666 memory (4 × 4 GB)
- Dedicated storage and remote-access role
- Selected after physical compatibility checks
Loading...
Please wait while the page loads.Back to Projects
Completed & LiveInfrastructure Case Study
Self-Hosted Private Cloud Infrastructure
A dedicated NAS and private-cloud platform with mirrored storage, controlled user access, automated provisioning, secure remote connectivity, and a portfolio-integrated Drive experience.
TrueNAS SCALEZFSNext.jsSupabaseCloudflareHTTPSLinux
- Status
- Operational
- Usable storage
- Approximately 1.76 TiB
- Resilience
- Two-drive ZFS mirror
- Access model
- Authorised users only
Final Outcome
Tested, operational, and serving authorised users
The physical build, mirrored storage, user access, provisioning workflow, remote connectivity, and operational checks have been completed. The service is live and working against the original project requirements.
Project Overview
From a storage idea to a live service
The project began with a simple goal: create a private, Google Drive-like environment under my control. It became a complete infrastructure exercise covering hardware, ZFS, identity, automation, networking, security, monitoring, and recovery planning.
Create a private file-storage service with a familiar web experience
Provide separate authorised accounts and configurable storage quotas
Use mirrored storage to tolerate a single data-drive failure
Enable encrypted remote access without publishing the NAS administration interface
Automate user-storage provisioning and document the complete build
Hardware Design
Purpose-selected storage hardware
The completed physical build uses a compact business-class system configured specifically for storage. The operating system, applications, and user data are separated, while two dedicated data drives provide mirrored capacity and single-drive fault tolerance.
Physical compatibility was treated as a design requirement
Drive bays, mounting points, SATA connectivity, power, clearance, cooling, and serviceability were verified before the final platform was commissioned.
Primary data storage
Two-drive ZFS mirror
- 2 × 2 TB Seagate Exos SATA HDDs
- Approximately 1.76 TiB usable capacity
- Single-drive fault tolerance
- Health checks and scheduled maintenance
Separated workloads
Dedicated solid-state storage
- 128 GB NVMe SSD for TrueNAS boot
- Separate 128 GB SATA SSD for applications
- User data isolated from system workloads
- System and application I/O kept off the data mirror
The Costliest Mistake
Good specifications did not mean suitable hardware
My first computer looked capable by processor, memory, and advertised storage specifications. A physical inspection revealed limitations in drive bays, clearance, mounting, power, SATA connectivity, airflow, and expansion. Replacing it added time and cost, but prevented an unreliable build and became one of the project's most valuable lessons.
System Architecture
Public experience, private infrastructure
The public application handles the user experience while the NAS and administration services remain separated behind controlled access and provisioning layers.
Portfolio Drive Gateway
The public web application provides sign-in, account checks, and the user-facing Drive experience without exposing the TrueNAS interface.
Authorisation Layer
Role and account checks restrict the service to approved users before any storage operation is permitted.
Provisioning Workflow
A controlled background workflow processes approved requests and provisions isolated user storage without manual service reloads.
TrueNAS SCALE
TrueNAS manages the storage pools, datasets, permissions, applications, monitoring, and maintenance tasks.
Mirrored User Storage
ZFS datasets provide separated storage areas, capacity controls, and a resilient two-drive mirror for user data.
Implementation Journey
From planning to operation
Each stage was completed and validated before the system was treated as operational.
Idea & Requirements
CompletedDefined the need for private storage, remote use, individual accounts, quotas, resilience, and portfolio integration.
Research & Hardware Selection
CompletedCompared systems, drive technologies, physical bays, SATA connections, power availability, cooling, and expansion limits.
Compatibility Correction
CompletedReplaced the original computer after discovering that its internal layout could not reliably support the planned storage design.
Assembly & TrueNAS Deployment
CompletedInstalled the selected components, deployed TrueNAS SCALE, verified every drive, and completed the initial system configuration.
Storage & Access Configuration
CompletedCreated the ZFS mirror, datasets, permissions, quotas, application storage, authenticated access, and provisioning workflow.
Testing & Operation
CompletedValidated storage health, user separation, provisioning, file operations, remote access, backup tasks, and monitoring before launch.
Security Approach
Reducing exposure by design
Public documentation remains intentionally high level. Internal addresses, identities, paths, credentials, and configuration details are excluded.
Encrypted access
External sessions use HTTPS so credentials and file transfers are encrypted in transit.
Separated administration
The user-facing service remains separate from the NAS management interface and internal configuration.
Isolated user storage
Individual datasets, permissions, and quotas restrict users to their assigned storage areas.
Controlled provisioning
Approved requests are processed through a restricted workflow rather than granting direct administrative access.
Recovery planning
Replication and recovery tasks support resilience, while an independent backup remains a separate requirement.
Sanitised documentation
Public documentation explains the design without exposing addresses, credentials, identifiers, or private paths.
Operational Evidence
Healthy, monitored, and in use
The sanitised dashboard records the final operating state while removing administrative identity, software-version, hostname, and network-address information.
- Operational TrueNAS system with healthy storage pools
- Two-drive ZFS mirror with no reported disk errors during final validation
- Authenticated, role-protected access for approved users
- Automated creation of isolated user storage and quotas
- Successful remote file access through the portfolio Drive experience
- Backup tasks, resource monitoring, and storage-health visibility
Learning Outcomes
The project was more valuable because it was not straightforward
Specifications do not prove compatibility
Processor and memory specifications looked suitable on the first system, but drive bays, clearance, cabling, power, and airflow determined whether it could actually become a reliable NAS.
Redundancy and backup solve different problems
A mirror improves availability after one drive fails. It does not protect against deletion, corruption, theft, or every system-level failure, so independent backups are still necessary.
Access design must start with security
Authentication, authorisation, encryption, user isolation, and administrative separation were treated as architecture requirements rather than additions after deployment.
Automation needs repeatable testing
Provisioning was tested across the complete workflow so new approved users could receive storage without manual reloads or inconsistent permissions.
Continuous Improvement
Live does not mean finished forever
The platform is operational, but reliability and security require continued testing, maintenance, and documentation.
Independent backups
Strengthen off-system backups and test restoration regularly rather than assuming stored copies are recoverable.
Automated alerts
Expand notifications for disk health, storage capacity, failed jobs, service availability, and unusual conditions.
Operational monitoring
Develop clearer long-term visibility into performance, capacity growth, temperatures, and service health.
Periodic security reviews
Review accounts, permissions, updates, recovery procedures, and external exposure on a defined schedule.
Power-loss protection
Add UPS integration and verify graceful shutdown and restart behaviour during power interruption tests.
Recovery documentation
Keep build, maintenance, restoration, and incident procedures accurate as the platform evolves.
Project Complete
A working service and a reusable home-lab platform
The most important outcome is not simply the storage capacity. It is the practical experience of researching, correcting mistakes, integrating systems, validating security boundaries, automating operations, and moving an infrastructure idea into reliable use. Future work will build on this foundation through stronger recovery testing, monitoring, power protection, and additional controlled self-hosted services.
View All Projects