CubyCrate Changelog & Release History

This document tracks official release history, technical enhancements, security fixes, and version compatibility across all published versions of CubyCrate.

Release Summary

Version	Release Channel	Release Date	Target Image	Multi-Arch
`v0.0.1-alpha.2`	`alpha` (Current)	Initial Official Release	`ghcr.io/septian-padli/cubycrate:v0.0.1-alpha.2`	`linux/amd64`, `linux/arm64`
`v0.0.1-alpha.1`	Deprecated	Historical Pre-Release	`cubycloud:v0.0.1-alpha.1`	Historical single-arch

v0.0.1-alpha.2 (Current Official Release)

v0.0.1-alpha.2 is the first official release published under the canonical CubyCrate name. It establishes the stable baseline architecture, multi-architecture container distribution, and unified storage engine.

Key Enhancements

1. Official Canonical Rebrand to CubyCrate

Canonical Repository: github.com/septian-padli/cubycrate
Container Distribution: ghcr.io/septian-padli/cubycrate:v0.0.1-alpha.2
Go Backend Module: github.com/septian-padli/cubycrate/apps/backend
Default Database Path: /app/data/db/cubycrate.db (SQLite in WAL mode)
Default Named Volume: cubycrate_data
Session Authentication Cookie: cubycrate_token
Uniform Storage Namespace: cubycrate/{instance_id}/{node_id}/blobs/{blob_id}

2. Multi-Architecture Container Support

Official container images published simultaneously for:
- linux/amd64 (standard 64-bit x86 servers)
- linux/arm64 (Raspberry Pi 4/5, AWS Graviton, Apple Silicon servers, Ampere Altra)
Built on minimal Alpine Linux base images with a compressed image footprint under 150 MB and 30–50 MB baseline idle memory usage.

3. S3-Compatible Object Storage Driver

Integration with AWS Go SDK v2 for full S3-compliant object store compatibility (AWS S3, Cloudflare R2, MinIO, Backblaze B2, Ceph RADOS Gateway).
Direct Presigned Browser Uploads: Generates 15-minute presigned PUT URLs allowing client browsers to stream files directly into S3 buckets, bypassing application server CPU and bandwidth.
Envelope Encryption: S3 access credentials stored in SQLite are protected with authenticated AES-256-GCM encryption using a 256-bit master key (/app/data/secrets/master.key or CUBYCRATE_MASTER_KEY).

4. Asynchronous Storage Node Detachment & Migration

Added background migration engine (node_migration task).
Non-empty storage nodes can be detached via administrator API (POST /api/v1/admin/storage-nodes/:id/detach), safely streaming blobs to another configured node.
Validates SHA-256 checksums and exact byte counts with a strict zero-deletion safety invariant on errors.

5. Streaming ZIP Batch Downloads

Download multiple files or entire folder trees as a single, dynamically streamed ZIP archive.
Archive streaming writes directly to the HTTP response stream without creating intermediate archive files on disk.

6. SQLite WAL Mode Optimization

Embedded SQLite configured with PRAGMA journal_mode=WAL and PRAGMA busy_timeout=5000.
Concurrent readers do not block write operations, ensuring responsive UI browsing during heavy uploads.

7. Dual Health Checks

Standardized probe endpoints at GET /health and GET /api/health.
Both endpoints execute comprehensive SQLite pings and verify all configured storage nodes, returning 503 Service Unavailable if any node is offline.

Clean Break Notice: Upgrading from v0.0.1-alpha.1

Why a Clean Break?

v0.0.1-alpha.1 was an experimental proof-of-concept release that used legacy database schemas, differing blob directory conventions, and old environment variable keys. To guarantee long-term schema stability and cryptographic integrity, v0.0.1-alpha.2 establishes a fresh baseline.

Migration Procedure for v0.0.1-alpha.1 Operators

Do not point v0.0.1-alpha.2 at existing cubycloud_data volumes.
Deploy v0.0.1-alpha.2 with a fresh cubycrate_data volume.
Complete the first-run administrator setup at /setup.
Export or download existing files from your v0.0.1-alpha.1 deployment.
Upload files into the new CubyCrate instance via the web interface or TUS resumable uploads.