How plakarkorp/plakar works
Project type and purpose
• This is a Go-based, CLI-driven backup & restore solution (“plakar”) with a built-in HTTP server and web UI.
• It ships as a single binary exposing subcommands (backup, restore, sync, ui, agent, scheduler, etc.), a plugin/connector architecture for multiple storage backends, and embeds an immutable data store called Kloset.
File structure
• Top-level binaries & CLI:
• main.go + subcommands/… (each subcommand is its own package)
• agent/ (daemon to manage repository locks and remote access)
• scheduler/ (task scheduling client & config)
• server/httpd/ (REST API + static-file server for UI)
• ui/v2/frontend (static assets of the web UI)
• Core libraries & layers:
• api/ (defines REST endpoints and proxying)
• appcontext/ (context management across calls)
• connectors/ (plugin directory grouping fs, s3, ftp, grpc, sqlite, http, null, ptar, stdio, tar, etc.)
• plugins/ (higher-level integration hooks)
• plakarfs/ (FUSE filesystem implementation for instant mounts)
• reporting/ (metric/event reporting)
• utils/ (config, flags, time, browser integration)
• Configuration & build:
• Makefile, Go modules, GitHub Actions workflows, goreleaser, scheduler/config.yaml
README for additional insights
• plakar offers CLI subcommands, built-in web UI (served by HTTP server), a plugin-style connector architecture, encryption, deduplication, and multiple “storage backends” (local FS, S3, FTP, etc.).
• Quickstart shows that plakar ties together repository creation, backup, restore, sync, and UI launch under one binary.