The Graveyard Book

How The Graveyard Book Actually Works

The Graveyard Book is a utility for managing data retention and archival workflows. You use it to take whatever you need to keep and set it aside so it doesn't interfere with your active environment, then bring it back when necessary. That's the whole premise. Most people overcomplicate this. They start trying to configure custom policies before they even have a working backup. The process is simpler than the documentation makes it look.

Installing and Setting Up The Graveyard Book

You download it from the official repository. At this point there's no single canonical source, but the project page on GitHub is the closest thing to an authority. The binary or archive will come with a README that covers installation for your platform. If you're on Linux or macOS, you can usually just extract it and run it. Windows users may need to adjust execution policies depending on their setup. Once installed, you run the init command to create your first graveyard. This sets up the directory structure where archived items live. The default location is ~/.graveyard unless you specify otherwise with the --config flag.

Core Workflow With The Graveyard Book

Here's how you actually use it day to day. You have data that needs to be preserved but isn't actively needed. Maybe it's old project files, maybe it's historical logs, maybe it's a dataset you loaded once and rarely query. You point The Graveyard Book at it, give it a label, and it creates a snapshot. The snapshot includes metadata — timestamps, file hashes, size — so you can verify integrity later when you restore. The restore process is where people run into trouble. You might encounter a situation where the snapshot references a path that no longer exists on your system. This happened to me when I was archiving an entire development environment and then the drive layout changed. The fix is straightforward: you run the restore with the --force-path flag, which remaps the archived paths to your current structure instead of failing outright.

I spent about two hours debugging a failed restore last year because I didn't realize the snapshot was using absolute paths from the original machine. Switching to relative paths before archiving would have prevented that entirely.

Advanced Usage Patterns

There are a few things beginners miss. The first is that The Graveyard Book supports differential snapshots. You don't need to archive everything from scratch each time. Once you have a base snapshot, subsequent snapshots only store what changed. This cuts storage requirements significantly after the initial full backup. If you're managing something like a database export that stays relatively stable, your delta sizes tend to drop to a few megabytes per run after the first week. The second is retention pruning. The tool can automatically clean up old snapshots beyond a certain age or count. I've seen people run The Graveyard Book for months without pruning and then wonder why their storage filled up unexpectedly. Setting a retention policy — even something simple like keeping the last five snapshots and anything older than 90 days — prevents this.

Limits and When It Fails

The Graveyard Book isn't a full backup solution. It doesn't handle incremental file-level backups across distributed systems. If you need that, you should be looking at something like BorgBackup or Restic instead. This tool is designed for a specific niche: discrete datasets that you want to archive and retrieve cleanly. It also doesn't encrypt by default. If your data is sensitive, you'll need to layer encryption on top, either by encrypting before archiving or by using a filesystem-level solution. I personally compress and encrypt my data with GPG before passing it to The Graveyard Book. It adds a step but it's not complicated. Another limitation: concurrent access. The tool locks snapshots during creation, which means you can't safely archive a directory that's actively being written to. You'll get corrupted snapshots if you do. The workaround is to freeze writes first or use a tool like lvm snapshot or zfs snapshot to create a point-in-time copy, then archive from that.

Getting Started

If you're working with static datasets and want something lightweight for archiving, The Graveyard Book is worth looking into. The Graveyard Book is still a niche tool, so documentation is sparse and community support is limited, but the core functionality works reliably when used within its intended scope. Grab it from the project's official page and start with a single directory. Don't try to archive your entire system on day one.