Bitcask implementation in Go

This project is an implementation of the Bitcask key-value store in Go, inspired by Coding Challenge #97 by John Crickett from Subsatck Coding Challenges.

What is Bitcask?

Bitcask is a high-performance, append-only key-value store originally developed by Justin Sheehy and David Smith (with inspiration from Eric Brewer). It uses log-structured data files and an in-memory key directory for fast reads and writes, making it simple yet powerful.

Motivation

This is a learning project to deepen my understanding of storage engines, file I/O, and database internals while practicing Go.

Setup

Git Hooks

To ensure code quality, there are Git hooks for linting and testing before pushing.

Run this command once to enable the custom hooks:

git config core.hooksPath .githooks

Development

Running the CLI

You can interact with the Bitcask database using go run:

Usage: gocask [options] <command> [args]

Options:
  --db <path>     Path to the database (default "./database")
  -h, --help      Show this help message

Commands (single-command mode):
  set <key> <value>     Store a value
  get <key>             Retrieve a value
  del <key>             Delete a value

Interactive mode:
  Simply run 'gocask' without commands to enter interactive REPL.
  Type commands like 'set <key> <value>', 'get <key>' or 'del <key>'.
  Use 'exit' or Ctrl+D to quit.

Examples:
  gocask set name Sirius --db ./mydb
  gocask get name --db ./mydb
  gocask # enter interactive mode

Run the test suite:

go test ./... -v

Run the linter:

golangci-lint run

This project also has a GitHub Actions CI workflow that runs tests and lint checks on every push and pull request to main.

Todo / Next Steps

• Basic in-memory Set and Get
• Set and Get with file persistence (append-only log)
• REPL
• Support for multiple data files
• Delete functionality using tombstones
• Merge/compaction functionality to clean up deleted and overwritten keys
• Optional / future enhancements:
  • Hint file for faster keydir loading.
  • Configurable maximum file size.
  • Automatic file rotation.
  • Concurrency safety.

References

• Bitcask: A Log-Structured Hash Table for Fast Key/Value Data
• Coding Challenges