litestream

name: litestream description: SQLite disaster recovery and streaming replication to cloud storage (S3, GCS, Azure, SFTP, NATS). Use this skill for configuring Litestream, deploying to cloud platforms, troubleshooting WAL replication issues, implementing point-in-time recovery, and setting up VFS read replicas. version: 0.5.5

Litestream is a disaster recovery tool for SQLite that runs as a background process, continuously replicating database changes to cloud storage. It monitors the SQLite Write-Ahead Log (WAL), converts changes to immutable LTX files, and uploads them to your chosen storage backend.

Key Concepts

WAL Monitoring: Watches SQLite Write-Ahead Log for changes at configurable intervals (default 1s)
LTX Files: Immutable files containing database page changes, never modified after creation
Level-Based Compaction: Multi-level system (L0-L9) that merges LTX files to reduce storage overhead
Point-in-Time Recovery: Restore to any transaction using TXID or timestamp
VFS Support: Read replicas directly from cloud storage without local database copy
MCP Server: AI tool integration for automated database management

Quick Start

Installation

# macOS
brew install litestream

# Linux (Debian/Ubuntu)
wget https://github.com/benbjohnson/litestream/releases/download/v0.5.5/litestream-v0.5.5-linux-amd64.deb
sudo dpkg -i litestream-v0.5.5-linux-amd64.deb

# Docker
docker pull litestream/litestream:0.5

Basic Replication

# Command line (single database)
litestream replicate /path/to/db.sqlite s3://bucket/db

# With config file
litestream replicate -config /etc/litestream.yml

Restore Database

# Restore latest backup
litestream restore -o /path/to/restored.db s3://bucket/db

# Point-in-time recovery
litestream restore -timestamp 2024-01-15T10:30:00Z -o /tmp/db s3://bucket/db

# Restore to specific transaction
litestream restore -txid 1000 -o /tmp/db s3://bucket/db

Check Status

# View all databases and their replication status
litestream status

# List available LTX files
litestream ltx /path/to/db.sqlite

Critical Rules

These rules are essential for correct Litestream operation:

Lock Page at 1GB: SQLite reserves the page at offset 0x40000000 (1GB). This page must always be skipped during replication and compaction. The page number varies by page size:
- 4KB pages: page 262145
- 8KB pages: page 131073
- 16KB pages: page 65537
LTX Files are Immutable: Once created, LTX files are never modified. New changes create new files.
Single Replica per Database: Each database can only replicate to one destination. Use multiple Litestream instances for multiple destinations.
WAL Mode Required: SQLite must be in WAL mode (PRAGMA journal_mode=WAL;)
Use litestream ltx: The litestream wal command is deprecated.

Storage Backends

Backend	URL Scheme	Example
AWS S3	`s3://`	`s3://bucket/path`
S3-Compatible (R2, Tigris, MinIO)	`s3://`	`s3://bucket/path?endpoint=host:port`
Google Cloud Storage	`gs://`	`gs://bucket/path`
Azure Blob Storage	`abs://`	`abs://container@account/path`
SFTP	`sftp://`	`sftp://user@host:22/path`
NATS JetStream	`nats://`	`nats://host:4222/bucket`
WebDAV	`webdav://`	`webdav://host/path`
Alibaba OSS	`oss://`	`oss://bucket/path`
Local File	(path)	`/var/backups/db`

Configuration File Example

dbs:
  - path: /data/app.db
    replica:
      url: s3://my-bucket/app-backup
      access-key-id: ${AWS_ACCESS_KEY_ID}
      secret-access-key: ${AWS_SECRET_ACCESS_KEY}
      region: us-east-1
      sync-interval: 1s

When to Use This Skill

Use Litestream skill when:

Configuring SQLite database replication to cloud storage
Setting up disaster recovery for SQLite applications
Troubleshooting replication issues (WAL growth, sync failures)
Implementing point-in-time recovery procedures
Deploying Litestream with Docker, Fly.io, Kubernetes, or systemd
Understanding WAL-based replication concepts

Skill Contents

concepts/ - Architecture, replication mechanics, LTX format, SQLite WAL, VFS support, compaction
configuration/ - Storage backend configurations (S3, GCS, Azure, SFTP, NATS, OSS, WebDAV, File)
commands/ - CLI command reference (replicate, restore, status, ltx, databases)
operations/ - Monitoring, troubleshooting, recovery, and heartbeat monitoring
deployment/ - Docker, Fly.io, Kubernetes, and systemd deployment guides
integrations/ - MCP server integration for AI tools
scripts/ - Validation and diagnostic helper scripts

Common Issues

WAL Growing Too Large

Check sync interval (default 1s may be too slow for write-heavy workloads)
Verify storage backend connectivity
Check for checkpoint blocking (long-running transactions)

Replication Lag

Monitor with litestream status
Check network connectivity to storage backend
Review sync interval configuration

Restore Failures

Verify backup exists: litestream ltx /path/to/db
Check storage backend credentials
Ensure target directory is writable