erigon-seg-retire

name	erigon-seg-retire
description	Run the Erigon segment retire command to build, merge, and clean snapshot files. Use this for snapshot publication readiness preparation.
allowed-tools	Bash, Read

Erigon Segment Retire Command

The erigon seg retire command prepares snapshot files for publication by performing build, merge, and cleanup operations on segment files.

Prerequisites

1. Erigon must be stopped - The command requires exclusive access to the datadir
2. Binary must be built - Run /erigon-build first if the binary doesn't exist
3. Datadir must exist - A synced Erigon datadir with blockchain data

Command

./build/bin/erigon seg retire --datadir=<path>

Replace <path> with the actual path to your Erigon data directory.

What the Command Does

The retire command performs these operations in sequence:

1. Build Missing Indices - Creates any missing indices for block and Caplin snapshots
2. Retire Blocks - Freezes blocks from the database into snapshot files
3. Remove Overlaps - Cleans up overlapping snapshot files
4. Prune Ancient Blocks - Removes block data from the database that has been successfully snapshotted
5. Build State History - Creates missing accessors and builds state history snapshot files
6. Prune State History - Removes state history that has been snapshotted
7. Merge and Cleanup - Merges smaller snapshot files into larger ones and removes overlaps

Usage in Snapshot Release Workflow

This command is part of the "Publishable v2" snapshot release flow:

1. Shutdown Erigon
2. Run seg retire (this command)
3. Optionally run seg clearIndexing
4. Run seg index
5. Create torrent files
6. Run integrity check
7. Run publishable command

Example

# Build the binary first (if needed)
make erigon

# Ensure Erigon is stopped, then run seg retire
./build/bin/erigon seg retire --datadir=/data/erigon

Important Notes

• File Lock: If Erigon is running, the command will fail due to file lock
• Long running: This command can take significant time on mainnet
• Disk I/O intensive: Performs extensive read/write operations
• Resource intensive: Designed to maximize resource utilization