Data Pruning

When a validator node is running, it participates in consensus to execute transactions and commit new data to the blockchain. Similarly, when fullnodes are running, they sync the new blockchain data through state synchronization. As the blockchain grows, storage disk space can be managed by pruning old blockchain data. Specifically, by pruning the ledger history: which contains old transactions. By default, ledger pruning is enabled on all nodes with a pruning window that can be configured. This document describes how you can configure the pruning behavior.

note

By default the ledger pruner keeps 150 million recent transactions. The approximate amount of disk space required for every 150M transactions is 200G. Unless bootstrapped from the genesis and configured to disable the pruner or a long prune window, the node doesn't carry the entirety of the ledger history. Majority of the nodes on both the testnet and mainnet have a partial history of 150 million transactions according to this configuration.

To manage these settings, edit the node configuration YAML files, for example, fullnode.yaml for fullnodes (validator or public) or validator.yaml for validator nodes, as shown below.

Disabling the ledger pruner

Add the following to the node configuration YAML file to disable the ledger pruner:

Proceed with caution

Disabling the ledger pruner can result in the storage disk filling up very quickly.

storage:
  storage_pruner_config:
    ledger_pruner_config:
      enable: false

Configuring the ledger pruning window

Add the following to the node configuration YAML file to make the node retain, for example, 1 billion transactions and their outputs, including events and write sets.

Proceed with caution

Setting the pruning window smaller than 100 million can lead to runtime errors and damage the health of the node.

storage:
  storage_pruner_config:
    ledger_pruner_config:
      prune_window: 1000000000

See the complete set of storage configuration settings in the Storage README.