feat(pruner, primitives): move prune batch sizes to ChainSpec

bin/reth/src/node/mod.rs

@@ -51,7 +51,6 @@ use reth_provider::{
     providers::BlockchainProvider, BlockHashReader, BlockReader, CanonStateSubscriptions,
     HeaderProvider, ProviderFactory, StageCheckpointReader,
 };
-use reth_prune::BatchSizes;
 use reth_revm::Factory;
 use reth_revm_inspectors::stack::Hook;
 use reth_rpc_engine_api::EngineApi;
@@ -429,7 +428,7 @@ impl<Ext: RethCliExt> NodeCommand<Ext> {
                 self.chain.clone(),
                 prune_config.block_interval,
                 prune_config.parts,
-                BatchSizes::default(),
+                self.chain.prune_batch_sizes,
             )
         });
 

crates/config/src/config.rs

@@ -283,14 +283,14 @@ impl Default for IndexHistoryConfig {
 #[serde(default)]
 pub struct PruneConfig {
     /// Minimum pruning interval measured in blocks.
-    pub block_interval: u64,
+    pub block_interval: usize,
     /// Pruning configuration for every part of the data that can be pruned.
     pub parts: PruneModes,
 }
 
 impl Default for PruneConfig {
     fn default() -> Self {
-        Self { block_interval: 10, parts: PruneModes::default() }
+        Self { block_interval: 5, parts: PruneModes::default() }
     }
 }
 

crates/consensus/beacon/src/engine/test_utils.rs

@@ -20,12 +20,12 @@ use reth_interfaces::{
     test_utils::{NoopFullBlockClient, TestConsensus},
 };
 use reth_payload_builder::test_utils::spawn_test_payload_service;
-use reth_primitives::{BlockNumber, ChainSpec, PruneModes, H256, U256};
+use reth_primitives::{BlockNumber, ChainSpec, PruneBatchSizes, PruneModes, H256, U256};
 use reth_provider::{
     providers::BlockchainProvider, test_utils::TestExecutorFactory, BlockExecutor, ExecutorFactory,
     ProviderFactory, StateProvider,
 };
-use reth_prune::{BatchSizes, Pruner};
+use reth_prune::Pruner;
 use reth_revm::Factory;
 use reth_rpc_types::engine::{ExecutionPayload, ForkchoiceState, ForkchoiceUpdated, PayloadStatus};
 use reth_stages::{
@@ -470,7 +470,7 @@ where
             self.base_config.chain_spec.clone(),
             5,
             PruneModes::default(),
-            BatchSizes::default(),
+            PruneBatchSizes::default(),
         );
 
         let (mut engine, handle) = BeaconConsensusEngine::new(

crates/primitives/src/chain/spec.rs

@@ -7,7 +7,7 @@ use crate::{
     header::Head,
     proofs::genesis_state_root,
     Address, BlockNumber, Chain, ForkFilter, ForkHash, ForkId, Genesis, Hardfork, Header,
-    SealedHeader, H160, H256, U256,
+    PruneBatchSizes, SealedHeader, H160, H256, U256,
 };
 use hex_literal::hex;
 use once_cell::sync::Lazy;
@@ -63,7 +63,8 @@ pub static MAINNET: Lazy<Arc<ChainSpec>> = Lazy::new(|| {
             11052984,
             H256(hex!("649bbc62d0e31342afea4e5cd82d4049e7e1ee912fc0889aa790803be39038c5")),
         )),
-        ..Default::default()
+        base_fee_params: BaseFeeParams::ethereum(),
+        prune_batch_sizes: PruneBatchSizes::mainnet(),
     }
     .into()
 });
@@ -104,7 +105,8 @@ pub static GOERLI: Lazy<Arc<ChainSpec>> = Lazy::new(|| {
             4367322,
             H256(hex!("649bbc62d0e31342afea4e5cd82d4049e7e1ee912fc0889aa790803be39038c5")),
         )),
-        ..Default::default()
+        base_fee_params: BaseFeeParams::ethereum(),
+        prune_batch_sizes: PruneBatchSizes::testnet(),
     }
     .into()
 });
@@ -149,7 +151,8 @@ pub static SEPOLIA: Lazy<Arc<ChainSpec>> = Lazy::new(|| {
             1273020,
             H256(hex!("649bbc62d0e31342afea4e5cd82d4049e7e1ee912fc0889aa790803be39038c5")),
         )),
-        ..Default::default()
+        base_fee_params: BaseFeeParams::ethereum(),
+        prune_batch_sizes: PruneBatchSizes::testnet(),
     }
     .into()
 });
@@ -203,7 +206,7 @@ pub struct BaseFeeParams {
 }
 
 impl BaseFeeParams {
-    /// Get the base fee parameters for ethereum mainnet
+    /// Get the base fee parameters for Ethereum mainnet
     pub const fn ethereum() -> BaseFeeParams {
         BaseFeeParams {
             max_change_denominator: EIP1559_DEFAULT_BASE_FEE_MAX_CHANGE_DENOMINATOR,
@@ -247,12 +250,18 @@ pub struct ChainSpec {
     /// The active hard forks and their activation conditions
     pub hardforks: BTreeMap<Hardfork, ForkCondition>,
 
-    /// The deposit contract deployed for PoS.
+    /// The deposit contract deployed for PoS
     #[serde(skip, default)]
     pub deposit_contract: Option<DepositContract>,
 
     /// The parameters that configure how a block's base fee is computed
     pub base_fee_params: BaseFeeParams,
+
+    /// The batch sizes for pruner, per block. In the actual pruner run it will be multiplied by
+    /// the amount of blocks between pruner runs to account for the difference in amount of new
+    /// data coming in.
+    #[serde(default)]
+    pub prune_batch_sizes: PruneBatchSizes,
 }
 
 impl Default for ChainSpec {
@@ -266,6 +275,7 @@ impl Default for ChainSpec {
             hardforks: Default::default(),
             deposit_contract: Default::default(),
             base_fee_params: BaseFeeParams::ethereum(),
+            prune_batch_sizes: Default::default(),
         }
     }
 }

crates/primitives/src/lib.rs

@@ -81,8 +81,8 @@ pub use net::{
 };
 pub use peer::{PeerId, WithPeerId};
 pub use prune::{
-    PruneCheckpoint, PruneMode, PruneModes, PrunePart, PrunePartError, ReceiptsLogPruneConfig,
-    MINIMUM_PRUNING_DISTANCE,
+    PruneBatchSizes, PruneCheckpoint, PruneMode, PruneModes, PrunePart, PrunePartError,
+    ReceiptsLogPruneConfig, MINIMUM_PRUNING_DISTANCE,
 };
 pub use receipt::{Receipt, ReceiptWithBloom, ReceiptWithBloomRef};
 pub use revm_primitives::JumpMap;

crates/primitives/src/prune/batch_sizes.rs

@@ -0,0 +1,83 @@
+use paste::paste;
+use serde::{Deserialize, Serialize};
+
+/// Batch sizes for configuring the pruner.
+/// The batch size for each prune part should be both large enough to prune the data which was
+/// generated with each new block, and small enough to not generate an excessive load on the
+/// database due to deletion of too many rows at once.
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize)]
+pub struct PruneBatchSizes {
+    /// Maximum number of receipts to prune, per block.
+    receipts: usize,
+    /// Maximum number of transaction lookup entries to prune, per block.
+    transaction_lookup: usize,
+    /// Maximum number of transaction senders to prune, per block.
+    transaction_senders: usize,
+    /// Maximum number of account history entries to prune, per block.
+    /// Measured in the number of `AccountChangeSet` table rows.
+    account_history: usize,
+    /// Maximum number of storage history entries to prune, per block.
+    /// Measured in the number of `StorageChangeSet` table rows.
+    storage_history: usize,
+}
+
+macro_rules! impl_prune_batch_size_methods {
+    ($(($human_name:expr, $name:ident)),+) => {
+        paste! {
+            impl PruneBatchSizes {
+                $(
+                    #[doc = concat!("Maximum number of ", $human_name, " to prune, accounting for the block interval.")]
+                    pub fn $name(&self, block_interval: usize) -> usize {
+                        self.$name * block_interval
+                    }
+
+                    #[doc = concat!("Set the maximum number of ", $human_name, " to prune per block.")]
+                    pub fn [<with_ $name>](mut self, batch_size: usize) -> Self {
+                        self.$name = batch_size;
+                        self
+                    }
+                )+
+            }
+        }
+    };
+}
+
+impl_prune_batch_size_methods!(
+    ("receipts", receipts),
+    ("transaction lookup entries", transaction_lookup),
+    ("transaction senders", transaction_senders),
+    ("account history entries", account_history),
+    ("storage history entries", storage_history)
+);
+
+impl PruneBatchSizes {
+    /// Default prune batch sizes for Ethereum mainnet.
+    /// These settings are sufficient to prune more data than generated with each new block.
+    pub const fn mainnet() -> Self {
+        Self {
+            receipts: 250,
+            transaction_lookup: 250,
+            transaction_senders: 250,
+            account_history: 1000,
+            storage_history: 1000,
+        }
+    }
+
+    /// Default prune batch sizes for Ethereum testnets.
+    /// These settings are sufficient to prune more data than generated with each new block.
+    pub const fn testnet() -> Self {
+        Self {
+            receipts: 100,
+            transaction_lookup: 100,
+            transaction_senders: 100,
+            account_history: 500,
+            storage_history: 500,
+        }
+    }
+}
+
+impl Default for PruneBatchSizes {
+    fn default() -> Self {
+        Self::mainnet()
+    }
+}

crates/primitives/src/prune/mod.rs

@@ -1,9 +1,11 @@
+mod batch_sizes;
 mod checkpoint;
 mod mode;
 mod part;
 mod target;
 
 use crate::{Address, BlockNumber};
+pub use batch_sizes::PruneBatchSizes;
 pub use checkpoint::PruneCheckpoint;
 pub use mode::PruneMode;
 pub use part::{PrunePart, PrunePartError};

crates/prune/src/lib.rs

@@ -4,4 +4,4 @@ mod pruner;
 
 use crate::metrics::Metrics;
 pub use error::PrunerError;
-pub use pruner::{BatchSizes, Pruner, PrunerResult, PrunerWithResult};
+pub use pruner::{Pruner, PrunerResult, PrunerWithResult};