add reth-prune-static-files

joshieDo · joshieDo · commit d422073bb93e · 2025-10-22T19:34:18.000+01:00
diff --git a/.github/assets/check_wasm.sh b/.github/assets/check_wasm.sh
@@ -68,6 +68,7 @@ exclude_crates=(
   reth-payload-builder # reth-metrics
   reth-provider # tokio
   reth-prune # tokio
+  reth-prune-static-files # reth-provider
   reth-stages-api # reth-provider, reth-prune
   reth-static-file # tokio
   reth-transaction-pool # c-kzg
diff --git a/Cargo.lock b/Cargo.lock
diff --git a/Cargo.toml b/Cargo.toml
@@ -94,6 +94,7 @@ members = [
     "crates/primitives-traits/",
     "crates/primitives/",
     "crates/prune/prune",
+    "crates/prune/static-files",
     "crates/prune/types",
     "crates/ress/protocol",
     "crates/ress/provider",
@@ -429,6 +430,7 @@ reth-primitives = { path = "crates/primitives", default-features = false }
 reth-primitives-traits = { path = "crates/primitives-traits", default-features = false }
 reth-provider = { path = "crates/storage/provider" }
 reth-prune = { path = "crates/prune/prune" }
+reth-prune-static-files = { path = "crates/prune/static-files" }
 reth-prune-types = { path = "crates/prune/types", default-features = false }
 reth-revm = { path = "crates/revm", default-features = false }
 reth-rpc = { path = "crates/rpc/rpc" }
diff --git a/crates/node/builder/src/launch/common.rs b/crates/node/builder/src/launch/common.rs
@@ -34,12 +34,11 @@ use crate::{
     hooks::OnComponentInitializedHook,
     BuilderContext, ExExLauncher, NodeAdapter, PrimitivesTy,
 };
-use alloy_consensus::BlockHeader as _;
 use alloy_eips::eip2124::Head;
 use alloy_primitives::{BlockNumber, B256};
 use eyre::Context;
 use rayon::ThreadPoolBuilder;
-use reth_chainspec::{Chain, EthChainSpec, EthereumHardfork, EthereumHardforks};
+use reth_chainspec::{Chain, EthChainSpec, EthereumHardforks};
 use reth_config::{config::EtlConfig, PruneConfig};
 use reth_consensus::noop::NoopConsensus;
 use reth_db_api::{database::Database, database_metrics::DatabaseMetrics};
@@ -67,8 +66,8 @@ use reth_node_metrics::{
 };
 use reth_provider::{
     providers::{NodeTypesForProvider, ProviderNodeTypes, StaticFileProvider},
-    BlockHashReader, BlockNumReader, BlockReaderIdExt, ProviderError, ProviderFactory,
-    ProviderResult, StageCheckpointReader, StaticFileProviderFactory,
+    BlockHashReader, BlockNumReader, ProviderError, ProviderFactory, ProviderResult,
+    StageCheckpointReader, StaticFileProviderFactory,
 };
 use reth_prune::{PruneModes, PrunerBuilder};
 use reth_rpc_builder::config::RethRpcServerConfig;
@@ -945,40 +944,6 @@ where
         Ok(None)
     }
 
-    /// Expire the pre-merge transactions if the node is configured to do so and the chain has a
-    /// merge block.
-    ///
-    /// If the node is configured to prune pre-merge transactions and it has synced past the merge
-    /// block, it will delete the pre-merge transaction static files if they still exist.
-    pub fn expire_pre_merge_transactions(&self) -> eyre::Result<()>
-    where
-        T: FullNodeTypes<Provider: StaticFileProviderFactory>,
-    {
-        if self.node_config().pruning.bodies_pre_merge &&
-            let Some(merge_block) = self
-                .chain_spec()
-                .ethereum_fork_activation(EthereumHardfork::Paris)
-                .block_number()
-        {
-            // Ensure we only expire transactions after we synced past the merge block.
-            let Some(latest) = self.blockchain_db().latest_header()? else { return Ok(()) };
-            if latest.number() > merge_block {
-                let provider = self.blockchain_db().static_file_provider();
-                if provider
-                    .get_lowest_transaction_static_file_block()
-                    .is_some_and(|lowest| lowest < merge_block)
-                {
-                    info!(target: "reth::cli", merge_block, "Expiring pre-merge transactions");
-                    provider.delete_transactions_below(merge_block)?;
-                } else {
-                    debug!(target: "reth::cli", merge_block, "No pre-merge transactions to expire");
-                }
-            }
-        }
-
-        Ok(())
-    }
-
     /// Returns the metrics sender.
     pub fn sync_metrics_tx(&self) -> UnboundedSender<MetricEvent> {
         self.right().db_provider_container.metrics_sender.clone()
diff --git a/crates/node/builder/src/launch/engine.rs b/crates/node/builder/src/launch/engine.rs
@@ -117,9 +117,6 @@ impl EngineNodeLauncher {
             })?
             .with_components(components_builder, on_component_initialized).await?;
 
-        // Try to expire pre-merge transaction history if configured
-        ctx.expire_pre_merge_transactions()?;
-
         // spawn exexs if any
         let maybe_exex_manager_handle = ctx.launch_exex(installed_exex).await?;
 
diff --git a/crates/node/core/src/args/pruning.rs b/crates/node/core/src/args/pruning.rs
@@ -126,8 +126,10 @@ impl PruningArgs {
                     receipts: Some(PruneMode::Distance(MINIMUM_PRUNING_DISTANCE)),
                     account_history: Some(PruneMode::Distance(MINIMUM_PRUNING_DISTANCE)),
                     storage_history: Some(PruneMode::Distance(MINIMUM_PRUNING_DISTANCE)),
-                    // TODO: set default to pre-merge block if available
-                    bodies_history: None,
+                    bodies_history: chain_spec
+                        .ethereum_fork_activation(EthereumHardfork::Paris)
+                        .block_number()
+                        .map(PruneMode::Before),
                     merkle_changesets: PruneMode::Distance(MINIMUM_PRUNING_DISTANCE),
                     #[expect(deprecated)]
                     receipts_log_filter: (),
diff --git a/crates/prune/prune/Cargo.toml b/crates/prune/prune/Cargo.toml
@@ -19,6 +19,7 @@ reth-errors.workspace = true
 reth-provider.workspace = true
 reth-tokio-util.workspace = true
 reth-config.workspace = true
+reth-prune-static-files.workspace = true
 reth-prune-types.workspace = true
 reth-primitives-traits.workspace = true
 reth-static-file-types.workspace = true
diff --git a/crates/prune/prune/src/segments/mod.rs b/crates/prune/prune/src/segments/mod.rs
@@ -15,8 +15,8 @@ pub use static_file::{
 use std::{fmt::Debug, ops::RangeInclusive};
 use tracing::error;
 pub use user::{
-    AccountHistory, MerkleChangeSets, Receipts as UserReceipts, SenderRecovery, StorageHistory,
-    TransactionLookup,
+    AccountHistory, BodiesAdapter, MerkleChangeSets, Receipts as UserReceipts, SenderRecovery,
+    StorageHistory, TransactionLookup,
 };
 
 /// A segment represents a pruning of some portion of the data.
diff --git a/crates/prune/prune/src/segments/set.rs b/crates/prune/prune/src/segments/set.rs
@@ -1,6 +1,6 @@
 use crate::segments::{
-    AccountHistory, MerkleChangeSets, Segment, SenderRecovery, StorageHistory, TransactionLookup,
-    UserReceipts,
+    AccountHistory, BodiesAdapter, MerkleChangeSets, Segment, SenderRecovery, StorageHistory,
+    TransactionLookup, UserReceipts,
 };
 use alloy_eips::eip2718::Encodable2718;
 use reth_db_api::{table::Value, transaction::DbTxMut};
@@ -68,7 +68,7 @@ where
             receipts,
             account_history,
             storage_history,
-            bodies_history: _,
+            bodies_history,
             merkle_changesets,
             receipts_log_filter: (),
         } = prune_modes;
@@ -92,6 +92,8 @@ where
             .segment_opt(transaction_lookup.map(TransactionLookup::new))
             // Sender recovery
             .segment_opt(sender_recovery.map(SenderRecovery::new))
+            // Bodies
+            .segment_opt(bodies_history.map(BodiesAdapter::new))
     }
 }
 
diff --git a/crates/prune/prune/src/segments/user/bodies.rs b/crates/prune/prune/src/segments/user/bodies.rs
@@ -0,0 +1,44 @@
+use crate::{
+    segments::{PruneInput, Segment},
+    PrunerError,
+};
+use reth_provider::{BlockReader, StaticFileProviderFactory};
+use reth_prune_static_files::Bodies;
+use reth_prune_types::{PruneMode, PrunePurpose, PruneSegment, SegmentOutput};
+
+/// Segment adapter for pruning transactions in static files.
+///
+/// This wraps the implementation from `reth_prune_static_files` and implements
+/// the `Segment` trait required by the pruner.
+#[derive(Debug)]
+pub struct BodiesAdapter {
+    inner: Bodies,
+}
+
+impl BodiesAdapter {
+    /// Creates a new [`BodiesAdapter`] segment with the given prune mode.
+    pub const fn new(mode: PruneMode) -> Self {
+        Self { inner: Bodies::new(mode) }
+    }
+}
+
+impl<Provider> Segment<Provider> for BodiesAdapter
+where
+    Provider: StaticFileProviderFactory + BlockReader,
+{
+    fn segment(&self) -> PruneSegment {
+        self.inner.segment()
+    }
+
+    fn mode(&self) -> Option<PruneMode> {
+        self.inner.mode()
+    }
+
+    fn purpose(&self) -> PrunePurpose {
+        self.inner.purpose()
+    }
+
+    fn prune(&self, provider: &Provider, input: PruneInput) -> Result<SegmentOutput, PrunerError> {
+        Ok(self.inner.prune(provider, input.to_block)?)
+    }
+}
diff --git a/crates/prune/prune/src/segments/user/mod.rs b/crates/prune/prune/src/segments/user/mod.rs
@@ -1,4 +1,5 @@
 mod account_history;
+mod bodies;
 mod history;
 mod merkle_change_sets;
 mod receipts;
@@ -7,6 +8,7 @@ mod storage_history;
 mod transaction_lookup;
 
 pub use account_history::AccountHistory;
+pub use bodies::BodiesAdapter;
 pub use merkle_change_sets::MerkleChangeSets;
 pub use receipts::Receipts;
 pub use sender_recovery::SenderRecovery;
diff --git a/crates/prune/static-files/Cargo.toml b/crates/prune/static-files/Cargo.toml
@@ -0,0 +1,25 @@
+[package]
+name = "reth-prune-static-files"
+version.workspace = true
+edition.workspace = true
+rust-version.workspace = true
+license.workspace = true
+homepage.workspace = true
+repository.workspace = true
+
+[lints]
+workspace = true
+
+[dependencies]
+# reth
+reth-prune-types.workspace = true
+reth-provider.workspace = true
+reth-static-file-types.workspace = true
+reth-storage-errors.workspace = true
+
+# alloy
+alloy-primitives.workspace = true
+
+[dev-dependencies]
+reth-provider = { workspace = true, features = ["test-utils"] }
+reth-prune-types = { workspace = true, features = ["test-utils"] }
diff --git a/crates/prune/static-files/src/bodies.rs b/crates/prune/static-files/src/bodies.rs
@@ -0,0 +1,75 @@
+//! Transaction pruning functionality for static files.
+
+use alloy_primitives::BlockNumber;
+use reth_provider::{BlockReader, StaticFileProviderFactory};
+use reth_prune_types::{
+    PruneMode, PruneProgress, PrunePurpose, PruneSegment, SegmentOutput, SegmentOutputCheckpoint,
+};
+use reth_static_file_types::StaticFileSegment;
+use reth_storage_errors::provider::ProviderResult;
+
+/// Segment responsible for pruning transactions in static files.
+///
+/// This segment is controlled by the `bodies_history` configuration.
+#[derive(Debug)]
+pub struct Bodies {
+    mode: Option<PruneMode>,
+}
+
+impl Bodies {
+    /// Creates a new [`Bodies`] segment with the given prune mode.
+    pub const fn new(mode: PruneMode) -> Self {
+        Self { mode: Some(mode) }
+    }
+
+    /// Returns the segment type.
+    pub const fn segment(&self) -> PruneSegment {
+        PruneSegment::Bodies
+    }
+
+    /// Returns the prune mode.
+    pub const fn mode(&self) -> Option<PruneMode> {
+        self.mode
+    }
+
+    /// Returns the purpose of this segment.
+    pub const fn purpose(&self) -> PrunePurpose {
+        PrunePurpose::User
+    }
+
+    /// Prunes transactions in static files based on the target block.
+    ///
+    /// Read `StaticFileProvider::delete_segment_below_block` for more on static file pruning.
+    pub fn prune<Provider>(
+        &self,
+        provider: &Provider,
+        target_block: BlockNumber,
+    ) -> ProviderResult<SegmentOutput>
+    where
+        Provider: StaticFileProviderFactory + BlockReader,
+    {
+        let deleted_headers = provider
+            .static_file_provider()
+            .delete_segment_below_block(StaticFileSegment::Transactions, target_block)?;
+
+        if deleted_headers.is_empty() {
+            return Ok(SegmentOutput::done())
+        }
+
+        let tx_ranges = deleted_headers.iter().filter_map(|header| header.tx_range());
+
+        let pruned = tx_ranges
+            .clone()
+            .map(|range| range.end().saturating_sub(range.start()).saturating_add(1))
+            .sum::<u64>() as usize;
+
+        Ok(SegmentOutput {
+            progress: PruneProgress::Finished,
+            pruned,
+            checkpoint: Some(SegmentOutputCheckpoint {
+                block_number: Some(target_block.saturating_sub(1)),
+                tx_number: tx_ranges.map(|range| range.end()).max(),
+            }),
+        })
+    }
+}
diff --git a/crates/prune/static-files/src/lib.rs b/crates/prune/static-files/src/lib.rs
@@ -0,0 +1,14 @@
+//! Pruning for static files.
+//!
+//! This crate provides pruning functionality for data stored in static files.
+
+#![doc(
+    html_logo_url = "https://raw.githubusercontent.com/paradigmxyz/reth/main/assets/reth-docs.png",
+    html_favicon_url = "https://avatars0.githubusercontent.com/u/97369466?s=256",
+    issue_tracker_base_url = "https://github.com/paradigmxyz/reth/issues/"
+)]
+#![cfg_attr(docsrs, feature(doc_cfg))]
+#![cfg_attr(not(test), warn(unused_crate_dependencies))]
+
+mod bodies;
+pub use bodies::Bodies;
diff --git a/crates/prune/types/src/segment.rs b/crates/prune/types/src/segment.rs
@@ -31,6 +31,8 @@ pub enum PruneSegment {
     /// Prune segment responsible for all rows in `AccountsTrieChangeSets` and
     /// `StoragesTrieChangeSets` table.
     MerkleChangeSets,
+    /// Prune segment responsible for bodies (transactions in static files).
+    Bodies,
 }
 
 #[cfg(test)]
@@ -53,6 +55,7 @@ impl PruneSegment {
             Self::AccountHistory |
             Self::StorageHistory |
             Self::MerkleChangeSets |
+            Self::Bodies |
             Self::Receipts => MINIMUM_PRUNING_DISTANCE,
         }
     }
diff --git a/crates/storage/provider/src/providers/static_file/manager.rs b/crates/storage/provider/src/providers/static_file/manager.rs
