Make ETL file size configurable (#6927)

Co-authored-by: Alexey Shekhirin <a.shekhirin@gmail.com>
Co-authored-by: joshieDo <ranriver@protonmail.com>

bin/reth/src/commands/debug_cmd/execution.rs

@@ -127,6 +127,7 @@ impl Command {
                     header_downloader,
                     body_downloader,
                     factory.clone(),
+                    stage_conf.etl.etl_file_size,
                 )
                 .set(SenderRecoveryStage {
                     commit_threshold: stage_conf.sender_recovery.commit_threshold,

bin/reth/src/commands/import.rs

@@ -178,6 +178,9 @@ impl ImportCommand {
             reth_revm::EvmProcessorFactory::new(self.chain.clone(), EthEvmConfig::default());
 
         let max_block = file_client.max_block().unwrap_or(0);
+
+        let etl_file_size = config.stages.etl.etl_file_size;
+
         let mut pipeline = Pipeline::builder()
             .with_tip_sender(tip_tx)
             // we want to sync all blocks the file client provides or 0 if empty
@@ -190,6 +193,7 @@ impl ImportCommand {
                     header_downloader,
                     body_downloader,
                     factory.clone(),
+                    etl_file_size,
                 )
                 .set(SenderRecoveryStage {
                     commit_threshold: config.stages.sender_recovery.commit_threshold,

bin/reth/src/commands/stage/run.rs

@@ -82,6 +82,10 @@ pub struct Command {
     #[arg(long)]
     batch_size: Option<u64>,
 
+    /// The maximum size in bytes of data held in memory before being flushed to disk as a file.
+    #[arg(long)]
+    etl_file_size: Option<usize>,
+
     /// Normally, running the stage requires unwinding for stages that already
     /// have been run, in order to not rewrite to the same database slots.
     ///
@@ -151,6 +155,8 @@ impl Command {
 
         let batch_size = self.batch_size.unwrap_or(self.to - self.from + 1);
 
+        let etl_file_size = self.etl_file_size.unwrap_or(500 * 1024 * 1024);
+
         let (mut exec_stage, mut unwind_stage): (Box<dyn Stage<_>>, Option<Box<dyn Stage<_>>>) =
             match self.stage {
                 StageEnum::Bodies => {
@@ -229,7 +235,7 @@ impl Command {
                     )
                 }
                 StageEnum::TxLookup => {
-                    (Box::new(TransactionLookupStage::new(batch_size, None)), None)
+                    (Box::new(TransactionLookupStage::new(batch_size, etl_file_size, None)), None)
                 }
                 StageEnum::AccountHashing => {
                     (Box::new(AccountHashingStage::new(1, batch_size)), None)

book/cli/reth/node.md

@@ -238,7 +238,7 @@ RPC:
       --rpc-max-tracing-requests <COUNT>
           Maximum number of concurrent tracing requests
 
-          [default: 8]
+          [default: 10]
 
       --rpc-max-blocks-per-filter <COUNT>
           Maximum number of blocks that could be scanned per filter request. (0 = entire chain)

book/cli/reth/stage/run.md

@@ -61,6 +61,9 @@ Options:
       --batch-size <BATCH_SIZE>
           Batch size for stage execution and unwind
 
+      --etl-file-size <ETL_FILE_SIZE>
+          Size for temporary file during ETL stages
+
   -s, --skip-unwind
           Normally, running the stage requires unwinding for stages that already have been run, in order to not rewrite to the same database slots.
 

book/run/config.md

@@ -221,6 +221,19 @@ The storage history indexing stage builds an index of what blocks a particular s
 commit_threshold = 100000
 ```
 
+### `etl`
+
+An ETL (extract, transform, load) data collector. Used mainly to insert data into `MDBX` in a sorted manner.
+
+```toml
+[stages.etl]
+# The maximum size in bytes of data held in memory before being flushed to disk as a file.
+#
+# Lower threshold corresponds to more frequent flushes,
+# but lowers temporary storage usage
+file_size = 524_288_000 # 500 * 1024 * 1024
+```
+
 ## The `[peers]` section
 
 The peers section is used to configure how the networking component of reth establishes and maintains connections to peers.

crates/config/src/config.rs

@@ -70,6 +70,8 @@ pub struct StageConfig {
     pub index_account_history: IndexHistoryConfig,
     /// Index Storage History stage configuration.
     pub index_storage_history: IndexHistoryConfig,
+    /// Common ETL related configuration.
+    pub etl: EtlConfig,
 }
 
 /// Header stage configuration.
@@ -235,7 +237,21 @@ impl Default for TransactionLookupConfig {
     }
 }
 
-/// History History stage configuration.
+/// Common ETL related configuration.
+#[derive(Debug, Clone, Copy, Deserialize, PartialEq, Serialize)]
+#[serde(default)]
+pub struct EtlConfig {
+    /// The maximum size in bytes of data held in memory before being flushed to disk as a file.
+    pub etl_file_size: usize,
+}
+
+impl Default for EtlConfig {
+    fn default() -> Self {
+        Self { etl_file_size: 500 * (1024 * 1024) }
+    }
+}
+
+/// History stage configuration.
 #[derive(Debug, Clone, Copy, Deserialize, PartialEq, Serialize)]
 #[serde(default)]
 pub struct IndexHistoryConfig {

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

@@ -406,6 +406,7 @@ where
                     header_downloader,
                     body_downloader,
                     executor_factory.clone(),
+                    500 * (1024 * 1024),
                 ))
             }
         };

crates/node-core/src/node_config.rs

@@ -837,6 +837,7 @@ impl NodeConfig {
                     header_downloader,
                     body_downloader,
                     factory.clone(),
+                    stage_config.etl.etl_file_size,
                 )
                 .set(SenderRecoveryStage {
                     commit_threshold: stage_config.sender_recovery.commit_threshold,
@@ -870,6 +871,7 @@ impl NodeConfig {
                 .set(MerkleStage::new_execution(stage_config.merkle.clean_threshold))
                 .set(TransactionLookupStage::new(
                     stage_config.transaction_lookup.chunk_size,
+                    stage_config.etl.etl_file_size,
                     prune_modes.transaction_lookup,
                 ))
                 .set(IndexAccountHistoryStage::new(

crates/stages/benches/criterion.rs

@@ -57,7 +57,7 @@ fn transaction_lookup(c: &mut Criterion) {
     let mut group = c.benchmark_group("Stages");
     // don't need to run each stage for that many times
     group.sample_size(10);
-    let stage = TransactionLookupStage::new(DEFAULT_NUM_BLOCKS, None);
+    let stage = TransactionLookupStage::new(DEFAULT_NUM_BLOCKS, 500 * 1024 * 1024, None);
 
     let db = setup::txs_testdata(DEFAULT_NUM_BLOCKS);
 

crates/stages/src/lib.rs

@@ -59,6 +59,7 @@
 //!             headers_downloader,
 //!             bodies_downloader,
 //!             executor_factory,
+//!             500*1024*1024,
 //!         )
 //!     )
 //!     .build(provider_factory, static_file_producer);

crates/stages/src/sets.rs

@@ -100,6 +100,7 @@ impl<Provider, H, B, EF> DefaultStages<Provider, H, B, EF> {
         header_downloader: H,
         body_downloader: B,
         executor_factory: EF,
+        etl_file_size: usize,
     ) -> Self
     where
         EF: ExecutorFactory,
@@ -111,6 +112,7 @@ impl<Provider, H, B, EF> DefaultStages<Provider, H, B, EF> {
                 consensus,
                 header_downloader,
                 body_downloader,
+                etl_file_size,
             ),
             executor_factory,
         }
@@ -162,6 +164,8 @@ pub struct OnlineStages<Provider, H, B> {
     header_downloader: H,
     /// The block body downloader
     body_downloader: B,
+    /// The size of temporary files in bytes for ETL data collector.
+    etl_file_size: usize,
 }
 
 impl<Provider, H, B> OnlineStages<Provider, H, B> {
@@ -172,8 +176,9 @@ impl<Provider, H, B> OnlineStages<Provider, H, B> {
         consensus: Arc<dyn Consensus>,
         header_downloader: H,
         body_downloader: B,
+        etl_file_size: usize,
     ) -> Self {
-        Self { provider, header_mode, consensus, header_downloader, body_downloader }
+        Self { provider, header_mode, consensus, header_downloader, body_downloader, etl_file_size }
     }
 }
 
@@ -198,9 +203,16 @@ where
         mode: HeaderSyncMode,
         header_downloader: H,
         consensus: Arc<dyn Consensus>,
+        etl_file_size: usize,
     ) -> StageSetBuilder<DB> {
         StageSetBuilder::default()
-            .add_stage(HeaderStage::new(provider, header_downloader, mode, consensus.clone()))
+            .add_stage(HeaderStage::new(
+                provider,
+                header_downloader,
+                mode,
+                consensus.clone(),
+                etl_file_size,
+            ))
             .add_stage(bodies)
     }
 }
@@ -219,6 +231,7 @@ where
                 self.header_downloader,
                 self.header_mode,
                 self.consensus.clone(),
+                self.etl_file_size,
             ))
             .add_stage(BodyStage::new(self.body_downloader))
     }

crates/stages/src/stages/headers.rs

@@ -74,15 +74,16 @@ where
         downloader: Downloader,
         mode: HeaderSyncMode,
         consensus: Arc<dyn Consensus>,
+        etl_file_size: usize,
     ) -> Self {
         Self {
             provider: database,
             downloader,
             mode,
             consensus,
             sync_gap: None,
-            hash_collector: Collector::new(100 * (1024 * 1024)),
-            header_collector: Collector::new(100 * (1024 * 1024)),
+            hash_collector: Collector::new(etl_file_size / 2),
+            header_collector: Collector::new(etl_file_size / 2),
             is_etl_ready: false,
         }
     }
@@ -419,6 +420,7 @@ mod tests {
                     (*self.downloader_factory)(),
                     HeaderSyncMode::Tip(self.channel.1.clone()),
                     self.consensus.clone(),
+                    500 * (1024 * 1024),
                 )
             }
         }

crates/stages/src/stages/tx_lookup.rs

@@ -32,19 +32,20 @@ pub struct TransactionLookupStage {
     /// The maximum number of lookup entries to hold in memory before pushing them to
     /// [`reth_etl::Collector`].
     chunk_size: u64,
+    etl_file_size: usize,
     prune_mode: Option<PruneMode>,
 }
 
 impl Default for TransactionLookupStage {
     fn default() -> Self {
-        Self { chunk_size: 5_000_000, prune_mode: None }
+        Self { chunk_size: 5_000_000, etl_file_size: 500 * 1024 * 1024, prune_mode: None }
     }
 }
 
 impl TransactionLookupStage {
     /// Create new instance of [TransactionLookupStage].
-    pub fn new(chunk_size: u64, prune_mode: Option<PruneMode>) -> Self {
-        Self { chunk_size, prune_mode }
+    pub fn new(chunk_size: u64, etl_file_size: usize, prune_mode: Option<PruneMode>) -> Self {
+        Self { chunk_size, etl_file_size, prune_mode }
     }
 }
 
@@ -99,7 +100,7 @@ impl<DB: Database> Stage<DB> for TransactionLookupStage {
         }
 
         // 500MB temporary files
-        let mut hash_collector: Collector<TxHash, TxNumber> = Collector::new(500 * (1024 * 1024));
+        let mut hash_collector: Collector<TxHash, TxNumber> = Collector::new(self.etl_file_size);
 
         debug!(
             target: "sync::stages::transaction_lookup",
@@ -397,12 +398,18 @@ mod tests {
     struct TransactionLookupTestRunner {
         db: TestStageDB,
         chunk_size: u64,
+        etl_file_size: usize,
         prune_mode: Option<PruneMode>,
     }
 
     impl Default for TransactionLookupTestRunner {
         fn default() -> Self {
-            Self { db: TestStageDB::default(), chunk_size: 1000, prune_mode: None }
+            Self {
+                db: TestStageDB::default(),
+                chunk_size: 1000,
+                etl_file_size: 500 * 1024 * 1024,
+                prune_mode: None,
+            }
         }
     }
 
@@ -449,7 +456,11 @@ mod tests {
         }
 
         fn stage(&self) -> Self::S {
-            TransactionLookupStage { chunk_size: self.chunk_size, prune_mode: self.prune_mode }
+            TransactionLookupStage {
+                chunk_size: self.chunk_size,
+                etl_file_size: self.etl_file_size,
+                prune_mode: self.prune_mode,
+            }
         }
     }