gnosis/lighthouse
1
0
Fork 0
mirror of https://github.com/sigp/lighthouse.git synced 2026-03-11 18:04:18 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add individual by_range sync requests (#6497)

Browse Source 
Part of
- https://github.com/sigp/lighthouse/issues/6258

To address PeerDAS sync issues we need to make individual by_range requests within a batch retriable. We should adopt the same pattern for lookup sync where each request (block/blobs/columns) is tracked individually within a "meta" request that group them all and handles retry logic.


  - Building on https://github.com/sigp/lighthouse/pull/6398

second step is to add individual request accumulators for `blocks_by_range`, `blobs_by_range`, and `data_columns_by_range`. This will allow each request to progress independently and be retried separately.

Most of the logic is just piping, excuse the large diff. This PR does not change the logic of how requests are handled or retried. This will be done in a future PR changing the logic of `RangeBlockComponentsRequest`.

### Before

- Sync manager receives block with `SyncRequestId::RangeBlockAndBlobs`
- Insert block into `SyncNetworkContext::range_block_components_requests`
- (If received stream terminators of all requests)
- Return `Vec<RpcBlock>`, and insert into `range_sync`

### Now

- Sync manager receives block with `SyncRequestId::RangeBlockAndBlobs`
- Insert block into `SyncNetworkContext:: blocks_by_range_requests`
- (If received stream terminator of this request)
- Return `Vec<SignedBlock>`, and insert into `SyncNetworkContext::components_by_range_requests `
- (If received a result for all requests)
- Return `Vec<RpcBlock>`, and insert into `range_sync`
This commit is contained in:
Lion - dapplion
2025-02-05 04:08:28 -03:00
committed by GitHub
parent 7bfdb33729
commit 2193f6a4d4
15 changed files with 776 additions and 502 deletions
Download Patch File Download Diff File Expand all files Collapse all files

203
beacon_node/network/src/sync/manager.rs
View File

@@ -36,7 +36,7 @@
use super::backfill_sync::{BackFillSync, ProcessResult, SyncStart};
use super::block_lookups::BlockLookups;
use super::network_context::{
BlockOrBlob, CustodyByRootResult, RangeRequestId, RpcEvent, SyncNetworkContext,
CustodyByRootResult, RangeBlockComponent, RangeRequestId, RpcEvent, SyncNetworkContext,
};
use super::peer_sampling::{Sampling, SamplingConfig, SamplingResult};
use super::peer_sync_info::{remote_sync_type, PeerSyncType};
@@ -47,7 +47,6 @@ use crate::status::ToStatusMessage;
use crate::sync::block_lookups::{
BlobRequestState, BlockComponent, BlockRequestState, CustodyRequestState, DownloadResult,
};
use crate::sync::block_sidecar_coupling::RangeBlockComponentsRequest;
use crate::sync::network_context::PeerGroup;
use beacon_chain::block_verification_types::AsBlock;
use beacon_chain::validator_monitor::timestamp_now;
@@ -57,8 +56,9 @@ use beacon_chain::{
use futures::StreamExt;
use lighthouse_network::rpc::RPCError;
use lighthouse_network::service::api_types::{
CustodyRequester, DataColumnsByRootRequestId, DataColumnsByRootRequester, Id, SamplingId,
SamplingRequester, SingleLookupReqId, SyncRequestId,
BlobsByRangeRequestId, BlocksByRangeRequestId, ComponentsByRangeRequestId, CustodyRequester,
DataColumnsByRangeRequestId, DataColumnsByRootRequestId, DataColumnsByRootRequester, Id,
SamplingId, SamplingRequester, SingleLookupReqId, SyncRequestId,
};
use lighthouse_network::types::{NetworkGlobals, SyncState};
use lighthouse_network::SyncInfo;
@@ -491,36 +491,14 @@ impl<T: BeaconChainTypes> SyncManager<T> {
SyncRequestId::DataColumnsByRoot(req_id) => {
self.on_data_columns_by_root_response(req_id, peer_id, RpcEvent::RPCError(error))
}
SyncRequestId::RangeBlockAndBlobs { id } => {
if let Some(sender_id) = self.network.range_request_failed(id) {
match sender_id {
RangeRequestId::RangeSync { chain_id, batch_id } => {
self.range_sync.inject_error(
&mut self.network,
peer_id,
batch_id,
chain_id,
id,
);
self.update_sync_state();
}
RangeRequestId::BackfillSync { batch_id } => match self
.backfill_sync
.inject_error(&mut self.network, batch_id, &peer_id, id)
{
Ok(_) => {}
Err(_) => self.update_sync_state(),
},
}
} else {
debug!(
self.log,
"RPC error for range request has no associated entry in network context, ungraceful disconnect";
"peer_id" => %peer_id,
"request_id" => %id,
"error" => ?error,
);
}
SyncRequestId::BlocksByRange(req_id) => {
self.on_blocks_by_range_response(req_id, peer_id, RpcEvent::RPCError(error))
}
SyncRequestId::BlobsByRange(req_id) => {
self.on_blobs_by_range_response(req_id, peer_id, RpcEvent::RPCError(error))
}
SyncRequestId::DataColumnsByRange(req_id) => {
self.on_data_columns_by_range_response(req_id, peer_id, RpcEvent::RPCError(error))
}
}
}
@@ -1051,14 +1029,13 @@ impl<T: BeaconChainTypes> SyncManager<T> {
SyncRequestId::SingleBlock { id } => self.on_single_block_response(
id,
peer_id,
match block {
Some(block) => RpcEvent::Response(block, seen_timestamp),
None => RpcEvent::StreamTermination,
},
RpcEvent::from_chunk(block, seen_timestamp),
),
SyncRequestId::BlocksByRange(id) => self.on_blocks_by_range_response(
id,
peer_id,
RpcEvent::from_chunk(block, seen_timestamp),
),
SyncRequestId::RangeBlockAndBlobs { id } => {
self.range_block_and_blobs_response(id, peer_id, block.into())
}
_ => {
crit!(self.log, "bad request id for block"; "peer_id" => %peer_id );
}
@@ -1094,14 +1071,13 @@ impl<T: BeaconChainTypes> SyncManager<T> {
SyncRequestId::SingleBlob { id } => self.on_single_blob_response(
id,
peer_id,
match blob {
Some(blob) => RpcEvent::Response(blob, seen_timestamp),
None => RpcEvent::StreamTermination,
},
RpcEvent::from_chunk(blob, seen_timestamp),
),
SyncRequestId::BlobsByRange(id) => self.on_blobs_by_range_response(
id,
peer_id,
RpcEvent::from_chunk(blob, seen_timestamp),
),
SyncRequestId::RangeBlockAndBlobs { id } => {
self.range_block_and_blobs_response(id, peer_id, blob.into())
}
_ => {
crit!(self.log, "bad request id for blob"; "peer_id" => %peer_id);
}
@@ -1120,19 +1096,14 @@ impl<T: BeaconChainTypes> SyncManager<T> {
self.on_data_columns_by_root_response(
req_id,
peer_id,
match data_column {
Some(data_column) => RpcEvent::Response(data_column, seen_timestamp),
None => RpcEvent::StreamTermination,
},
);
}
SyncRequestId::RangeBlockAndBlobs { id } => {
self.range_block_and_blobs_response(
id,
peer_id,
BlockOrBlob::CustodyColumns(data_column),
RpcEvent::from_chunk(data_column, seen_timestamp),
);
}
SyncRequestId::DataColumnsByRange(id) => self.on_data_columns_by_range_response(
id,
peer_id,
RpcEvent::from_chunk(data_column, seen_timestamp),
),
_ => {
crit!(self.log, "bad request id for data_column"; "peer_id" => %peer_id);
}
@@ -1188,6 +1159,54 @@ impl<T: BeaconChainTypes> SyncManager<T> {
}
}
fn on_blocks_by_range_response(
&mut self,
id: BlocksByRangeRequestId,
peer_id: PeerId,
block: RpcEvent<Arc<SignedBeaconBlock<T::EthSpec>>>,
) {
if let Some(resp) = self.network.on_blocks_by_range_response(id, peer_id, block) {
self.on_range_components_response(
id.parent_request_id,
peer_id,
RangeBlockComponent::Block(resp),
);
}
}
fn on_blobs_by_range_response(
&mut self,
id: BlobsByRangeRequestId,
peer_id: PeerId,
blob: RpcEvent<Arc<BlobSidecar<T::EthSpec>>>,
) {
if let Some(resp) = self.network.on_blobs_by_range_response(id, peer_id, blob) {
self.on_range_components_response(
id.parent_request_id,
peer_id,
RangeBlockComponent::Blob(resp),
);
}
}
fn on_data_columns_by_range_response(
&mut self,
id: DataColumnsByRangeRequestId,
peer_id: PeerId,
data_column: RpcEvent<Arc<DataColumnSidecar<T::EthSpec>>>,
) {
if let Some(resp) = self
.network
.on_data_columns_by_range_response(id, peer_id, data_column)
{
self.on_range_components_response(
id.parent_request_id,
peer_id,
RangeBlockComponent::CustodyColumns(resp),
);
}
}
fn on_custody_by_root_result(
&mut self,
requester: CustodyRequester,
@@ -1230,27 +1249,26 @@ impl<T: BeaconChainTypes> SyncManager<T> {
/// Handles receiving a response for a range sync request that should have both blocks and
/// blobs.
fn range_block_and_blobs_response(
fn on_range_components_response(
&mut self,
id: Id,
range_request_id: ComponentsByRangeRequestId,
peer_id: PeerId,
block_or_blob: BlockOrBlob<T::EthSpec>,
range_block_component: RangeBlockComponent<T::EthSpec>,
) {
if let Some(resp) = self
.network
.range_block_and_blob_response(id, block_or_blob)
.range_block_component_response(range_request_id, range_block_component)
{
let epoch = resp.sender_id.batch_id();
match resp.responses {
match resp {
Ok(blocks) => {
match resp.sender_id {
match range_request_id.requester {
RangeRequestId::RangeSync { chain_id, batch_id } => {
self.range_sync.blocks_by_range_response(
&mut self.network,
peer_id,
chain_id,
batch_id,
id,
range_request_id.id,
blocks,
);
self.update_sync_state();
@@ -1260,7 +1278,7 @@ impl<T: BeaconChainTypes> SyncManager<T> {
&mut self.network,
batch_id,
&peer_id,
id,
range_request_id.id,
blocks,
) {
Ok(ProcessResult::SyncCompleted) => self.update_sync_state(),
@@ -1274,36 +1292,25 @@ impl<T: BeaconChainTypes> SyncManager<T> {
}
}
}
Err(e) => {
// Re-insert the request so we can retry
self.network.insert_range_blocks_and_blobs_request(
id,
resp.sender_id,
RangeBlockComponentsRequest::new(
resp.expects_blobs,
resp.expects_custody_columns,
None,
vec![],
self.chain.spec.max_blobs_per_block(epoch) as usize,
),
);
// inform range that the request needs to be treated as failed
// With time we will want to downgrade this log
warn!(
self.log,
"Blocks and blobs request for range received invalid data";
"peer_id" => %peer_id,
"sender_id" => ?resp.sender_id,
"error" => e.clone()
);
let id = SyncRequestId::RangeBlockAndBlobs { id };
self.network.report_peer(
peer_id,
PeerAction::MidToleranceError,
"block_blob_faulty_batch",
);
self.inject_error(peer_id, id, RPCError::InvalidData(e))
}
Err(_) => match range_request_id.requester {
RangeRequestId::RangeSync { chain_id, batch_id } => {
self.range_sync.inject_error(
&mut self.network,
peer_id,
batch_id,
chain_id,
range_request_id.id,
);
self.update_sync_state();
}
RangeRequestId::BackfillSync { batch_id } => match self
.backfill_sync
.inject_error(&mut self.network, batch_id, &peer_id, range_request_id.id)
{
Ok(_) => {}
Err(_) => self.update_sync_state(),
},
},
}
}
}