feat: implement resume-resharding tool #12796
+211
−2
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -1 +1,2 @@ | ||
| pub mod commands; | ||
| mod resume_resharding; |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,174 @@ | ||
| use crate::commands::ResumeReshardingCmd; | ||
| use anyhow::anyhow; | ||
| use near_async::messaging::{CanSend, IntoMultiSender}; | ||
| use near_async::time::Clock; | ||
| use near_chain::flat_storage_resharder::FlatStorageReshardingTaskResult; | ||
| use near_chain::rayon_spawner::RayonAsyncComputationSpawner; | ||
| use near_chain::resharding::types::{ | ||
| FlatStorageShardCatchupRequest, FlatStorageSplitShardRequest, MemtrieReloadRequest, | ||
| }; | ||
| use near_chain::types::ChainConfig; | ||
| use near_chain::{Chain, ChainGenesis, ChainStore, DoomslugThresholdMode}; | ||
| use near_chain_configs::MutableConfigValue; | ||
| use near_epoch_manager::shard_tracker::{ShardTracker, TrackedConfig}; | ||
| use near_epoch_manager::EpochManager; | ||
| use near_store::adapter::StoreAdapter; | ||
| use near_store::flat::FlatStorageReshardingStatus; | ||
| use near_store::{ShardUId, StoreOpener}; | ||
| use nearcore::{NearConfig, NightshadeRuntime, NightshadeRuntimeExt}; | ||
| use std::collections::VecDeque; | ||
| use std::path::{Path, PathBuf}; | ||
| use std::sync::{Arc, Mutex}; | ||
| use tracing::info; | ||
|
|
||
| pub(crate) fn resume_resharding( | ||
| cmd: &ResumeReshardingCmd, | ||
| home_dir: &PathBuf, | ||
| config: &NearConfig, | ||
| opener: StoreOpener, | ||
| ) -> anyhow::Result<()> { | ||
| let (chain, executor) = create_chain_and_executor(config, home_dir, opener)?; | ||
| let resharder = chain.resharding_manager.flat_storage_resharder.clone(); | ||
|
|
||
| let shard_uid = ShardUId::new(3, cmd.shard_id); // version is fixed at 3 in resharding V3 | ||
| let resharding_status = get_resharding_status_for_shard(&shard_uid, &chain)?; | ||
|
|
||
| chain.runtime_adapter.get_flat_storage_manager().create_flat_storage_for_shard(shard_uid)?; | ||
|
|
||
| resharder.resume(shard_uid, &resharding_status)?; | ||
|
|
||
| while executor.run()? { | ||
| info!(target: "resharding", "running next task"); | ||
| } | ||
|
|
||
| info!(target: "resharding", "operation completed"); | ||
| Ok(()) | ||
| } | ||
|
|
||
| fn create_chain_and_executor( | ||
| config: &NearConfig, | ||
| home_dir: &Path, | ||
| opener: StoreOpener, | ||
| ) -> anyhow::Result<(Chain, Arc<SerialExecutor>)> { | ||
| let node_storage = opener.open_in_mode(near_store::Mode::ReadWriteExisting).unwrap(); | ||
| let epoch_manager = EpochManager::new_arc_handle( | ||
| node_storage.get_hot_store(), | ||
| &config.genesis.config, | ||
| Some(home_dir), | ||
| ); | ||
| let runtime_adapter = NightshadeRuntime::from_config( | ||
| home_dir, | ||
| node_storage.get_hot_store(), | ||
| &config, | ||
| epoch_manager.clone(), | ||
| )?; | ||
| let shard_tracker = | ||
| ShardTracker::new(TrackedConfig::from_config(&config.client_config), epoch_manager.clone()); | ||
| let chain_genesis = ChainGenesis::new(&config.genesis.config); | ||
| let chain_config = ChainConfig { | ||
| save_trie_changes: config.client_config.save_trie_changes, | ||
| background_migration_threads: config.client_config.client_background_migration_threads, | ||
| resharding_config: config.client_config.resharding_config.clone(), | ||
| }; | ||
| let executor = Arc::new(SerialExecutor::new(ChainStore::new( | ||
| node_storage.get_hot_store(), | ||
| chain_genesis.height, | ||
| false, | ||
| chain_genesis.transaction_validity_period, | ||
| ))); | ||
| let chain = Chain::new( | ||
| Clock::real(), | ||
| epoch_manager, | ||
| shard_tracker, | ||
| runtime_adapter, | ||
| &chain_genesis, | ||
| DoomslugThresholdMode::TwoThirds, | ||
| chain_config, | ||
| None, | ||
| Arc::new(RayonAsyncComputationSpawner), | ||
| MutableConfigValue::new(None, "validator_signer"), | ||
| executor.as_multi_sender(), | ||
| )?; | ||
| Ok((chain, executor)) | ||
| } | ||
|
|
||
| fn get_resharding_status_for_shard( | ||
| shard_uid: &ShardUId, | ||
| chain: &Chain, | ||
| ) -> anyhow::Result<FlatStorageReshardingStatus> { | ||
| use near_store::flat::FlatStorageStatus::*; | ||
| let flat_storage_status = | ||
| chain.runtime_adapter.store().flat_store().get_flat_storage_status(*shard_uid)?; | ||
| match &flat_storage_status { | ||
| Disabled | Empty | Creation(_) | Ready(_) => Err(anyhow!( | ||
| "resharding is not in progress! flat storage status: {:?}", | ||
| flat_storage_status | ||
| )), | ||
| Resharding(status) => Ok(status.clone()), | ||
| } | ||
| } | ||
|
|
||
| /// Executor that runs tasks in sequence. | ||
| struct SerialExecutor { | ||
| chain_store: Arc<Mutex<ChainStore>>, | ||
| tasks: Mutex<VecDeque<Box<dyn Fn(&ChainStore) -> FlatStorageReshardingTaskResult + Send>>>, | ||
| } | ||
|
|
||
| impl SerialExecutor { | ||
| fn new(chain_store: ChainStore) -> Self { | ||
| Self { chain_store: Arc::new(Mutex::new(chain_store)), tasks: Mutex::new(VecDeque::new()) } | ||
| } | ||
|
|
||
| /// Runs the next task. | ||
| /// | ||
| /// Returns false if there are no tasks. | ||
| fn run(&self) -> anyhow::Result<bool> { | ||
| let task = self.tasks.lock().unwrap().pop_front(); | ||
| match task { | ||
| Some(task) => { | ||
| match task(&self.chain_store.lock().unwrap()) { | ||
| FlatStorageReshardingTaskResult::Successful { num_batches_done } => { | ||
| info!(target: "resharding", num_batches_done, "task completed"); | ||
| } | ||
| FlatStorageReshardingTaskResult::Failed => { | ||
| return Err(anyhow!("resharding task has failed!")); | ||
| } | ||
| FlatStorageReshardingTaskResult::Cancelled => { | ||
| info!(target: "resharding", "task cancelled") | ||
| } | ||
| FlatStorageReshardingTaskResult::Postponed => { | ||
| info!(target: "resharding", "task postponed - retrying"); | ||
| self.tasks.lock().unwrap().push_back(task); | ||
| } | ||
| } | ||
| Ok(true) | ||
| } | ||
| None => Ok(false), | ||
| } | ||
| } | ||
| } | ||
|
|
||
| impl CanSend<FlatStorageSplitShardRequest> for SerialExecutor { | ||
| fn send(&self, msg: FlatStorageSplitShardRequest) { | ||
| let resharder = msg.resharder.clone(); | ||
| let task = | ||
| Box::new(move |chain_store: &ChainStore| resharder.split_shard_task(chain_store)); | ||
| self.tasks.lock().unwrap().push_back(task); | ||
| } | ||
| } | ||
|
|
||
| impl CanSend<FlatStorageShardCatchupRequest> for SerialExecutor { | ||
| fn send(&self, msg: FlatStorageShardCatchupRequest) { | ||
| let resharder = msg.resharder.clone(); | ||
| let task = Box::new(move |chain_store: &ChainStore| { | ||
| resharder.shard_catchup_task(msg.shard_uid, chain_store) | ||
| }); | ||
| self.tasks.lock().unwrap().push_back(task); | ||
| } | ||
| } | ||
|
|
||
| impl CanSend<MemtrieReloadRequest> for SerialExecutor { | ||
| fn send(&self, _: MemtrieReloadRequest) { | ||
| // no op | ||
| } | ||
| } |
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I had to freeze the parent flat storage in the end, because otherwise something in chain was overriding its status.