If turbine_disabled is true, the commit discards turbine packets
earlier in the pipeline so that they won't interfere with the deduper
and the packets can get through once turbine is enabled again.
This is a prerequisite of:
https://github.com/solana-labs/solana/pull/30786
so that local-cluster tests pass.
For duplicate block detection, for each (slot, shred-index, shred-type)
we need to allow 2 different shreds to be retransmitted.
The commit implements this using two bloom-filter dedupers:
* Shreds are deduplicated using the 1st deduper.
* If a shred is not a duplicate, then we check if:
(slot, shred-index, shred-type, k)
is not a duplicate for either k = 0 or k = 1 using the 2nd deduper,
and if so then the shred is retransmitted.
This allows to achieve larger capactiy compared to current LRU-cache.
generalizes Deduper to work with any hashable type
Current Deduper is hard-coded only for Packet type. In order to use
Deduper in retransmit-stage, we need to dedup types other than Packet.
The commit generalizes Deduper to any hashable type.
* Add new vote state version that replaces Lockout with LandedVote to allow vote latency to be tracked in a future change. Includes a feature to be enabled which will when enabled cause the vote state to be written in the new form.
* Update feature set key to one owned by ashwin
---------
Co-authored-by: Ashwin Sekar <ashwin@solana.com>
removes the false_positive_rate field from the Deduper
Deduper.false_positive_rate field is misleading because it is not
enforced until maybe_reset is called. But then maybe_reset can be
invoked with an explicit argument.
* Clean up orphaned account snapshot hardlink dirs
* fix compilation issues
* debugged, now working. seeing the orphaned directories deleted
* change back to eprintln + exit for account_path error
* changed eprintln to panic for now
* add test_clean_orphaned_account_snapshot_dirs for codecov check
* address a few comments and nit isseus
* directly unzip, skipped the intermediate array of tuples
* let set_up_account_run_and_snapshot_paths return Result
* 'proper' typo, and comment on return
* use map_err
* use for loop in clean_orphaned_account_snapshot_dirs, removed panic
* add test_set_up_account_run_and_snapshot_paths
* minor, replace .for_each with .all
* rename set_up_account_run_and_snapshot_paths to create_all_accounts_run_and_snapshot_dirs
* remove unnecessary closure return type
* change to for loop
* change match to unwrap_or_else
* remove create_dir_all(&account_path) in create_all
* minor comment cleanup
Current Deduper implementation uses many bits per entry:
https://github.com/solana-labs/solana/blob/65cd55261/perf/src/deduper.rs#L70-L73
and may be saturated quickly. It also lacks api to specify desired false
positive rate.
The commit instead uses an atomic bloom filter with K hash functions.
The false positive rate is obtained by tracking popcount of bits.
* Add RewardsMessage enum
* Cache and update max_complete_rewards_slot
* Plumb max_complete_rewards_slot into JsonRpcRequestProcesseor
* Use max_complete_rewards_slot to check get_block requests
* Use max_complete_rewards_slot to limit Bigtable uploads
* Plumb max_complete_rewards_slot into RpcSubscriptions
* Use max_complete_rewards_slot to limit block subscriptions
* Nit: fix test
* Add helper function to create BlockstoreOptions from ValidatorConfig
* Make backup_and_clear_blockstore() honor ValidatorConfig options
backup_and_clear_blockstore() opens a Blockstore session; however, it
is currently using Blockstore::open(). This Blockstore method uses
BlockstoreOption::default() under the hood. As a result, any validator
args that adjust Blockstore settings are not considered in
backup_and_clear_blockstore().
This is especially problematic if the non-default value of
--rocksdb-shred-compaction is being used. In this case,
backup_and_clear_blockstore() was opening the wrong directory and
incorrectly finding an empty ledger.
This change plumbs any blockstore configuration to
backup_and_clear_blockstore().
* Move AdminRpcRequestMetadataPostInit to solana-core
* Move AdminRpcRequestMetadataPostInit write to just before wait_for_supermajority
* Pass AdminRpcRequestMetadataPostInit in TestValidatorGenesis
* Fixup local-cluster
* introduce workspace.package
* introduce workspace.dependencies
* read version from root cargo.toml
* pass check when version = { workspace = true }
* don't bump version when version = { workspace = true }
* including workspace Cargo.toml when bump version
* programs/sbf use workspace inheritance
* fix increasing cargo version ignore program/sbf/Cargo.toml
* measure! to measure_us!
* Consistent naming of transaction_recorder
* Remove outdated comment - Instant cannot be None
* use local
* Remove measure! import
* Add accounts hard-link files into the bank snapshot directory
* Small adjustments and fixes.
* Address some of the review issues
* Fix compilation issues
* Change the latest slot snapshot storage from VecDeque to Option
* IoWithSourceAndFile and expanded comments on accounts
* last_slot_snapshot_storages in return value
* Update comments following the review input
* rename dir_accounts_hard_links to hard_link_path
* Add dir_full_state flag for add_bank_snapshot
* Let appendvec files hardlinking work with multiple accounts paths across multiple partitions
* Fixes for rebasing
* fix tests which generates account_path without adding run/
* rebasing fixes
* fix account path test failures
* fix test test_concurrent_snapshot_packaging
* review comments. renamed the path setup function
* Addressed most of the review comments
* update with more review comments
* handle error from create_accounts_run_and_snapshot_dirs
* fix rebasing duplicate
* minor accounts_dir path cleanup
* minor cleanup, remove commented code
* misc review comments
* build error fix
* Fix test_incremental_snapshot_download_with_crossing_full_snapshot_interval_at_startup
* fix build error on MAX_BANK_SNAPSHOTS_TO_RETAIN
* rebase fix, update hardlink filename
* minor comment spelling fix
* rebasing fixes
* fix rebase issues; with_extension
* comments changes for review
* misc minor review issues
* bank.fill_bank_with_ticks_for_tests
* error handling on appendvec path
* fix use_jit
* minor comments refining
* Remove type AccountStorages
* get_account_path_from_appendvec_path return changed to Option
* removed appendvec_path.to_path_buf in create_accounts_run_and_snapshot_dirs
* add test_get_snapshot_accounts_hardlink_dir
* update last_snapshot_storages comment
* update last_snapshot_storages comment
* symlink map_err
* simplify test_get_snapshot_accounts_hardlink_dir with fake paths
* log last_snapshot_storages at the end of the loop
Snapshot construction and interaction code was a bit more manual than necessary, even causing a bug to slip past a review. Separated snapshot construction from the diffing of two snapshots.
This should make the logic clearer.
A recent change to this service to store the number of non-vote
transactions introduced a bug in the computation of the number of
transactions during the time interval. This resulted in bogus values
being stored in Blockstore and eventually getting served through RPC for
the TPS chart on explorer.
Dynamic dispatch forces heap allocation and adds extra overhead.
Dynamic casting as in the ones below, lacks compile-time type safety:
https://github.com/solana-labs/solana/blob/eeb622c4e/quic-client/src/lib.rs#L172-L175https://github.com/solana-labs/solana/blob/eeb622c4e/udp-client/src/lib.rs#L52-L55
The commit removes all instances of Any, Box<dyn ...>, and Arc<dyn ...>,
and instead uses generic and associated types.
There are only two protocols QUIC and UDP; and the code which has to
work with both protocols can use a trivial thin enum wrapper.
With respect to connection-cache specifically:
* connection-cache/ConnectionCache is a single protocol cache which
allows to use either QUIC or UDP without any build dependency on the
other protocol.
* client/ConnectionCache is an enum wrapper around both protocols and
can be used in the code which has to work with both QUIC and UDP.
Co-authored-by: Tyera Eulberg <tyera@solana.com>
tpu-client/tpu_connection_cache is refactored out the module and moved to connection-cache/connection_cache and the logic in client/connection_cache is consolidated to connection-cache/connection_cache as well. client/connection_cache only has a thin wrapper which forward calls to connection-cache/connection_cache and deal with constructions of quic/udp connection cache for clients using them both.2.
The TpuConnection is refactored to ClientConnection to make it generic and functions renamed to be proper for other workflows. eg. tpu_addr -> server_addr, send_transaction --> send_data and etc...
The enum dispatch is removed so that we can make the bulk of code of quic and udp agnostic of each other. The client is possible to load quic or udp only into its runtime.
The generic type parameter in the tpu-client/tpu_connection_cache is removed in order to create both quic and udp connection cache and use the object to send transactions with multiple branching when sending data. The generic type parameters and associated types are dropped in other types in order to make the trait "object safe" for this purpose.
I have annotated the code explaining the reasoning and the refactoring source -> destination.
There is no functional changes
bench-tps has been performed for rpc-client, thin-client and tpu-client. And it is found the performance number largely match the ones before the refactoring.