Given the 32:32 erasure recovery schema, current implementation requires
exactly 32 data shreds to generate coding shreds for the batch (except
for the final erasure batch in each slot).
As a result, when serializing ledger entries to data shreds, if the
number of data shreds is not a multiple of 32, the coding shreds for the
last batch cannot be generated until there are more data shreds to
complete the batch to 32 data shreds. This adds latency in generating
and broadcasting coding shreds.
In addition, with Merkle variants for shreds, data shreds cannot be
signed and broadcasted until coding shreds are also generated. As a
result *both* code and data shreds will be delayed before broadcast if
we still require exactly 32 data shreds for each batch.
This commit instead always generates and broadcast coding shreds as soon
as there any number of data shreds available. When serializing entries
to shreds:
* if the number of resulting data shreds is less than 32, then more
coding shreds will be generated so that the resulting erasure batch
has the same recovery probabilities as a 32:32 batch.
* if the number of data shreds is more than 32, then the data shreds are
split uniformly into erasure batches with _at least_ 32 data shreds in
each batch. Each erasure batch will have the same number of code and
data shreds.
For example:
* If there are 19 data shreds, 27 coding shreds are generated. The
resulting 19(data):27(code) erasure batch has the same recovery
probabilities as a 32:32 batch.
* If there are 107 data shreds, they are split into 3 batches of 36:36,
36:36 and 35:35 data:code shreds each.
A consequence of this change is that code and data shreds indices will
no longer align as there will be more coding shreds than data shreds
(not only in the last batch in each slot but also in the intermediate
ones);
- Forward packets by prioritization in desc order
- Add support of cost-tracking by transaction requested compute units
- Hook up account buckets to forwarder
- Add metrics for forwardable batches count
- Remove redundant invalid packets filtering at end of slot since forwarder will do the same when batch forwardable packets
- Add bench test for forwarding
Fully deserializing shreds in window-service before sending them to
retransmit stage adds latency to shreds propagation.
This commit instead channels through the payload and relies on only
partial deserialization of a few required fields: slot, shred-index,
shred-type.
Following commits will skip shreds deserializaton before retransmit, and
so we will only have a ShredId and not a fully deserialized shred to
obtain the shuffling seed from.
* Define shuffle to prep using same shuffle for multiple slices
* Determine transaction indexes and plumb to execute_batch
* Pair transaction_index with transaction in TransactionStatusService
* Add new ReplicaTransactionInfoVersion
* Plumb transaction_indexes through BankingStage
* Prepare BankingStage to receive transaction indexes from PohRecorder
* Determine transaction indexes in PohRecorder; add field to WorkingBank
* Add PohRecorder::record unit test
* Only pass starting_transaction_index around PohRecorder
* Add helper structs to simplify test DashMap
* Pass entry and starting-index into process_entries_with_callback together
* Add tx-index checks to test_rebatch_transactions
* Revert shuffle definition and use zip/unzip
* Only zip/unzip if randomize
* Add confirm_slot_entries test
* Review nits
* Add type alias to make sender docs more clear
In prepration of
https://github.com/solana-labs/solana/pull/25807
which reworks erasure batch sizes, this commit:
* adds a helper function mapping the number of data shreds to the
erasure batch size.
* adds ProcessShredsStats to Shredder::entries_to_shreds in order to
replace and remove entries_to_data_shreds from the public interface.
* client: Remove static connection cache, plumb it instead
* Add TpuClient::new_with_connection_cache to not break downstream
* Refactor get_connection and RwLock into ConnectionCache
* Fix merge conflicts from new async TpuClient
* Remove `ConnectionCache::set_use_quic`
* Move DEFAULT_TPU_USE_QUIC to client, use ConnectionCache::default()
Upcoming changes to PacketBatch to support variable sized packets will
modify the internals of PacketBatch. So, this change removes usage of
the internal packet struct and instead uses accessors (which are
currently just wrappers of Vector functions but will change down the
road).
* - get prioritization fee from compute_budget instruction;
- update compute_budget::process_instruction function to take instruction iter to support sanitized versioned message;
- updated runtime.md
* update transaction fee calculation for prioritization fee rate as lamports per 10K CUs
* review changes
* fix test
* fix a bpf test
* fix bpf test
* patch feedback
* fix clippy
* fix bpf test
* feedback
* rename prioritization fee rate to compute unit price
* feedback
Co-authored-by: Justin Starry <justin@solana.com>
Shred::new_empty_data_shred returns an invalid shred (i.e.
shred.sanitize() returns error). The method is only used in tests and
can be easily replaced with Shred::new_from_data. To keep the shred api
surface small, this commit removes this method.
As shown by the added benchmark, current code does worse if there is a
spam address plus a lot of unique addresses.
on current master:
test bench_packet_discard_many_senders ... bench: 1,997,960 ns/iter (+/- 103,715)
test bench_packet_discard_mixed_senders ... bench: 14,256,116 ns/iter (+/- 534,865)
test bench_packet_discard_single_sender ... bench: 1,306,809 ns/iter (+/- 61,992)
with this commit:
test bench_packet_discard_many_senders ... bench: 1,644,025 ns/iter (+/- 83,715)
test bench_packet_discard_mixed_senders ... bench: 1,089,789 ns/iter (+/- 86,324)
test bench_packet_discard_single_sender ... bench: 955,234 ns/iter (+/- 55,953)
* - qos_service metrics tagged with leader thread ids to separate gossip/tpu votes and transactions;
- qos_service metrics is reported with bank slot;
- replaced timer-based reporting with signal via channel; removed async report test as qos_service now lives within a thread
* - add tpu live packets (eg, not buffered packets) states to qos metrics reporting
The indices for erasure coding shreds are tied to data shreds:
https://github.com/solana-labs/solana/blob/90f41fd9b/ledger/src/shred.rs#L921
However with the upcoming changes to erasure schema, there will be more
erasure coding shreds than data shreds and we can no longer infer coding
shreds indices from data shreds.
The commit adds constructs to track coding shreds indices explicitly.
next-shred-index is already readily available from returned data shreds.
The commit simplifies the api for upcoming changes to erasure coding
schema which will require explicit tracking of indices for coding shreds
as well as data shreds.
- decouple cost_model from cost_tracker; allowing one cost_model
instance being shared within a validator;
- update cost_model api to calculate_cost(&self...)->transaction_cost
* Add separate vote processing tpu port
* Add feature to send to tpu vote port
* Add vote rejecting sigverify mode
* use packet.meta.is_simple_vote_tx in place of deserialization
* consolidate code that identifies vote tx atcommon path for cpu and gpu
* new key for feature set
* banking forward tpu vote
* add tpu vote port to dockerfile and other review changes
* Simplify thread id compare
* fix a test; updated cluster_info ABI change
Co-authored-by: Tao Zhu <tao@solana.com>
Co-authored-by: sakridge <sakridge@gmail.com>
* windows: Make solana-test-validator work
The important changes to get this going on Windows:
* ledger lock needs to be done on a file instead of the directory
* IPC service needs to use the Windows pipe naming scheme
* always disable the JIT
* file logging not possible yet because we can't redirect stderr,
but this will change once env_logger fixes the pipe output target!
* Integrate review feedback
* Move transaction sanitization earlier in the pipeline
* Renamed HashedTransaction to SanitizedTransaction
* Implement deref for sanitized transaction
* bring back process_transactions test method
* Use sanitized transactions for cost model calculation
Broadcast stage and retransmit stage should arrange nodes on turbine
broadcast tree in exactly same order. Additionally any changes to this
ordering (e.g. updating how unstaked nodes are handled) requires feature
gating to keep the cluster in sync.
Current implementation is scattered out over several public methods and
exposes too much of implementation details (e.g. usize indices into
peers vector) which makes code changes and checking for feature
activations more difficult.
This commit encapsulates turbine peer computations into a new struct,
and only exposes two public methods, get_broadcast_peer and
get_retransmit_peers, for call-sites.
* Add stats and counter around cost model ops, mainly:
- calculate transaction cost
- check transaction can fit in a block
- update block cost tracker after transactions are added to block
- replay_stage to update/insert execution cost to table
* Change mutex on cost_tracker to RwLock
* removed cloning cost_tracker for local use, as the metrics show clone is very expensive.
* acquire and hold locks for block of TXs, instead of acquire and release per transaction;
* remove redundant would_fit check from cost_tracker update execution path
* refactor cost checking with less frequent lock acquiring
* avoid many Transaction_cost heap allocation when calculate cost, which
is in the hot path - executed per transaction.
* create hashmap with new_capacity to reduce runtime heap realloc.
* code review changes: categorize stats, replace explicit drop calls, concisely initiate to default
* address potential deadlock by acquiring locks one at time
* replay stage feeds back realtime per-program execution cost to cost model;
* program cost execution table is initialized into empty table, no longer populated with hardcoded numbers;
* changed cost unit to microsecond, using value collected from mainnet;
* add ExecuteCostTable with fixed capacity for security concern, when its limit is reached, programs with old age AND less occurrence will be pushed out to make room for new programs.
* Create solana-poh crate
* Move BigTableUploadService to solana-ledger
* Add solana-rpc to workspace
* Move dependencies to solana-rpc
* Move remaining rpc modules to solana-rpc
* Single use statement solana-poh
* Single use statement solana-rpc
* * Add following to banking_stage:
1. CostModel as immutable ref shared between threads, to provide estimated cost for transactions.
2. CostTracker which is shared between threads, tracks transaction costs for each block.
* replace hard coded program ID with id() calls
* Add Account Access Cost as part of TransactionCost. Account Access cost are weighted differently between read and write, signed and non-signed.
* Establish instruction_execution_cost_table, add function to update or insert instruction cost, unit tested. It is read-only for now; it allows Replay to insert realtime instruction execution costs to the table.
* add test for cost_tracker atomically try_add operation, serves as safety guard for future changes
* check cost against local copy of cost_tracker, return transactions that would exceed limit as unprocessed transaction to be buffered; only apply bank processed transactions cost to tracker;
* bencher to new banking_stage with max cost limit to allow cost model being hit consistently during bench iterations
* Move gossip modules to solana-gossip
* Update Protocol abi digest due to move
* Move gossip benches and hook up CI
* Remove unneeded Result entries
* Single use statements
For all code paths (gossip push, pull, purge, etc) that remove or
override a crds value, it is necessary to record hash of values purged
from crds table, in order to exclude them from subsequent pull-requests;
otherwise the next pull request will likely return outdated values,
wasting bandwidth:
https://github.com/solana-labs/solana/blob/ed51cde37/core/src/crds_gossip_pull.rs#L486-L491
Currently this is done all over the place in multiple modules, and this
has caused bugs in the past where purged values were not recorded.
This commit encapsulated this bookkeeping into crds module, so that any
code path which removes or overrides a crds value, also records the hash
of purged value in-place.
It is crucial that VersionedCrdsValue::insert_timestamp does not go
backward in time:
https://github.com/solana-labs/solana/blob/ec37a843a/core/src/crds.rs#L67-L79
Otherwise methods such as get_votes and get_epoch_slots_since will
break, which will break their downstream flow, including vote-listener
and optimistic confirmation:
https://github.com/solana-labs/solana/blob/ec37a843a/core/src/cluster_info.rs#L1197-L1215https://github.com/solana-labs/solana/blob/ec37a843a/core/src/cluster_info.rs#L1274-L1298
For that, Crds::new_versioned is intended to be called "atomically" with
Crds::insert_verioned (as the comment already says so):
https://github.com/solana-labs/solana/blob/ec37a843a/core/src/crds.rs#L126-L129
However, currently this is violated in the code. For example,
filter_pull_responses creates VersionedCrdsValues (with the current
timestamp), then acquires an exclusive lock on gossip, then
process_pull_responses writes those values to the crds table:
https://github.com/solana-labs/solana/blob/ec37a843a/core/src/cluster_info.rs#L2375-L2392
Depending on the workload and lock contention, the insert_timestamps may
well be in the past when these values finally are inserted into gossip.
To avoid such scenarios, this commit:
* removes Crds::new_versioned and Crd::insert_versioned.
* makes VersionedCrdsValue constructor private, only invoked in
Crds::insert, so that insert_timestamp is populated right before
insert.
This will improve insert_timestamp monotonicity as long as Crds::insert
is not called with a stalled timestamp. Following commits may further
improve this by calling timestamp() inside Crds::insert, and/or
switching to std::time::Instant which guarantees monotonicity.
Number of parity coding shreds is always less than the number of data
shreds in FEC blocks:
https://github.com/solana-labs/solana/blob/6907a2366/ledger/src/shred.rs#L719
Data shreds are batched in chunks of 32 shreds each:
https://github.com/solana-labs/solana/blob/6907a2366/ledger/src/shred.rs#L714
However the very last batch of data shreds in a slot can be small, in
which case the loss rate can be exacerbated.
This commit expands the number of coding shreds in the last FEC block in
slots to: 64 - number of data shreds; so that FEC blocks are always 64
data and parity coding shreds each.
As a consequence of this, the last FEC block has more parity coding
shreds than data shreds. So for some shred indices we will have a coding
shred but no data shreds. This should not cause any kind of overlapping
FEC blocks as in:
https://github.com/solana-labs/solana/pull/10095
since this is done only for the very last batch in a slot, and the next
slot will reset the shred index.
* Track transaction check time separately from account loads
* banking packet process metrics
* Remove signature clone in status cache lookup
* Reduce allocations when converting packets to transactions
* Add blake3 hash of transaction messages in status cache
* Bug fixes
* fix tests and run fmt
* Address feedback
* fix simd tx entry verification
* Fix rebase
* Feedback
* clean up
* Add tests
* Remove feature switch and fall back to signature check
* Bump programs/bpf Cargo.lock
* clippy
* nudge benches
* Bump `BankSlotDelta` frozen ABI hash`
* Add blake3 to sdk/programs/Cargo.lock
* nudge bpf tests
* short circuit status cache checks
Co-authored-by: Trent Nelson <trent@solana.com>
behzad nouri
Nicholas Clarke
Tao Zhu
Tyera Eulberg
Jon Cinque
HaoranYi
carllin
steviez
Brooks Prumo
Justin Starry
sakridge
Tao Zhu
Jeff Biseda
buffalu
anatoly yakovenko
Eric Warehime
Michael Vines
Jeff Washington (jwash)
Alexander Meißner
Tyera Eulberg