rocksdb compaction can cause long stalls, so
make it more configurable to try and reduce those stalls
and also to coordinate between multiple nodes to not induce
stall at the same time.
* introduce get largest accounts cache
* remove cache size and change hash key
* remove eq and hash derivation from commitment config
* add slot to the cache
* Plumb allow-dead-slots to ledger-tool verify
* ledger-tool cleanup and add some useful missing args
Print root slots and how many unrooted past last root.
* Update tonic & prost, and regenerate proto
* Reignore doc code
* Revert pull #14367, but pin tokio to v0.2 for jsonrpc
* Bump backoff and goauth -> and therefore tokio
* Bump tokio in faucet, net-utils
* Bump remaining tokio, plus tarpc
* Add validator flag to opt in to cpi and logs storage
* Default TestValidator to opt-in; allow using in multinode-demo
* No clone
Co-authored-by: Carl Lin <carl@solana.com>
* Deprecate commitment variants
* Add new CommitmentConfig builders
* Add helpers to avoid allowing deprecated variants
* Remove deprecated transaction-status code
* Include new commitment variants in runtime commitment; allow deprecated as long as old variants persist
* Remove deprecated banks code
* Remove deprecated variants in core; allow deprecated in rpc/rpc-subscriptions for now
* Heavier hand with rpc/rpc-subscription commitment
* Remove deprecated variants from local-cluster
* Remove deprecated variants from various tools
* Remove deprecated variants from validator
* Update docs
* Remove deprecated client code
* Add new variants to cli; remove deprecated variants as possible
* Don't send new commitment variants to old clusters
* Retain deprecated method in test_validator_saves_tower
* Fix clippy matches! suggestion for BPF solana-sdk legacy compile test
* Refactor node version check to handle commitment variants and transaction encoding
* Hide deprecated variants from cli help
* Add cli App comments
* Expose not-found bigtable error
* Add custom rpc error for bigtable data not found
* Return custom rpc error when bigtable block is not found
* Generalize long-term storage
* Configure bigtable's timeout when read-only
* Review comments
* Apply nits (thanks!)
Co-authored-by: Michael Vines <mvines@gmail.com>
* Timeout in the streamed decoding as well
Co-authored-by: Michael Vines <mvines@gmail.com>
* Add optimistically_confirmed field to TransactionStatus
* Update docs
* Convert new field to confirmation_status
* Update docs to confirmationStatus
* Update variants
* Update docs
* Just Confirmed
* fill in timing gaps in replay_stage
* add replay_stage bank_count metric
* formatting
* handle another gap
* cleanup wait_receive_time to be more straightforward
Crds values buffered when responding to pull-requests can be very large taking a lot of memory.
Added a limit for number of buffered crds values based on outbound data budget.
record_labels returns all the possible labels for a record identified by
a pubkey, used in updating timestamp of crds values:
https://github.com/solana-labs/solana/blob/1792100e2/core/src/crds_value.rs#L560-L577https://github.com/solana-labs/solana/blob/1792100e2/core/src/crds.rs#L240-L251
The code relies on CrdsValueLabel to be limited to a small deterministic
set of possible values for a fixed pubkey. As we expand crds values to
include duplicate shreds, this limits what the duplicate proofs can be
keyed by in the table.
In addition the computation of these labels is inefficient and will
become more so as duplicate shreds and more types of crds values are
added. An alternative is to maintain an index of all crds values
associated with a pubkey.
* Move bank drop to AccountsBackgroundService
* Send to ABS on drop instead, protects against other places banks are dropped
* Fix Abi
* test
Co-authored-by: Carl Lin <carl@solana.com>
* feat: store pre / post token balances
* move helper functions into separate include
* move token balance functionality to transaction-status crate
* fix blockstore processor test
* fix bigtable legacy test
* add caching to decimals
If a node "a" receives instance-info from node "b1" it will override any
instance-info associated with "b1" pubkey in its crds table. This makes
it less likely that when "b1" receives crds values from "a" (either
through pull or push), it sees other instances of itself (because node
"a" discarded them when it received "b1" instance info).
In order for the crds table to contain all instance-info associated with
the same pubkey at the same time, we need to add the instance tokens to
the keys in the crds table (i.e. the CrdsValueLabel).
* Adds a CLI option to the validator to enable just-in-time compilation of BPF.
* Refactoring to use bpf_loader_program instead of feature_set to pass JIT flag from the validator CLI to the executor.
Gossip and other places repeatedly de-serialize vote-state stored in
vote accounts. Ideally the first de-serialization should cache the
result.
This commit adds new VoteAccount type which lazily de-serializes
VoteState from Account data and caches the result internally.
Serialize and Deserialize traits are manually implemented to match
existing code. So, despite changes to frozen_abi, this commit should be
backward compatible.
* runtime: Add `FeeCalculator` resolution method to `HashAgeKind`
* runtime: Plumb fee-collected accounts for failed nonce tx rollback
* runtime: Use fee-collected nonce/fee account for nonced TX error rollback
* runtime: Add test for failed nonced TX accounts rollback
* Fee payer test
* fixup: replace nonce account when it pays the fee
* fixup: nonce fee-payer collect test
* fixup: fixup: clippy/fmt for replace...
* runtime: Test for `HashAgeKind::fee_calculator()`
* Clippy
Co-authored-by: Trent Nelson <trent@solana.com>
process_pull_requests acquires a write lock on crds table to update
records timestamp for each of the pull-request callers:
https://github.com/solana-labs/solana/blob/3087c9049/core/src/crds_gossip_pull.rs#L287-L300
However, pull-requests overlap a lot in callers and this function ends
up doing a lot of redundant duplicate work.
This commit obtains unique callers before acquiring an exclusive lock on
crds table.
* Fix fragile tests in prep of stake rewrite pr
* Restore BOOTSTRAP_VALIDATOR_LAMPORTS where appropriate
* Further clean up
* Further clean up
* Aligh with other call site change
* Remove false warn!
* fix ci!
Validator logs show that prune messages are dropped because they exceed
packet data size:
https://github.com/solana-labs/solana/blob/f25c969ad/perf/src/packet.rs#L90-L92
This can exacerbate gossip traffic by redundantly increasing push
messages across network. The workaround is to break prunes into smaller
chunks and send over in multiple messages.
* Add TestValidator::new_with_fees constructor, and warning for low bootstrap_validator_lamports
* Add logging to solana-tokens integration test to help catch low bootstrap_validator_lamports in the future
* Reasonable TestValidator mint_lamports
split_gossip_messages:
https://github.com/solana-labs/solana/blob/a97c04b40/core/src/cluster_info.rs#L1536-L1574
splits crds-values into chunks to fit into a gossip packet. However it is
using a global upper-bound for the header-size across all protocols:
https://github.com/solana-labs/solana/blob/a97c04b40/core/src/cluster_info.rs#L90-L93
This can be wasteful as the specific gossip protocol can have smaller
header than this upper-bound (e.g. Protocol::PushMessage is 170 bytes
smaller). Adding more crds-values in one gossip packet can avoid the
overheads of separate packets and reduce total number of bytes sent over
the wire.
This commit updates the splitting function to take a max-chunk-size
argument. At call-site, this value is set to the size of the protocol
which the values are sent over.
In several places in gossip code, the entire crds table is scanned only
to filter out nodes' contact infos. Currently on mainnet, crds table is
of size ~70k, while there are only ~470 nodes. So the full table scan is
inefficient. Instead we may maintain an index of only nodes' contact
infos.
The --rpc-pubsub-enable-vote-subscription flag may be used to enable it.
The current vote subscription is problematic because it emits a
notification for *every* vote, so hundreds a second in a real cluster.
Critically it's also missing information about *who* is voting,
rendering all those notifications practically useless.
Until these two issues can be resolved, the vote subscription is not
much more than a potential DoS vector.
* Discard pre hard fork persisted tower if hard-forking
* Relax config.require_tower
* Add cluster test
* nits
* Remove unnecessary check
Co-authored-by: Ryo Onodera <ryoqun@gmail.com>
Co-authored-by: Carl Lin <carl@solana.com>
* Fix slow/stuck unstaking due to toggling in epoch
* nits
* nits
* Add stake_program_v2 feature status check to cli
Co-authored-by: Tyera Eulberg <tyera@solana.com>
Packet::from_data is ignoring serialization errors:
https://github.com/solana-labs/solana/blob/d08c3232e/sdk/src/packet.rs#L42-L48
This is likely never useful as the packet will be sent over the wire
taking bandwidth but at the receiving end will either fail to
deserialize or it will be invalid.
This commit will propagate the errors out of the function to the
call-site, allowing the call-site to handle the error.
* Fix tower/blockstore unsync due to external causes
* Add and clean up long comments
* Clean up test
* Comment about warped_slot_history
* Run test_future_tower with master-only/master-slave
* Update comments about false leader condition
https://hackerone.com/reports/991106
> It’s possible to use UDP gossip protocol to amplify DDoS attacks. An attacker
> can spoof IP address in UDP packet when sending PullRequest to the node.
> There's no any validation if provided source IP address is not spoofed and
> the node can send much larger PullResponse to victim's IP. As I checked,
> PullRequest is about 290 bytes, while PullResponse is about 10 kB. It means
> that amplification is about 34x. This way an attacker can easily perform DDoS
> attack both on Solana node and third-party server.
>
> To prevent it, need for example to implement ping-pong mechanism similar as
> in Ethereum: Before accepting requests from remote client needs to validate
> his IP. Local node sends Ping packet to the remote node and it needs to reply
> with Pong packet that contains hash of matching Ping packet. Content of Ping
> packet is unpredictable. If hash from Pong packet matches, local node can
> remember IP where Ping packet was sent as correct and allow further
> communication.
>
> More info:
> https://github.com/ethereum/devp2p/blob/master/discv4.md#endpoint-proof
> https://github.com/ethereum/devp2p/blob/master/discv4.md#wire-protocol
The commit adds a PingCache, which maintains records of remote nodes
which have returned a valid response to a ping message, and on-the-fly
ping messages pending a pong response from the remote node.
When handling pull-requests, those from addresses which have not passed
the ping-pong check are filtered out, and additionally ping packets are
added for addresses which need to be (re)verified.