Trait sc_rpc_api::state::StateApiServer
source · pub trait StateApiServer<Hash>: Sized + Send + Sync + 'static {
Show 16 methods
// Required methods
fn call(
&self,
name: String,
bytes: Bytes,
hash: Option<Hash>
) -> RpcResult<Bytes>;
fn storage_keys(
&self,
prefix: StorageKey,
hash: Option<Hash>
) -> RpcResult<Vec<StorageKey>>;
fn storage_pairs(
&self,
prefix: StorageKey,
hash: Option<Hash>
) -> RpcResult<Vec<(StorageKey, StorageData)>>;
fn storage_keys_paged(
&self,
prefix: Option<StorageKey>,
count: u32,
start_key: Option<StorageKey>,
hash: Option<Hash>
) -> RpcResult<Vec<StorageKey>>;
fn storage(
&self,
key: StorageKey,
hash: Option<Hash>
) -> RpcResult<Option<StorageData>>;
fn storage_hash(
&self,
key: StorageKey,
hash: Option<Hash>
) -> RpcResult<Option<Hash>>;
fn storage_size<'life0, 'async_trait>(
&'life0 self,
key: StorageKey,
hash: Option<Hash>
) -> Pin<Box<dyn Future<Output = RpcResult<Option<u64>>> + Send + 'async_trait>>
where Self: 'async_trait,
'life0: 'async_trait;
fn metadata(&self, hash: Option<Hash>) -> RpcResult<Bytes>;
fn runtime_version(&self, hash: Option<Hash>) -> RpcResult<RuntimeVersion>;
fn query_storage(
&self,
keys: Vec<StorageKey>,
block: Hash,
hash: Option<Hash>
) -> RpcResult<Vec<StorageChangeSet<Hash>>>;
fn query_storage_at(
&self,
keys: Vec<StorageKey>,
at: Option<Hash>
) -> RpcResult<Vec<StorageChangeSet<Hash>>>;
fn read_proof(
&self,
keys: Vec<StorageKey>,
hash: Option<Hash>
) -> RpcResult<ReadProof<Hash>>;
fn trace_block(
&self,
block: Hash,
targets: Option<String>,
storage_keys: Option<String>,
methods: Option<String>
) -> RpcResult<TraceBlockResponse>;
fn subscribe_runtime_version(
&self,
subscription_sink: SubscriptionSink
) -> SubscriptionResult;
fn subscribe_storage(
&self,
subscription_sink: SubscriptionSink,
keys: Option<Vec<StorageKey>>
) -> SubscriptionResult;
// Provided method
fn into_rpc(self) -> RpcModule<Self>
where Hash: Send + Sync + 'static + DeserializeOwned + Serialize { ... }
}
Expand description
Server trait implementation for the StateApi
RPC API.
Required Methods§
sourcefn call(
&self,
name: String,
bytes: Bytes,
hash: Option<Hash>
) -> RpcResult<Bytes>
fn call( &self, name: String, bytes: Bytes, hash: Option<Hash> ) -> RpcResult<Bytes>
Call a method from the runtime API at a block’s state.
sourcefn storage_keys(
&self,
prefix: StorageKey,
hash: Option<Hash>
) -> RpcResult<Vec<StorageKey>>
fn storage_keys( &self, prefix: StorageKey, hash: Option<Hash> ) -> RpcResult<Vec<StorageKey>>
Returns the keys with prefix, leave empty to get all the keys.
sourcefn storage_pairs(
&self,
prefix: StorageKey,
hash: Option<Hash>
) -> RpcResult<Vec<(StorageKey, StorageData)>>
fn storage_pairs( &self, prefix: StorageKey, hash: Option<Hash> ) -> RpcResult<Vec<(StorageKey, StorageData)>>
Returns the keys with prefix, leave empty to get all the keys
sourcefn storage_keys_paged(
&self,
prefix: Option<StorageKey>,
count: u32,
start_key: Option<StorageKey>,
hash: Option<Hash>
) -> RpcResult<Vec<StorageKey>>
fn storage_keys_paged( &self, prefix: Option<StorageKey>, count: u32, start_key: Option<StorageKey>, hash: Option<Hash> ) -> RpcResult<Vec<StorageKey>>
Returns the keys with prefix with pagination support.
Up to count
keys will be returned.
If start_key
is passed, return next keys in storage in lexicographic order.
sourcefn storage(
&self,
key: StorageKey,
hash: Option<Hash>
) -> RpcResult<Option<StorageData>>
fn storage( &self, key: StorageKey, hash: Option<Hash> ) -> RpcResult<Option<StorageData>>
Returns a storage entry at a specific block’s state.
sourcefn storage_hash(
&self,
key: StorageKey,
hash: Option<Hash>
) -> RpcResult<Option<Hash>>
fn storage_hash( &self, key: StorageKey, hash: Option<Hash> ) -> RpcResult<Option<Hash>>
Returns the hash of a storage entry at a block’s state.
sourcefn storage_size<'life0, 'async_trait>(
&'life0 self,
key: StorageKey,
hash: Option<Hash>
) -> Pin<Box<dyn Future<Output = RpcResult<Option<u64>>> + Send + 'async_trait>>where
Self: 'async_trait,
'life0: 'async_trait,
fn storage_size<'life0, 'async_trait>( &'life0 self, key: StorageKey, hash: Option<Hash> ) -> Pin<Box<dyn Future<Output = RpcResult<Option<u64>>> + Send + 'async_trait>>where Self: 'async_trait, 'life0: 'async_trait,
Returns the size of a storage entry at a block’s state.
sourcefn metadata(&self, hash: Option<Hash>) -> RpcResult<Bytes>
fn metadata(&self, hash: Option<Hash>) -> RpcResult<Bytes>
Returns the runtime metadata as an opaque blob.
sourcefn runtime_version(&self, hash: Option<Hash>) -> RpcResult<RuntimeVersion>
fn runtime_version(&self, hash: Option<Hash>) -> RpcResult<RuntimeVersion>
Get the runtime version.
sourcefn query_storage(
&self,
keys: Vec<StorageKey>,
block: Hash,
hash: Option<Hash>
) -> RpcResult<Vec<StorageChangeSet<Hash>>>
fn query_storage( &self, keys: Vec<StorageKey>, block: Hash, hash: Option<Hash> ) -> RpcResult<Vec<StorageChangeSet<Hash>>>
Query historical storage entries (by key) starting from a block given as the second parameter.
NOTE: The first returned result contains the initial state of storage for all keys. Subsequent values in the vector represent changes to the previous state (diffs). WARNING: The time complexity of this query is O(|keys|*dist(block, hash)), and the memory complexity is O(dist(block, hash)) – use with caution.
sourcefn query_storage_at(
&self,
keys: Vec<StorageKey>,
at: Option<Hash>
) -> RpcResult<Vec<StorageChangeSet<Hash>>>
fn query_storage_at( &self, keys: Vec<StorageKey>, at: Option<Hash> ) -> RpcResult<Vec<StorageChangeSet<Hash>>>
Query storage entries (by key) at a block hash given as the second parameter. NOTE: Each StorageChangeSet in the result corresponds to exactly one element – the storage value under an input key at the input block hash.
sourcefn read_proof(
&self,
keys: Vec<StorageKey>,
hash: Option<Hash>
) -> RpcResult<ReadProof<Hash>>
fn read_proof( &self, keys: Vec<StorageKey>, hash: Option<Hash> ) -> RpcResult<ReadProof<Hash>>
Returns proof of storage entries at a specific block’s state.
sourcefn trace_block(
&self,
block: Hash,
targets: Option<String>,
storage_keys: Option<String>,
methods: Option<String>
) -> RpcResult<TraceBlockResponse>
fn trace_block( &self, block: Hash, targets: Option<String>, storage_keys: Option<String>, methods: Option<String> ) -> RpcResult<TraceBlockResponse>
The traceBlock
RPC provides a way to trace the re-execution of a single
block, collecting Spans and Events from both the client and the relevant WASM runtime.
The Spans and Events are conceptually equivalent to those from the Tracing crate.
The structure of the traces follows that of the block execution pipeline, so meaningful interpretation of the traces requires an understanding of the Substrate chain’s block execution.
Link to conceptual map of trace structure for Polkadot and Kusama block execution.
Node requirements
- Fully synced archive node (i.e. a node that is not actively doing a “major” sync).
- Tracing enabled WASM runtimes for all runtime versions for which tracing is desired.
Node recommendations
- Use fast SSD disk storage.
- Run node flags to increase DB read speed (i.e.
--state-cache-size
,--db-cache
).
Creating tracing enabled WASM runtimes
- Checkout commit of chain version to compile with WASM traces
- diener can help to peg commit of substrate to what the chain expects.
- Navigate to the
runtime
folder/package of the chain - Add feature
with-tracing = ["frame-executive/with-tracing", "sp-io/with-tracing"]
under[features]
to theruntime
packages’Cargo.toml
. - Compile the runtime with
cargo build --release --features with-tracing
- Tracing-enabled WASM runtime should be found in
./target/release/wbuild/{{chain}}-runtime
and be called something like{{your_chain}}_runtime.compact.wasm
. This can be renamed/modified however you like, as long as it retains the.wasm
extension. - Run the node with the wasm blob overrides by placing them in a folder with all your runtimes, and passing the path of this folder to your chain, e.g.:
./target/release/polkadot --wasm-runtime-overrides /home/user/my-custom-wasm-runtimes
You can also find some pre-built tracing enabled wasm runtimes in substrate-archive
RPC Usage
The RPC allows for two filtering mechanisms: tracing targets and storage key prefixes. The filtering of spans and events takes place after they are all collected; so while filters do not reduce time for actual block re-execution, they reduce the response payload size.
Note: storage events primarily come from primitives/state-machine/src/ext.rs. The default filters can be overridden, see the params section for details.
curl
example
- Get tracing spans and events
curl \
-H "Content-Type: application/json" \
-d '{"id":1, "jsonrpc":"2.0", "method": "state_traceBlock", \
"params": ["0xb246acf1adea1f801ce15c77a5fa7d8f2eb8fed466978bcee172cc02cf64e264", "pallet,frame,state", "", ""]}' \
http://localhost:9933/
- Get tracing events with all
storage_keys
curl \
-H "Content-Type: application/json" \
-d '{"id":1, "jsonrpc":"2.0", "method": "state_traceBlock", \
"params": ["0xb246acf1adea1f801ce15c77a5fa7d8f2eb8fed466978bcee172cc02cf64e264", "state", "", ""]}' \
http://localhost:9933/
- Get tracing events with
storage_keys
(‘f0c365c3cf59d671eb72da0e7a4113c4’)
curl \
-H "Content-Type: application/json" \
-d '{"id":1, "jsonrpc":"2.0", "method": "state_traceBlock", \
"params": ["0xb246acf1adea1f801ce15c77a5fa7d8f2eb8fed466978bcee172cc02cf64e264", "state", "f0c365c3cf59d671eb72da0e7a4113c4", ""]}' \
http://localhost:9933/
- Get tracing events with
storage_keys
(‘f0c365c3cf59d671eb72da0e7a4113c4’) and method (‘Put’)
curl \
-H "Content-Type: application/json" \
-d '{"id":1, "jsonrpc":"2.0", "method": "state_traceBlock", \
"params": ["0xb246acf1adea1f801ce15c77a5fa7d8f2eb8fed466978bcee172cc02cf64e264", "state", "f0c365c3cf59d671eb72da0e7a4113c4", "Put"]}' \
http://localhost:9933/
- Get tracing events with all
storage_keys
and method (‘Put’)
curl \
-H "Content-Type: application/json" \
-d '{"id":1, "jsonrpc":"2.0", "method": "state_traceBlock", \
"params": ["0xb246acf1adea1f801ce15c77a5fa7d8f2eb8fed466978bcee172cc02cf64e264", "state", "", "Put"]}' \
http://localhost:9933/
Params
block
(param index 0): Hash of the block to trace.targets
(param index 1): String of comma separated (no spaces) targets. Specified targets match with trace targets by prefix (i.e if a target is in the beginning of a trace target it is considered a match). If an empty string is specified no targets will be filtered out. The majority of targets correspond to Rust module names, and the ones that do not are typically “hardcoded” into span or event location somewhere in the Substrate source code. (“Non-hardcoded” targets typically come from frame support macros.)storage_keys
(param index 2): String of comma separated (no spaces) hex encoded (no0x
prefix) storage keys. If an empty string is specified no events will be filtered out. If anything other than an empty string is specified, events will be filtered by storage key (so non-storage events will not show up). You can specify any length of a storage key prefix (i.e. if a specified storage key is in the beginning of an events storage key it is considered a match). Example: for balance tracking on Polkadot & Kusama you would likely want to track changes to account balances with the frame_system::Account storage item, which is a map fromAccountId
toAccountInfo
. The key filter for this would be the storage prefix for the map:26aa394eea5630e07c48ae0c9558cef7b99d880ec681799c0cf30e8886371da9
methods
(param index 3): String of comma separated (no spaces) tracing event method. If an empty string is specified no events will be filtered out. If anything other than an empty string is specified, events will be filtered by method (so non-method events will not show up).
Additionally you would want to track the extrinsic index, which is under the
:extrinsic_index
key. The key for this would be the aforementioned string as bytes
in hex: 3a65787472696e7369635f696e646578
.
The following are some resources to learn more about storage keys in substrate:
substrate storage, transparent keys in substrate,
querying substrate storage via rpc.
Maximum payload size
The maximum payload size allowed is 15mb. Payloads over this size will return a object with a simple error message. If you run into issues with payload size you can narrow down the traces using a smaller set of targets and/or storage keys.
If you are having issues with maximum payload size you can use the flag
-ltracing=trace
to get some logging during tracing.
sourcefn subscribe_runtime_version(
&self,
subscription_sink: SubscriptionSink
) -> SubscriptionResult
fn subscribe_runtime_version( &self, subscription_sink: SubscriptionSink ) -> SubscriptionResult
New runtime version subscription
sourcefn subscribe_storage(
&self,
subscription_sink: SubscriptionSink,
keys: Option<Vec<StorageKey>>
) -> SubscriptionResult
fn subscribe_storage( &self, subscription_sink: SubscriptionSink, keys: Option<Vec<StorageKey>> ) -> SubscriptionResult
New storage subscription