rpc v2: backpressure chainHead_v1_storage #5741
Conversation
substrate/client/rpc-spec-v2/src/chain_head/chain_head_storage.rs
Outdated
Show resolved
Hide resolved
substrate/client/rpc-spec-v2/src/chain_head/chain_head_storage.rs
Outdated
Show resolved
Hide resolved
|
The CI pipeline was cancelled due to failure one of the required jobs. |
niklasad1
changed the title
Storage::query_iter
niklasad1
changed the title
substrate/client/rpc-spec-v2/src/chain_head/subscription/inner.rs
Outdated
Show resolved
Hide resolved
substrate/client/rpc-spec-v2/src/chain_head/subscription/inner.rs
Outdated
Show resolved
Hide resolved
substrate/client/rpc-spec-v2/src/chain_head/subscription/inner.rs
Outdated
Show resolved
Hide resolved
| impl OperationState { | ||
| pub fn stop(&mut self) { | ||
| if !self.stop.is_stopped() { | ||
| self.operations.lock().remove(&self.operation_id); |
There was a problem hiding this comment.
annoying to lock the mutex here instead of AtomicBool but this is needed to have an async notification when operation is stopped.
couldn't find a better way for this
There was a problem hiding this comment.
I think this is acceptable since the operationStop shouldn't be happening too often if at all. We are also acquiring the mutex on dropping RegisteredOperations to clean up the tracing of operation IDs.
substrate/client/rpc-spec-v2/src/chain_head/chain_head_storage.rs
Outdated
Show resolved
Hide resolved
|
bot fmt |
|
@niklasad1 https://gitlab.parity.io/parity/mirrors/polkadot-sdk/-/jobs/7424018 was started for your command Comment |
|
@niklasad1 Command |
| operation_id: operation.operation_id(), | ||
| })); | ||
| } | ||
|
|
||
| /// Generate the block events for the `chainHead_storage` method. | ||
| pub async fn generate_events( |
There was a problem hiding this comment.
I wasn't happy with this generate_events and generate_events_iter, so now I have merged them which provides proper backpressure to all storage queries
|
bot fmt |
|
@niklasad1 https://gitlab.parity.io/parity/mirrors/polkadot-sdk/-/jobs/7453031 was started for your command Comment |
|
@niklasad1 Command |
| ) -> Result<(), tokio::task::JoinError> { | ||
| let this = self.clone(); | ||
|
|
||
| tokio::task::spawn_blocking(move || { |
There was a problem hiding this comment.
Nice! We are doing a spawn_blocking here instead of in chain_head_unstable_storage. Looks good to me! Thanks for simplifying this module!
Close #5589
This PR makes it possible for
rpc_v2::Storage::query_iter_paginatedto be "backpressured" which is achieved by having a channel where the result is sent back and when this channel is "full" we pause the iteration.The chainHead_follow has an internal channel which doesn't represent the actual connection and that is set to a very small number (16). Recall that the JSON-RPC server has a dedicate buffer for each connection by default of 64.
Notes
archive_storagealso depends onrpc_v2::Storage::query_iter_paginatedI had to tweak the method to support limits as well. The reason is that archive_storage won't get backpressured properly because it's not an subscription. (it would much easier if it would be a subscription in rpc v2 spec because nothing against querying huge amount storage keys)query_iter_paginateddoesn't necessarily return the storage "in order" such asquery_iter_paginated(vec![("key1", hash), ("key2", value)], ...)could return them in arbitrary order because it's wrapped in FuturesUnordered but I could change that if we want to process it inorder (it's slower)chainHead_v1_storage callrather than the rpc max message limit which 10MB and only allowed to max 16 callschainHead_v1_xconcurrently (this should be fine)Benchmarks using subxt on localhost
The reason for this is because as Josep explained in the issue is that one is only allowed query five storage items per call and clients has make lots of calls to drive it forward..