tvl-depot/tvix/store/src/directoryservice/mod.rs

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

67 lines
2.5 KiB
Rust
Raw Normal View History

use crate::{proto, B3Digest, Error};
mod from_addr;
feat(tvix/store/directorysvc): add gRPC client This provides a GRPCDirectoryService struct implementing DirectoryService, allowing a client to Directory objects from a (remote) tvix-store. Remote in this case is anything outside the current process, be it another process, or an endpoint on the network. To keep the sync interface in the `DirectoryService` trait, a handle to some tokio runtime needs to be passed into the constructor, and the two methods use `self.tokio_handle.spawn` to start an async function, and `self.tokio_handle.block_on` to wait for its completion. The client handle, called `grpc_client` itself is easy to clone, and treats concurrent requests internally. This means, even though we keep the `DirectoryService` trait sync, there's nothing preventing it from being used concurrently, let's say from multiple threads. There's still two limitations for now: 1) The trait doesn't make use of the `recursive` request, which currently leads to a N+1 query problem. This can be fixed by `GRPCDirectoryService` having a reference to another `DirectoryService` acting as the local side. I want to wait for general store composition code to pop up before manually coding this here. 2) It's currently only possible to put() leaf directory nodes, as the request normally requires uploading a whole closure. We might want to add another batch function to upload a whole closure, and/or do this batching in certain cases. This still needs some more thinking. Change-Id: I7ffec791610b72c0960cf5307cefbb12ec946dc9 Reviewed-on: https://cl.tvl.fyi/c/depot/+/8336 Tested-by: BuildkiteCI Autosubmit: flokli <flokli@flokli.de> Reviewed-by: tazjin <tazjin@tvl.su>
2023-03-23 13:49:57 +01:00
mod grpc;
mod memory;
mod sled;
mod traverse;
mod utils;
pub use self::from_addr::from_addr;
feat(tvix/store/directorysvc): add gRPC client This provides a GRPCDirectoryService struct implementing DirectoryService, allowing a client to Directory objects from a (remote) tvix-store. Remote in this case is anything outside the current process, be it another process, or an endpoint on the network. To keep the sync interface in the `DirectoryService` trait, a handle to some tokio runtime needs to be passed into the constructor, and the two methods use `self.tokio_handle.spawn` to start an async function, and `self.tokio_handle.block_on` to wait for its completion. The client handle, called `grpc_client` itself is easy to clone, and treats concurrent requests internally. This means, even though we keep the `DirectoryService` trait sync, there's nothing preventing it from being used concurrently, let's say from multiple threads. There's still two limitations for now: 1) The trait doesn't make use of the `recursive` request, which currently leads to a N+1 query problem. This can be fixed by `GRPCDirectoryService` having a reference to another `DirectoryService` acting as the local side. I want to wait for general store composition code to pop up before manually coding this here. 2) It's currently only possible to put() leaf directory nodes, as the request normally requires uploading a whole closure. We might want to add another batch function to upload a whole closure, and/or do this batching in certain cases. This still needs some more thinking. Change-Id: I7ffec791610b72c0960cf5307cefbb12ec946dc9 Reviewed-on: https://cl.tvl.fyi/c/depot/+/8336 Tested-by: BuildkiteCI Autosubmit: flokli <flokli@flokli.de> Reviewed-by: tazjin <tazjin@tvl.su>
2023-03-23 13:49:57 +01:00
pub use self::grpc::GRPCDirectoryService;
pub use self::memory::MemoryDirectoryService;
pub use self::sled::SledDirectoryService;
pub use self::traverse::traverse_to;
pub use self::utils::DirectoryTraverser;
/// The base trait all Directory services need to implement.
/// This is a simple get and put of [crate::proto::Directory], returning their
/// digest.
pub trait DirectoryService: Send + Sync {
/// Create a new instance by passing in a connection URL.
fn from_url(url: &url::Url) -> Result<Self, Error>
where
Self: Sized;
/// Get looks up a single Directory message by its digest.
/// In case the directory is not found, Ok(None) is returned.
fn get(&self, digest: &B3Digest) -> Result<Option<proto::Directory>, Error>;
/// Get uploads a single Directory message, and returns the calculated
/// digest, or an error.
fn put(&self, directory: proto::Directory) -> Result<B3Digest, Error>;
/// Looks up a closure of [proto::Directory].
/// Ideally this would be a `impl Iterator<Item = Result<proto::Directory, Error>>`,
/// and we'd be able to add a default implementation for it here, but
/// we can't have that yet.
fn get_recursive(
&self,
root_directory_digest: &B3Digest,
) -> Box<dyn Iterator<Item = Result<proto::Directory, Error>> + Send>;
/// Allows persisting a closure of [proto::Directory], which is a graph of
/// connected Directory messages.
fn put_multiple_start(&self) -> Box<dyn DirectoryPutter>;
}
/// Provides a handle to put a closure of connected [proto::Directory] elements.
///
/// The consumer can periodically call [DirectoryPutter::put], starting from the
/// leaves. Once the root is reached, [DirectoryPutter::close] can be called to
/// retrieve the root digest (or an error).
pub trait DirectoryPutter {
/// Put a individual [proto::Directory] into the store.
/// Error semantics and behaviour is up to the specific implementation of
/// this trait.
/// Due to bursting, the returned error might refer to an object previously
/// sent via `put`.
fn put(&mut self, directory: proto::Directory) -> Result<(), Error>;
/// Close the stream, and wait for any errors.
fn close(&mut self) -> Result<B3Digest, Error>;
/// Return whether the stream is closed or not.
/// Used from some [DirectoryService] implementations only.
fn is_closed(&self) -> bool;
}