feat(scanner): add bounded incarnation-scoped ACK receiver (#7182)

* chore(deps): refresh SDKs and pin clock skew regression coverage

Refresh compatible dependencies for Scanner/Heal V2 batch 1 and verify
the production S3 retry/signing path with a deterministic clock.

Co-Authored-By: heihutu <heihutu@gmail.com>
Co-Authored-By: zhi22915 <qiuzgang@gmail.com>

* feat(scanner): add bounded incarnation-scoped ACK receiver

Refs rustfs/backlog#2265 and rustfs/backlog#2240.

Co-Authored-By: heihutu <heihutu@gmail.com>
Co-Authored-By: zhi22915 <qiuzgang@gmail.com>

---------

Co-authored-by: heihutu <heihutu@gmail.com>
Co-authored-by: zhi22915 <qiuzgang@gmail.com>
This commit is contained in:
houseme
2026-09-05 16:54:55 +08:00
committed by GitHub
parent 0d1b312673
commit acfeef55ab
17 changed files with 1097 additions and 24 deletions
+10 -9
View File
@@ -196,15 +196,16 @@ pub mod bucket {
pub use crate::bucket::metadata_sys::ConfigWriteLockProbe;
pub use crate::bucket::metadata_sys::{
BucketMetadataMutationGuard, BucketMetadataSys, ObjectLockConfigState, acquire_bucket_metadata_transaction_lock,
acquire_bucket_metadata_transaction_lock_for_incarnation, capture_bucket_metadata_incarnation, delete,
delete_if_incarnation, delete_under_transaction_lock, get, get_accelerate_config, get_bucket_policy,
get_bucket_policy_raw, get_bucket_targets_config, get_config_from_disk, get_cors_config, get_durability_config,
get_global_bucket_metadata_sys, get_lifecycle_config, get_logging_config, get_notification_config,
get_object_lock_config, get_object_lock_config_state, get_on_demand_migration_config, get_public_access_block_config,
get_quota_config, get_replication_config, get_request_payment_config, get_sse_config, get_tagging_config,
get_versioning_config, get_website_config, init_bucket_metadata_sys, list_bucket_targets, reload_bucket_metadata,
remove_bucket_metadata, set_bucket_metadata, update, update_bucket_targets_under_transaction_lock,
update_config_with, update_if_incarnation, update_quota_if_incarnation, update_under_transaction_lock,
acquire_bucket_metadata_transaction_lock_for_incarnation, acquire_scanner_bucket_incarnation_fence,
capture_bucket_metadata_incarnation, delete, delete_if_incarnation, delete_under_transaction_lock, get,
get_accelerate_config, get_bucket_policy, get_bucket_policy_raw, get_bucket_targets_config, get_config_from_disk,
get_cors_config, get_durability_config, get_global_bucket_metadata_sys, get_lifecycle_config, get_logging_config,
get_notification_config, get_object_lock_config, get_object_lock_config_state, get_on_demand_migration_config,
get_public_access_block_config, get_quota_config, get_replication_config, get_request_payment_config, get_sse_config,
get_tagging_config, get_versioning_config, get_website_config, init_bucket_metadata_sys, list_bucket_targets,
reload_bucket_metadata, remove_bucket_metadata, set_bucket_metadata, update,
update_bucket_targets_under_transaction_lock, update_config_with, update_if_incarnation, update_quota_if_incarnation,
update_under_transaction_lock,
};
}
+114 -7
View File
@@ -655,6 +655,12 @@ pub struct BucketMetadataMutationGuard {
}
impl BucketMetadataMutationGuard {
/// Returns the storage-verified identity while both incarnation fences remain valid.
pub fn checked_bucket_incarnation(&self) -> Result<(&str, Uuid)> {
self.ensure_valid(&self.bucket)?;
Ok((&self.bucket, self.incarnation_id))
}
fn ensure_valid(&self, bucket: &str) -> Result<()> {
if self.bucket != bucket {
return Err(Error::other("bucket metadata mutation guard does not match bucket"));
@@ -674,6 +680,29 @@ async fn acquire_config_write_guard_for_incarnation(
sys: Arc<RwLock<BucketMetadataSys>>,
bucket: &str,
expected_incarnation_id: Option<Uuid>,
) -> Result<BucketMetadataMutationGuard> {
acquire_config_write_guard_with_migration(sys, bucket, expected_incarnation_id, true).await
}
/// Scanner probes must not create an incarnation to make a capability available.
pub async fn acquire_scanner_bucket_incarnation_fence(
bucket: &str,
expected_incarnation_id: Uuid,
expected_owner_id: Uuid,
) -> Result<BucketMetadataMutationGuard> {
super::utils::check_valid_bucket_name(bucket)?;
let sys = get_bucket_metadata_sys()?;
if expected_owner_id.is_nil() || sys.read().await.api.id != expected_owner_id || expected_incarnation_id.is_nil() {
return Err(Error::other("scanner bucket incarnation owner does not match"));
}
acquire_config_write_guard_with_migration(sys, bucket, Some(expected_incarnation_id), false).await
}
async fn acquire_config_write_guard_with_migration(
sys: Arc<RwLock<BucketMetadataSys>>,
bucket: &str,
expected_incarnation_id: Option<Uuid>,
migrate: bool,
) -> Result<BucketMetadataMutationGuard> {
let metadata_sys = sys.read().await.clone();
let lifecycle_guard = metadata_sys.api.acquire_bucket_lifecycle_read_lock(bucket).await?;
@@ -681,13 +710,15 @@ async fn acquire_config_write_guard_for_incarnation(
// Legacy buckets are migrated while the lifecycle fence prevents a
// same-name replacement. The second read under the write transaction is
// the CAS source of truth for the actual rewrite.
await_bucket_namespace_operation(
Some(&lifecycle_guard),
bucket,
"bucket config incarnation migration",
metadata_sys.get_bucket_incarnation_id(bucket),
)
.await?;
if migrate {
await_bucket_namespace_operation(
Some(&lifecycle_guard),
bucket,
"bucket config incarnation migration",
metadata_sys.get_bucket_incarnation_id(bucket),
)
.await?;
}
let transaction_guard = await_bucket_namespace_operation(
Some(&lifecycle_guard),
bucket,
@@ -3176,6 +3207,82 @@ mod tests {
);
}
#[tokio::test]
async fn scoped_dirty_usage_incarnation_probe_does_not_migrate_legacy_metadata() {
let (dirs, store) = isolated_store_over_temp_disks().await;
let sys = Arc::new(RwLock::new(BucketMetadataSys::new(store.clone())));
let bucket = "scoped-ack-legacy";
for dir in &dirs {
std::fs::create_dir_all(dir.path().join(bucket)).expect("create legacy bucket");
}
let mut metadata = BucketMetadata::new(bucket);
metadata.bucket_incarnation_id = Uuid::nil();
sys.read()
.await
.persist_and_set(metadata)
.await
.expect("persist legacy metadata");
assert!(
acquire_config_write_guard_with_migration(sys.clone(), bucket, Some(Uuid::new_v4()), false)
.await
.is_err()
);
assert!(load_bucket_incarnation(store, bucket).await.expect("read sidecar").is_none());
assert!(
sys.read()
.await
.get_config_from_disk(bucket)
.await
.expect("read metadata")
.bucket_incarnation_id
.is_nil()
);
}
#[tokio::test(flavor = "multi_thread", worker_threads = 4)]
#[serial]
async fn scoped_dirty_usage_incarnation_rejects_deleted_and_recreated_bucket() {
let (_dirs, store) = isolated_store_over_temp_disks().await;
init_bucket_metadata_sys(store.clone(), Vec::new()).await;
let sys = bucket_metadata_sys_of(&store.ctx).expect("metadata owner");
let bucket = "scoped-ack-recreated";
store
.make_bucket(bucket, &MakeBucketOptions::default())
.await
.expect("create bucket");
let old = store.bucket_incarnation_id_from_disk(bucket).await.expect("old incarnation");
let guard = acquire_config_write_guard_with_migration(sys.clone(), bucket, Some(old), false)
.await
.expect("trusted incarnation fence");
assert_eq!(guard.checked_bucket_incarnation().expect("valid fences"), (bucket, old));
drop(guard);
store
.delete_bucket(bucket, &DeleteBucketOptions::default())
.await
.expect("delete bucket");
assert!(
acquire_config_write_guard_with_migration(sys.clone(), bucket, Some(old), false)
.await
.is_err()
);
store
.make_bucket(bucket, &MakeBucketOptions::default())
.await
.expect("recreate bucket");
let new = store.bucket_incarnation_id_from_disk(bucket).await.expect("new incarnation");
assert_ne!(old, new);
assert!(
acquire_config_write_guard_with_migration(sys.clone(), bucket, Some(old), false)
.await
.is_err()
);
assert!(
acquire_config_write_guard_with_migration(sys, bucket, Some(new), false)
.await
.is_ok()
);
}
#[tokio::test]
async fn old_node_metadata_rewrite_cannot_replace_bucket_incarnation_sidecar() {
let (dirs, ecstore) = isolated_store_over_temp_disks().await;
+19
View File
@@ -30,6 +30,7 @@ use rustfs_protos::{
ChannelClass, create_new_channel, get_channel_for_class,
proto_gen::node_service::{
heal_control_service_client::HealControlServiceClient, node_service_client::NodeServiceClient,
scanner_control_service_client::ScannerControlServiceClient,
tier_mutation_control_service_client::TierMutationControlServiceClient,
},
};
@@ -60,6 +61,24 @@ pub async fn node_service_time_out_client(
node_service_time_out_client_for_class(addr, interceptor, ChannelClass::Control).await
}
pub(crate) async fn scanner_control_time_out_client(
addr: &str,
interceptor: TonicInterceptor,
) -> crate::error::Result<ScannerControlServiceClient<InterceptedService<AuthenticatedChannel, TonicInterceptor>>> {
let interceptor = interceptor.with_rpc_audience(addr)?;
let channel = match runtime_sources::cached_node_channel(addr).await {
Some(channel) => channel,
None => create_new_channel(addr)
.await
.map_err(|err| crate::error::Error::other(err.to_string()))?,
};
let channel = ReplayScopeChannel::new(channel, interceptor.replay_scope_audience());
let limit = rustfs_protos::scoped_dirty_usage::SCOPED_DIRTY_USAGE_MAX_REQUEST_BYTES as usize;
Ok(ScannerControlServiceClient::with_interceptor(channel, interceptor)
.max_decoding_message_size(limit)
.max_encoding_message_size(limit))
}
pub async fn heal_control_time_out_client(
addr: &str,
interceptor: TonicInterceptor,
@@ -2050,6 +2050,53 @@ impl PeerRestClient {
.await
}
/// Probe only: scoped ACK production requires a durable per-bucket proof.
pub async fn scanner_scoped_dirty_usage_capability(
&self,
owner_id: String,
instance_id: String,
entries: Vec<rustfs_protos::proto_gen::node_service::ScannerScopedDirtyUsageEntry>,
) -> Result<bool> {
use rustfs_protos::scoped_dirty_usage::*;
let payload = rustfs_protos::proto_gen::node_service::ScannerScopedDirtyUsageAckRequest {
challenge: Uuid::new_v4().as_bytes().to_vec().into(),
protocol_version: SCOPED_DIRTY_USAGE_PROTOCOL_VERSION,
owner_id,
instance_id,
scope: SCOPED_DIRTY_USAGE_BUCKET_SCOPE,
probe_only: true,
entries,
};
let canonical = canonical_scoped_dirty_usage_request(&payload).map_err(|err| Error::other(err.to_string()))?;
self.finalize_result(
async {
let mut client = super::client::scanner_control_time_out_client(
&self.grid_host,
TonicInterceptor::Signature(gen_tonic_signature_interceptor()),
)
.await?;
let mut request = Request::new(payload.clone());
set_tonic_canonical_body_digest(&mut request, &canonical)?;
let response = client.scanner_scoped_dirty_usage_ack(request).await?.into_inner();
let body = canonical_scoped_dirty_usage_response(&canonical, &response)
.map_err(|_| Error::other("scoped dirty usage capability response is too large"))?;
verify_tonic_rpc_response_proof(&body, response.response_proof.as_ref())?;
if response.protocol_version != SCOPED_DIRTY_USAGE_PROTOCOL_VERSION
|| response.owner_id != payload.owner_id
|| response.instance_id != payload.instance_id
|| response.max_entries != SCOPED_DIRTY_USAGE_MAX_ENTRIES
|| response.max_request_bytes != SCOPED_DIRTY_USAGE_MAX_REQUEST_BYTES
|| response.cleared != 0
{
return Err(Error::other("scoped dirty usage capability response does not match request"));
}
Ok(response.supported)
}
.await,
)
.await
}
pub async fn acknowledge_scanner_dirty_usage(&self, instance_id: String, generation: u64) -> Result<ScannerPeerActivity> {
let result = self
.scanner_activity_request_with_protocol(instance_id.clone(), generation, SCANNER_ACTIVITY_PROTOCOL_VERSION)
@@ -1283,6 +1283,54 @@ pub struct ScannerDirtyUsageSnapshotResponse {
#[prost(bytes = "bytes", tag = "7")]
pub response_proof: ::prost::bytes::Bytes,
}
/// Receiver-only protocol. Producers must retain whole-cycle ACK until they
/// have a durable per-bucket publication proof.
#[derive(Clone, PartialEq, Eq, Hash, ::prost::Message)]
pub struct ScannerScopedDirtyUsageEntry {
#[prost(string, tag = "1")]
pub bucket: ::prost::alloc::string::String,
#[prost(bytes = "bytes", tag = "2")]
pub bucket_incarnation: ::prost::bytes::Bytes,
#[prost(uint64, tag = "3")]
pub generation: u64,
}
#[derive(Clone, PartialEq, ::prost::Message)]
pub struct ScannerScopedDirtyUsageAckRequest {
#[prost(bytes = "bytes", tag = "1")]
pub challenge: ::prost::bytes::Bytes,
#[prost(uint32, tag = "2")]
pub protocol_version: u32,
#[prost(string, tag = "3")]
pub owner_id: ::prost::alloc::string::String,
#[prost(string, tag = "4")]
pub instance_id: ::prost::alloc::string::String,
/// Only scope 1 (a complete bucket) is supported; zero is invalid.
#[prost(uint32, tag = "5")]
pub scope: u32,
#[prost(bool, tag = "6")]
pub probe_only: bool,
#[prost(message, repeated, tag = "7")]
pub entries: ::prost::alloc::vec::Vec<ScannerScopedDirtyUsageEntry>,
}
#[derive(Clone, PartialEq, Eq, Hash, ::prost::Message)]
pub struct ScannerScopedDirtyUsageAckResponse {
#[prost(uint32, tag = "1")]
pub protocol_version: u32,
#[prost(string, tag = "2")]
pub owner_id: ::prost::alloc::string::String,
#[prost(string, tag = "3")]
pub instance_id: ::prost::alloc::string::String,
#[prost(bool, tag = "4")]
pub supported: bool,
#[prost(uint32, tag = "5")]
pub max_entries: u32,
#[prost(uint32, tag = "6")]
pub max_request_bytes: u32,
#[prost(uint64, tag = "7")]
pub cleared: u64,
#[prost(bytes = "bytes", tag = "8")]
pub response_proof: ::prost::bytes::Bytes,
}
/// A short-lived storage-owned read admission used only around a final
/// scanner metadata publication. It is intentionally separate from the
/// ScannerActivity observation wire so v6/v7 rolling compatibility remains
@@ -6282,6 +6330,244 @@ pub mod node_service_server {
}
}
/// Generated client implementations.
pub mod scanner_control_service_client {
#![allow(unused_variables, dead_code, missing_docs, clippy::wildcard_imports, clippy::let_unit_value)]
use tonic::codegen::http::Uri;
use tonic::codegen::*;
#[derive(Debug, Clone)]
pub struct ScannerControlServiceClient<T> {
inner: tonic::client::Grpc<T>,
}
impl ScannerControlServiceClient<tonic::transport::Channel> {
/// Attempt to create a new client by connecting to a given endpoint.
pub async fn connect<D>(dst: D) -> Result<Self, tonic::transport::Error>
where
D: TryInto<tonic::transport::Endpoint>,
D::Error: Into<StdError>,
{
let conn = tonic::transport::Endpoint::new(dst)?.connect().await?;
Ok(Self::new(conn))
}
}
impl<T> ScannerControlServiceClient<T>
where
T: tonic::client::GrpcService<tonic::body::Body>,
T::Error: Into<StdError>,
T::ResponseBody: Body<Data = Bytes> + std::marker::Send + 'static,
<T::ResponseBody as Body>::Error: Into<StdError> + std::marker::Send,
{
pub fn new(inner: T) -> Self {
let inner = tonic::client::Grpc::new(inner);
Self { inner }
}
pub fn with_origin(inner: T, origin: Uri) -> Self {
let inner = tonic::client::Grpc::with_origin(inner, origin);
Self { inner }
}
pub fn with_interceptor<F>(inner: T, interceptor: F) -> ScannerControlServiceClient<InterceptedService<T, F>>
where
F: tonic::service::Interceptor,
T::ResponseBody: Default,
T: tonic::codegen::Service<
http::Request<tonic::body::Body>,
Response = http::Response<<T as tonic::client::GrpcService<tonic::body::Body>>::ResponseBody>,
>,
<T as tonic::codegen::Service<http::Request<tonic::body::Body>>>::Error:
Into<StdError> + std::marker::Send + std::marker::Sync,
{
ScannerControlServiceClient::new(InterceptedService::new(inner, interceptor))
}
/// Compress requests with the given encoding.
///
/// This requires the server to support it otherwise it might respond with an
/// error.
#[must_use]
pub fn send_compressed(mut self, encoding: CompressionEncoding) -> Self {
self.inner = self.inner.send_compressed(encoding);
self
}
/// Enable decompressing responses.
#[must_use]
pub fn accept_compressed(mut self, encoding: CompressionEncoding) -> Self {
self.inner = self.inner.accept_compressed(encoding);
self
}
/// Limits the maximum size of a decoded message.
///
/// Default: `4MB`
#[must_use]
pub fn max_decoding_message_size(mut self, limit: usize) -> Self {
self.inner = self.inner.max_decoding_message_size(limit);
self
}
/// Limits the maximum size of an encoded message.
///
/// Default: `usize::MAX`
#[must_use]
pub fn max_encoding_message_size(mut self, limit: usize) -> Self {
self.inner = self.inner.max_encoding_message_size(limit);
self
}
pub async fn scanner_scoped_dirty_usage_ack(
&mut self,
request: impl tonic::IntoRequest<super::ScannerScopedDirtyUsageAckRequest>,
) -> std::result::Result<tonic::Response<super::ScannerScopedDirtyUsageAckResponse>, tonic::Status> {
self.inner
.ready()
.await
.map_err(|e| tonic::Status::unknown(format!("Service was not ready: {}", e.into())))?;
let codec = tonic_prost::ProstCodec::default();
let path = http::uri::PathAndQuery::from_static("/node_service.ScannerControlService/ScannerScopedDirtyUsageAck");
let mut req = request.into_request();
req.extensions_mut()
.insert(GrpcMethod::new("node_service.ScannerControlService", "ScannerScopedDirtyUsageAck"));
self.inner.unary(req, path, codec).await
}
}
}
/// Generated server implementations.
pub mod scanner_control_service_server {
#![allow(unused_variables, dead_code, missing_docs, clippy::wildcard_imports, clippy::let_unit_value)]
use tonic::codegen::*;
/// Generated trait containing gRPC methods that should be implemented for use with ScannerControlServiceServer.
#[async_trait]
pub trait ScannerControlService: std::marker::Send + std::marker::Sync + 'static {
async fn scanner_scoped_dirty_usage_ack(
&self,
request: tonic::Request<super::ScannerScopedDirtyUsageAckRequest>,
) -> std::result::Result<tonic::Response<super::ScannerScopedDirtyUsageAckResponse>, tonic::Status>;
}
#[derive(Debug)]
pub struct ScannerControlServiceServer<T> {
inner: Arc<T>,
accept_compression_encodings: EnabledCompressionEncodings,
send_compression_encodings: EnabledCompressionEncodings,
max_decoding_message_size: Option<usize>,
max_encoding_message_size: Option<usize>,
}
impl<T> ScannerControlServiceServer<T> {
pub fn new(inner: T) -> Self {
Self::from_arc(Arc::new(inner))
}
pub fn from_arc(inner: Arc<T>) -> Self {
Self {
inner,
accept_compression_encodings: Default::default(),
send_compression_encodings: Default::default(),
max_decoding_message_size: None,
max_encoding_message_size: None,
}
}
pub fn with_interceptor<F>(inner: T, interceptor: F) -> InterceptedService<Self, F>
where
F: tonic::service::Interceptor,
{
InterceptedService::new(Self::new(inner), interceptor)
}
/// Enable decompressing requests with the given encoding.
#[must_use]
pub fn accept_compressed(mut self, encoding: CompressionEncoding) -> Self {
self.accept_compression_encodings.enable(encoding);
self
}
/// Compress responses with the given encoding, if the client supports it.
#[must_use]
pub fn send_compressed(mut self, encoding: CompressionEncoding) -> Self {
self.send_compression_encodings.enable(encoding);
self
}
/// Limits the maximum size of a decoded message.
///
/// Default: `4MB`
#[must_use]
pub fn max_decoding_message_size(mut self, limit: usize) -> Self {
self.max_decoding_message_size = Some(limit);
self
}
/// Limits the maximum size of an encoded message.
///
/// Default: `usize::MAX`
#[must_use]
pub fn max_encoding_message_size(mut self, limit: usize) -> Self {
self.max_encoding_message_size = Some(limit);
self
}
}
impl<T, B> tonic::codegen::Service<http::Request<B>> for ScannerControlServiceServer<T>
where
T: ScannerControlService,
B: Body + std::marker::Send + 'static,
B::Error: Into<StdError> + std::marker::Send + 'static,
{
type Response = http::Response<tonic::body::Body>;
type Error = std::convert::Infallible;
type Future = BoxFuture<Self::Response, Self::Error>;
fn poll_ready(&mut self, _cx: &mut Context<'_>) -> Poll<std::result::Result<(), Self::Error>> {
Poll::Ready(Ok(()))
}
fn call(&mut self, req: http::Request<B>) -> Self::Future {
match req.uri().path() {
"/node_service.ScannerControlService/ScannerScopedDirtyUsageAck" => {
#[allow(non_camel_case_types)]
struct ScannerScopedDirtyUsageAckSvc<T: ScannerControlService>(pub Arc<T>);
impl<T: ScannerControlService> tonic::server::UnaryService<super::ScannerScopedDirtyUsageAckRequest>
for ScannerScopedDirtyUsageAckSvc<T>
{
type Response = super::ScannerScopedDirtyUsageAckResponse;
type Future = BoxFuture<tonic::Response<Self::Response>, tonic::Status>;
fn call(&mut self, request: tonic::Request<super::ScannerScopedDirtyUsageAckRequest>) -> Self::Future {
let inner = Arc::clone(&self.0);
let fut = async move {
<T as ScannerControlService>::scanner_scoped_dirty_usage_ack(&inner, request).await
};
Box::pin(fut)
}
}
let accept_compression_encodings = self.accept_compression_encodings;
let send_compression_encodings = self.send_compression_encodings;
let max_decoding_message_size = self.max_decoding_message_size;
let max_encoding_message_size = self.max_encoding_message_size;
let inner = self.inner.clone();
let fut = async move {
let method = ScannerScopedDirtyUsageAckSvc(inner);
let codec = tonic_prost::ProstCodec::default();
let mut grpc = tonic::server::Grpc::new(codec)
.apply_compression_config(accept_compression_encodings, send_compression_encodings)
.apply_max_message_size_config(max_decoding_message_size, max_encoding_message_size);
let res = grpc.unary(method, req).await;
Ok(res)
};
Box::pin(fut)
}
_ => Box::pin(async move {
let mut response = http::Response::new(tonic::body::Body::default());
let headers = response.headers_mut();
headers.insert(tonic::Status::GRPC_STATUS, (tonic::Code::Unimplemented as i32).into());
headers.insert(http::header::CONTENT_TYPE, tonic::metadata::GRPC_CONTENT_TYPE);
Ok(response)
}),
}
}
}
impl<T> Clone for ScannerControlServiceServer<T> {
fn clone(&self) -> Self {
let inner = self.inner.clone();
Self {
inner,
accept_compression_encodings: self.accept_compression_encodings,
send_compression_encodings: self.send_compression_encodings,
max_decoding_message_size: self.max_decoding_message_size,
max_encoding_message_size: self.max_encoding_message_size,
}
}
}
/// Generated gRPC service name
pub const SERVICE_NAME: &str = "node_service.ScannerControlService";
impl<T> tonic::server::NamedService for ScannerControlServiceServer<T> {
const NAME: &'static str = SERVICE_NAME;
}
}
/// Generated client implementations.
pub mod heal_control_service_client {
#![allow(unused_variables, dead_code, missing_docs, clippy::wildcard_imports, clippy::let_unit_value)]
use tonic::codegen::http::Uri;
+2
View File
@@ -541,6 +541,8 @@ pub fn canonical_scanner_activity_v7_response_body(
Ok(body)
}
pub mod scoped_dirty_usage;
pub fn canonical_scanner_dirty_usage_snapshot_request_body(
request: &proto_gen::node_service::ScannerDirtyUsageSnapshotRequest,
) -> Result<Vec<u8>, std::num::TryFromIntError> {
+34
View File
@@ -903,6 +903,36 @@ message ScannerDirtyUsageSnapshotResponse {
bytes response_proof = 7;
}
// Receiver-only protocol. Producers must retain whole-cycle ACK until they
// have a durable per-bucket publication proof.
message ScannerScopedDirtyUsageEntry {
string bucket = 1;
bytes bucket_incarnation = 2;
uint64 generation = 3;
}
message ScannerScopedDirtyUsageAckRequest {
bytes challenge = 1;
uint32 protocol_version = 2;
string owner_id = 3;
string instance_id = 4;
// Only scope 1 (a complete bucket) is supported; zero is invalid.
uint32 scope = 5;
bool probe_only = 6;
repeated ScannerScopedDirtyUsageEntry entries = 7;
}
message ScannerScopedDirtyUsageAckResponse {
uint32 protocol_version = 1;
string owner_id = 2;
string instance_id = 3;
bool supported = 4;
uint32 max_entries = 5;
uint32 max_request_bytes = 6;
uint64 cleared = 7;
bytes response_proof = 8;
}
// A short-lived storage-owned read admission used only around a final
// scanner metadata publication. It is intentionally separate from the
// ScannerActivity observation wire so v6/v7 rolling compatibility remains
@@ -1245,6 +1275,10 @@ service NodeService {
rpc GetLiveEvents(GetLiveEventsRequest) returns (GetLiveEventsResponse) {}; // auth-policy: read-only
}
service ScannerControlService {
rpc ScannerScopedDirtyUsageAck(ScannerScopedDirtyUsageAckRequest) returns (ScannerScopedDirtyUsageAckResponse) {}; // auth-policy: body-bound
}
service HealControlService {
rpc HealControl(HealControlRequest) returns (HealControlResponse) {};
}
+213
View File
@@ -0,0 +1,213 @@
// Copyright 2024 RustFS Team
// Licensed under the Apache License, Version 2.0.
//! Bounded, authenticated receiver contract for per-bucket dirty acknowledgements.
use crate::CanonicalBodyBuilder;
use crate::proto_gen::node_service::{ScannerScopedDirtyUsageAckRequest, ScannerScopedDirtyUsageAckResponse};
use prost::Message;
pub const SCOPED_DIRTY_USAGE_PROTOCOL_VERSION: u32 = 1;
pub const SCOPED_DIRTY_USAGE_BUCKET_SCOPE: u32 = 1;
pub const SCOPED_DIRTY_USAGE_MAX_ENTRIES: u32 = 32;
pub const SCOPED_DIRTY_USAGE_MAX_REQUEST_BYTES: u32 = 8192;
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub enum ScopedDirtyUsageRequestError {
UnsupportedProtocol,
UnsupportedScope,
InvalidIdentity,
InvalidGeneration,
InvalidEntries,
TooLarge,
}
impl std::fmt::Display for ScopedDirtyUsageRequestError {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
f.write_str(match self {
Self::UnsupportedProtocol => "unsupported scoped dirty usage protocol",
Self::UnsupportedScope => "unsupported scoped dirty usage scope",
Self::InvalidIdentity => "invalid scoped dirty usage identity",
Self::InvalidGeneration => "invalid scoped dirty usage generation",
Self::InvalidEntries => "scoped dirty usage entries must be nonempty and strictly ordered",
Self::TooLarge => "scoped dirty usage request exceeds its budget",
})
}
}
impl std::error::Error for ScopedDirtyUsageRequestError {}
pub fn validate_scoped_dirty_usage_request(
request: &ScannerScopedDirtyUsageAckRequest,
) -> Result<(), ScopedDirtyUsageRequestError> {
use ScopedDirtyUsageRequestError as E;
if request.entries.len() > SCOPED_DIRTY_USAGE_MAX_ENTRIES as usize
|| request.encoded_len() > SCOPED_DIRTY_USAGE_MAX_REQUEST_BYTES as usize
{
return Err(E::TooLarge);
}
if request.protocol_version != SCOPED_DIRTY_USAGE_PROTOCOL_VERSION {
return Err(E::UnsupportedProtocol);
}
if request.scope != SCOPED_DIRTY_USAGE_BUCKET_SCOPE {
return Err(E::UnsupportedScope);
}
if request.challenge.len() != 16 || request.owner_id.len() != 36 || request.instance_id.len() != 32 {
return Err(E::InvalidIdentity);
}
if request.entries.is_empty() || request.entries.windows(2).any(|pair| pair[0].bucket >= pair[1].bucket) {
return Err(E::InvalidEntries);
}
for entry in &request.entries {
if entry.bucket.is_empty()
|| entry.bucket.len() > 63
|| entry.bucket_incarnation.len() != 16
|| entry.bucket_incarnation.iter().all(|byte| *byte == 0)
{
return Err(E::InvalidIdentity);
}
if entry.generation == 0 || entry.generation == u64::MAX {
return Err(E::InvalidGeneration);
}
}
Ok(())
}
pub fn canonical_scoped_dirty_usage_request(
request: &ScannerScopedDirtyUsageAckRequest,
) -> Result<Vec<u8>, ScopedDirtyUsageRequestError> {
validate_scoped_dirty_usage_request(request)?;
let mut body = CanonicalBodyBuilder::new(b"rustfs-scoped-dirty-usage-ack-request-v1\0");
let encode = |_: std::num::TryFromIntError| ScopedDirtyUsageRequestError::TooLarge;
body.push_bytes(request.challenge.as_ref()).map_err(encode)?;
body.push_u32(request.protocol_version);
body.push_str(&request.owner_id).map_err(encode)?;
body.push_str(&request.instance_id).map_err(encode)?;
body.push_u32(request.scope);
body.push_bool(request.probe_only);
body.push_count(request.entries.len()).map_err(encode)?;
for entry in &request.entries {
body.push_str(&entry.bucket).map_err(encode)?;
body.push_bytes(entry.bucket_incarnation.as_ref()).map_err(encode)?;
body.push_u64(entry.generation);
}
Ok(body.finish())
}
pub fn canonical_scoped_dirty_usage_response(
request_body: &[u8],
response: &ScannerScopedDirtyUsageAckResponse,
) -> Result<Vec<u8>, std::num::TryFromIntError> {
let mut body = CanonicalBodyBuilder::new(b"rustfs-scoped-dirty-usage-ack-response-v1\0");
body.push_bytes(request_body)?;
body.push_u32(response.protocol_version);
body.push_str(&response.owner_id)?;
body.push_str(&response.instance_id)?;
body.push_bool(response.supported);
body.push_u32(response.max_entries);
body.push_u32(response.max_request_bytes);
body.push_u64(response.cleared);
Ok(body.finish())
}
#[cfg(test)]
mod tests {
use super::*;
use crate::proto_gen::node_service::ScannerScopedDirtyUsageEntry;
fn request() -> ScannerScopedDirtyUsageAckRequest {
ScannerScopedDirtyUsageAckRequest {
challenge: vec![1; 16].into(),
protocol_version: 1,
owner_id: "11111111-1111-1111-1111-111111111111".into(),
instance_id: "a".repeat(32),
scope: 1,
probe_only: false,
entries: vec![ScannerScopedDirtyUsageEntry {
bucket: "photos".into(),
bucket_incarnation: vec![2; 16].into(),
generation: 8,
}],
}
}
#[test]
fn scoped_dirty_usage_binds_every_request_field() {
let base = request();
let baseline = canonical_scoped_dirty_usage_request(&base).expect("valid request");
for field in 0..9 {
let mut changed = base.clone();
match field {
0 => changed.challenge = vec![3; 16].into(),
1 => changed.protocol_version += 1,
2 => changed.owner_id = "22222222-2222-2222-2222-222222222222".into(),
3 => changed.instance_id = "b".repeat(32),
4 => changed.scope += 1,
5 => changed.probe_only = true,
6 => changed.entries[0].bucket = "videos".into(),
7 => changed.entries[0].bucket_incarnation = vec![3; 16].into(),
_ => changed.entries[0].generation += 1,
}
assert!(canonical_scoped_dirty_usage_request(&changed).map_or(true, |body| body != baseline));
}
}
#[test]
fn scoped_dirty_usage_binds_capability_and_ack_to_exact_request() {
let request = canonical_scoped_dirty_usage_request(&request()).expect("valid request");
let response = ScannerScopedDirtyUsageAckResponse {
protocol_version: 1,
owner_id: "owner".into(),
instance_id: "process".into(),
supported: true,
max_entries: 32,
max_request_bytes: 8192,
cleared: 1,
response_proof: vec![1; 32].into(),
};
let baseline = canonical_scoped_dirty_usage_response(&request, &response).expect("valid response");
for field in 0..7 {
let mut changed = response.clone();
match field {
0 => changed.protocol_version += 1,
1 => changed.owner_id.push('x'),
2 => changed.instance_id.push('x'),
3 => changed.supported = false,
4 => changed.max_entries += 1,
5 => changed.max_request_bytes += 1,
_ => changed.cleared += 1,
}
assert_ne!(
canonical_scoped_dirty_usage_response(&request, &changed).expect("response variant"),
baseline
);
}
assert_ne!(
canonical_scoped_dirty_usage_response(b"another request", &response).expect("request variant"),
baseline
);
}
#[test]
fn scoped_dirty_usage_rejects_overflow_unknown_and_duplicate_entries() {
let base = request();
let mut invalid = base.clone();
invalid.entries = vec![base.entries[0].clone(); SCOPED_DIRTY_USAGE_MAX_ENTRIES as usize + 1];
assert_eq!(validate_scoped_dirty_usage_request(&invalid), Err(ScopedDirtyUsageRequestError::TooLarge));
invalid = base.clone();
invalid.entries[0].bucket = "x".repeat(SCOPED_DIRTY_USAGE_MAX_REQUEST_BYTES as usize);
assert_eq!(validate_scoped_dirty_usage_request(&invalid), Err(ScopedDirtyUsageRequestError::TooLarge));
invalid = base.clone();
invalid.entries.push(base.entries[0].clone());
assert_eq!(
validate_scoped_dirty_usage_request(&invalid),
Err(ScopedDirtyUsageRequestError::InvalidEntries)
);
invalid = base;
invalid.entries[0].bucket_incarnation = vec![0; 16].into();
assert_eq!(
validate_scoped_dirty_usage_request(&invalid),
Err(ScopedDirtyUsageRequestError::InvalidIdentity)
);
}
}
+3 -2
View File
@@ -90,8 +90,9 @@ pub use scanner::{
};
pub use scanner_io::{
ScannerDirtyUsageAckError, ScannerDirtyUsageBucket, ScannerDirtyUsageSnapshot, ScannerDirtyUsageState,
acknowledge_dirty_usage_generation, clear_dirty_usage_bucket, record_dirty_usage_bucket, record_scanner_maintenance_change,
scanner_activity_epoch, scanner_dirty_usage_snapshot, scanner_dirty_usage_state, scanner_maintenance_generation,
acknowledge_dirty_usage_generation, acknowledge_scoped_dirty_usage, clear_dirty_usage_bucket, record_dirty_usage_bucket,
record_scanner_maintenance_change, scanner_activity_epoch, scanner_dirty_usage_snapshot, scanner_dirty_usage_state,
scanner_maintenance_generation,
};
pub use sleeper::{DynamicSleeper, SCANNER_IDLE_MODE, SCANNER_SLEEPER};
use std::sync::atomic::{AtomicU64, Ordering};
+3 -2
View File
@@ -883,8 +883,9 @@ pub(crate) use cache::{
};
pub use dirty_usage::{
ScannerDirtyUsageAckError, ScannerDirtyUsageBucket, ScannerDirtyUsageSnapshot, ScannerDirtyUsageState,
acknowledge_dirty_usage_generation, clear_dirty_usage_bucket, record_dirty_usage_bucket, record_scanner_maintenance_change,
scanner_activity_epoch, scanner_dirty_usage_snapshot, scanner_dirty_usage_state, scanner_maintenance_generation,
acknowledge_dirty_usage_generation, acknowledge_scoped_dirty_usage, clear_dirty_usage_bucket, record_dirty_usage_bucket,
record_scanner_maintenance_change, scanner_activity_epoch, scanner_dirty_usage_snapshot, scanner_dirty_usage_state,
scanner_maintenance_generation,
};
#[cfg(test)]
pub(crate) use dirty_usage::{clear_dirty_usage_buckets_for_tests, dirty_usage_buckets_for_tests};
@@ -52,6 +52,112 @@ pub enum ScannerDirtyUsageAckError {
ProcessChanged,
#[error("scanner dirty usage generation cannot be acknowledged")]
InvalidGeneration,
#[error("scanner dirty usage bucket incarnation fence is unavailable")]
IncarnationUnavailable,
}
/// A scoped ACK requires storage-owned lifecycle and incarnation fences.
/// Callers must only send ACKs backed by durable per-bucket publication.
pub fn acknowledge_scoped_dirty_usage(
instance_id: &str,
entries: &[(&crate::storage_api::EcstoreBucketMetadataMutationGuard, u64)],
probe_only: bool,
) -> std::result::Result<u64, ScannerDirtyUsageAckError> {
// Lock order: sorted bucket lifecycle/metadata fences (caller), then dirty map.
// No await or storage operation occurs while the dirty map is locked.
let (cleared, pending) = {
let mut dirty = dirty_usage_buckets();
let checked = entries
.iter()
.map(|(guard, generation)| {
guard
.checked_bucket_incarnation()
.map(|(bucket, _)| (bucket, *generation))
.map_err(|_| ScannerDirtyUsageAckError::IncarnationUnavailable)
})
.collect::<std::result::Result<Vec<_>, _>>()?;
let cleared = apply_scoped_dirty_usage_ack(
instance_id,
scanner_activity_epoch(),
DIRTY_USAGE_BUCKET_GENERATION.load(Ordering::Acquire),
&mut dirty,
&checked,
probe_only,
)?;
if cleared > 0 {
advance_generation(&DIRTY_USAGE_BUCKET_GENERATION);
}
(cleared, dirty.len())
};
if !probe_only {
global_metrics().record_scanner_dirty_usage_cycle_clear(usize_to_u64_saturated(cleared), usize_to_u64_saturated(pending));
}
Ok(usize_to_u64_saturated(cleared))
}
fn apply_scoped_dirty_usage_ack(
instance_id: &str,
current_instance: &str,
current_generation: u64,
dirty: &mut DirtyUsageBuckets,
entries: &[(&str, u64)],
probe_only: bool,
) -> std::result::Result<usize, ScannerDirtyUsageAckError> {
if instance_id != current_instance {
return Err(ScannerDirtyUsageAckError::ProcessChanged);
}
if current_generation == u64::MAX
|| entries
.iter()
.any(|(_, generation)| *generation == 0 || *generation == u64::MAX || *generation > current_generation)
{
return Err(ScannerDirtyUsageAckError::InvalidGeneration);
}
let mut cleared = 0;
if !probe_only {
for (bucket, generation) in entries {
if dirty.get(*bucket) == Some(generation) {
dirty.remove(*bucket);
cleared += 1;
}
}
}
Ok(cleared)
}
#[cfg(test)]
mod scoped_dirty_usage_tests {
use super::*;
#[test]
fn scoped_dirty_usage_preserves_uncovered_newer_and_replayed_generations() {
let mut dirty = HashMap::from([("hot".to_string(), 7), ("cold".to_string(), 8)]);
assert_eq!(apply_scoped_dirty_usage_ack("p", "p", 8, &mut dirty, &[("cold", 8)], true), Ok(0));
assert_eq!(dirty.len(), 2);
assert_eq!(apply_scoped_dirty_usage_ack("p", "p", 8, &mut dirty, &[("cold", 8)], false), Ok(1));
assert_eq!(dirty.get("hot"), Some(&7));
assert_eq!(apply_scoped_dirty_usage_ack("p", "p", 8, &mut dirty, &[("cold", 8)], false), Ok(0));
dirty.insert("cold".to_string(), 9);
assert_eq!(apply_scoped_dirty_usage_ack("p", "p", 9, &mut dirty, &[("cold", 8)], false), Ok(0));
assert_eq!(dirty.get("cold"), Some(&9));
}
#[test]
fn scoped_dirty_usage_rejects_restart_and_invalid_batch_before_clearing() {
let original = HashMap::from([("hot".to_string(), 7), ("cold".to_string(), 8)]);
let mut dirty = original.clone();
assert_eq!(
apply_scoped_dirty_usage_ack("old", "new", 8, &mut dirty, &[("cold", 8)], false),
Err(ScannerDirtyUsageAckError::ProcessChanged)
);
for generation in [0, 9, u64::MAX] {
assert_eq!(
apply_scoped_dirty_usage_ack("p", "p", 8, &mut dirty, &[("cold", 8), ("hot", generation)], false),
Err(ScannerDirtyUsageAckError::InvalidGeneration)
);
assert_eq!(dirty, original);
}
}
}
pub(super) fn dirty_usage_buckets() -> MutexGuard<'static, DirtyUsageBuckets> {
+2 -2
View File
@@ -38,8 +38,8 @@ pub(crate) use rustfs_ecstore::api::bucket::lifecycle::lifecycle::object_opts_fr
#[cfg(test)]
pub(crate) use rustfs_ecstore::api::bucket::metadata_sys::init_bucket_metadata_sys as ecstore_init_bucket_metadata_sys;
pub(crate) use rustfs_ecstore::api::bucket::metadata_sys::{
get_lifecycle_config as ecstore_get_lifecycle_config, get_object_lock_config as ecstore_get_object_lock_config,
get_replication_config as ecstore_get_replication_config,
BucketMetadataMutationGuard as EcstoreBucketMetadataMutationGuard, get_lifecycle_config as ecstore_get_lifecycle_config,
get_object_lock_config as ecstore_get_object_lock_config, get_replication_config as ecstore_get_replication_config,
};
pub(crate) use rustfs_ecstore::api::bucket::replication::{
ReplicateObjectInfo, ReplicationConfig as EcstoreReplicationConfig,