mirror of
https://github.com/rustfs/rustfs.git
synced 2026-08-11 07:36:53 +00:00
feat(heal): aggregate replacement recovery status (#5916)
Add a replacement recovery peer RPC so Admin v4 can distinguish definitive cluster proofs from unsupported, unavailable, or conflicting peer state without extending the existing background heal v3/v1 status protocol. Co-authored-by: heihutu <heihutu@gmail.com>
This commit is contained in:
@@ -848,6 +848,13 @@ impl NodeService for MinimalLockNodeService {
|
||||
Err(Status::unimplemented("lock-only test server"))
|
||||
}
|
||||
|
||||
async fn replacement_recovery_status(
|
||||
&self,
|
||||
_request: Request<rustfs_protos::proto_gen::node_service::ReplacementRecoveryStatusRequest>,
|
||||
) -> Result<Response<rustfs_protos::proto_gen::node_service::ReplacementRecoveryStatusResponse>, Status> {
|
||||
Err(Status::unimplemented("lock-only test server"))
|
||||
}
|
||||
|
||||
async fn get_metacache_listing(
|
||||
&self,
|
||||
_request: Request<rustfs_protos::proto_gen::node_service::GetMetacacheListingRequest>,
|
||||
|
||||
@@ -44,9 +44,9 @@ use rustfs_protos::proto_gen::node_service::{
|
||||
GetPartitionsRequest, GetProcInfoRequest, GetSeLinuxInfoRequest, GetSysConfigRequest, GetSysErrorsRequest,
|
||||
HealControlRequest, LoadBucketMetadataRequest, LoadGroupRequest, LoadPolicyMappingRequest, LoadPolicyRequest,
|
||||
LoadRebalanceMetaRequest, LoadServiceAccountRequest, LoadTransitionTierConfigRequest, LoadUserRequest,
|
||||
LocalStorageInfoRequest, Mss, ReloadPoolMetaRequest, ReloadSiteReplicationConfigRequest, ScannerActivityRequest,
|
||||
ScannerActivityResponse, ServerInfoRequest, SignalServiceRequest, SignalServiceResponse, StartDecommissionRequest,
|
||||
StartProfilingRequest, StopRebalanceRequest, TierMutationAbortRequest, TierMutationCommitRequest,
|
||||
LocalStorageInfoRequest, Mss, ReloadPoolMetaRequest, ReloadSiteReplicationConfigRequest, ReplacementRecoveryStatusRequest,
|
||||
ScannerActivityRequest, ScannerActivityResponse, ServerInfoRequest, SignalServiceRequest, SignalServiceResponse,
|
||||
StartDecommissionRequest, StartProfilingRequest, StopRebalanceRequest, TierMutationAbortRequest, TierMutationCommitRequest,
|
||||
TierMutationControlResponse, TierMutationPeerState, TierMutationPrepareRequest, node_service_client::NodeServiceClient,
|
||||
tier_mutation_control_service_client::TierMutationControlServiceClient,
|
||||
};
|
||||
@@ -78,6 +78,7 @@ pub const SERVICE_SIGNAL_RELOAD_DYNAMIC: u64 = 2;
|
||||
/// reload signal transport.
|
||||
pub const KMS_SIGNAL_SUBSYSTEM: &str = "kms";
|
||||
const BACKGROUND_HEAL_STATUS_MAX_MESSAGE_SIZE: usize = 64 * 1024;
|
||||
const REPLACEMENT_RECOVERY_STATUS_MAX_MESSAGE_SIZE: usize = 64 * 1024;
|
||||
const HEAL_CONTROL_FINGERPRINT_MAX_SIZE: usize = 256;
|
||||
const HEAL_CONTROL_PAYLOAD_MAX_SIZE: usize = 64 * 1024;
|
||||
const PEER_REST_RECOVERY_MAX_ATTEMPTS: u32 = 60;
|
||||
@@ -1083,6 +1084,38 @@ impl PeerRestClient {
|
||||
.await
|
||||
}
|
||||
|
||||
pub async fn replacement_recovery_status(&self) -> Result<Option<Vec<u8>>> {
|
||||
self.finalize_result(
|
||||
async {
|
||||
let mut client = self
|
||||
.get_client()
|
||||
.await?
|
||||
.max_decoding_message_size(REPLACEMENT_RECOVERY_STATUS_MAX_MESSAGE_SIZE);
|
||||
let response = match client
|
||||
.replacement_recovery_status(Request::new(ReplacementRecoveryStatusRequest::default()))
|
||||
.await
|
||||
{
|
||||
Ok(response) => response.into_inner(),
|
||||
Err(status) if status.code() == tonic::Code::Unimplemented => {
|
||||
// RUSTFS_COMPAT_TODO(replacement-recovery-status-v1): old peers cannot prove replacement completion during rolling upgrades. Remove after the minimum supported RustFS peer version implements ReplacementRecoveryStatus.
|
||||
return Ok(None);
|
||||
}
|
||||
Err(status) => return Err(status.into()),
|
||||
};
|
||||
if !response.success {
|
||||
return Err(Error::other(
|
||||
response
|
||||
.error_info
|
||||
.unwrap_or_else(|| "peer replacement recovery status failed without an error".to_string()),
|
||||
));
|
||||
}
|
||||
Ok(Some(response.recovery_status.to_vec()))
|
||||
}
|
||||
.await,
|
||||
)
|
||||
.await
|
||||
}
|
||||
|
||||
pub async fn prepare_tier_mutation(&self, mutation_id: Uuid, canonical_payload: Bytes) -> Result<PeerTierMutationOutcome> {
|
||||
self.tier_mutation_control(TierMutationRpcPhase::Prepare, mutation_id, canonical_payload)
|
||||
.await
|
||||
|
||||
@@ -1221,6 +1221,17 @@ pub struct BackgroundHealStatusResponse {
|
||||
#[prost(string, optional, tag = "3")]
|
||||
pub error_info: ::core::option::Option<::prost::alloc::string::String>,
|
||||
}
|
||||
#[derive(Clone, Copy, PartialEq, Eq, Hash, ::prost::Message)]
|
||||
pub struct ReplacementRecoveryStatusRequest {}
|
||||
#[derive(Clone, PartialEq, Eq, Hash, ::prost::Message)]
|
||||
pub struct ReplacementRecoveryStatusResponse {
|
||||
#[prost(bool, tag = "1")]
|
||||
pub success: bool,
|
||||
#[prost(bytes = "bytes", tag = "2")]
|
||||
pub recovery_status: ::prost::bytes::Bytes,
|
||||
#[prost(string, optional, tag = "3")]
|
||||
pub error_info: ::core::option::Option<::prost::alloc::string::String>,
|
||||
}
|
||||
#[derive(Clone, PartialEq, Eq, Hash, ::prost::Message)]
|
||||
pub struct HealControlRequest {
|
||||
#[prost(uint32, tag = "1")]
|
||||
@@ -2692,6 +2703,21 @@ pub mod node_service_client {
|
||||
.insert(GrpcMethod::new("node_service.NodeService", "BackgroundHealStatus"));
|
||||
self.inner.unary(req, path, codec).await
|
||||
}
|
||||
pub async fn replacement_recovery_status(
|
||||
&mut self,
|
||||
request: impl tonic::IntoRequest<super::ReplacementRecoveryStatusRequest>,
|
||||
) -> std::result::Result<tonic::Response<super::ReplacementRecoveryStatusResponse>, 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.NodeService/ReplacementRecoveryStatus");
|
||||
let mut req = request.into_request();
|
||||
req.extensions_mut()
|
||||
.insert(GrpcMethod::new("node_service.NodeService", "ReplacementRecoveryStatus"));
|
||||
self.inner.unary(req, path, codec).await
|
||||
}
|
||||
pub async fn get_metacache_listing(
|
||||
&mut self,
|
||||
request: impl tonic::IntoRequest<super::GetMetacacheListingRequest>,
|
||||
@@ -3179,6 +3205,10 @@ pub mod node_service_server {
|
||||
&self,
|
||||
request: tonic::Request<super::BackgroundHealStatusRequest>,
|
||||
) -> std::result::Result<tonic::Response<super::BackgroundHealStatusResponse>, tonic::Status>;
|
||||
async fn replacement_recovery_status(
|
||||
&self,
|
||||
request: tonic::Request<super::ReplacementRecoveryStatusRequest>,
|
||||
) -> std::result::Result<tonic::Response<super::ReplacementRecoveryStatusResponse>, tonic::Status>;
|
||||
async fn get_metacache_listing(
|
||||
&self,
|
||||
request: tonic::Request<super::GetMetacacheListingRequest>,
|
||||
@@ -5479,6 +5509,34 @@ pub mod node_service_server {
|
||||
};
|
||||
Box::pin(fut)
|
||||
}
|
||||
"/node_service.NodeService/ReplacementRecoveryStatus" => {
|
||||
#[allow(non_camel_case_types)]
|
||||
struct ReplacementRecoveryStatusSvc<T: NodeService>(pub Arc<T>);
|
||||
impl<T: NodeService> tonic::server::UnaryService<super::ReplacementRecoveryStatusRequest> for ReplacementRecoveryStatusSvc<T> {
|
||||
type Response = super::ReplacementRecoveryStatusResponse;
|
||||
type Future = BoxFuture<tonic::Response<Self::Response>, tonic::Status>;
|
||||
fn call(&mut self, request: tonic::Request<super::ReplacementRecoveryStatusRequest>) -> Self::Future {
|
||||
let inner = Arc::clone(&self.0);
|
||||
let fut = async move { <T as NodeService>::replacement_recovery_status(&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 = ReplacementRecoveryStatusSvc(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)
|
||||
}
|
||||
"/node_service.NodeService/GetMetacacheListing" => {
|
||||
#[allow(non_camel_case_types)]
|
||||
struct GetMetacacheListingSvc<T: NodeService>(pub Arc<T>);
|
||||
|
||||
@@ -851,6 +851,14 @@ message BackgroundHealStatusResponse {
|
||||
optional string error_info = 3;
|
||||
}
|
||||
|
||||
message ReplacementRecoveryStatusRequest {}
|
||||
|
||||
message ReplacementRecoveryStatusResponse {
|
||||
bool success = 1;
|
||||
bytes recovery_status = 2;
|
||||
optional string error_info = 3;
|
||||
}
|
||||
|
||||
message HealControlRequest {
|
||||
uint32 version = 1;
|
||||
string topology_fingerprint = 2;
|
||||
@@ -1084,6 +1092,7 @@ service NodeService {
|
||||
rpc SignalService(SignalServiceRequest) returns (SignalServiceResponse) {}; // auth-policy: body-bound
|
||||
rpc ScannerActivity(ScannerActivityRequest) returns (ScannerActivityResponse) {}; // auth-policy: body-bound
|
||||
rpc BackgroundHealStatus(BackgroundHealStatusRequest) returns (BackgroundHealStatusResponse) {}; // auth-policy: read-only
|
||||
rpc ReplacementRecoveryStatus(ReplacementRecoveryStatusRequest) returns (ReplacementRecoveryStatusResponse) {}; // auth-policy: read-only
|
||||
rpc GetMetacacheListing(GetMetacacheListingRequest) returns (GetMetacacheListingResponse) {}; // auth-policy: unimplemented
|
||||
rpc UpdateMetacacheListing(UpdateMetacacheListingRequest) returns (UpdateMetacacheListingResponse) {}; // auth-policy: unimplemented
|
||||
rpc ReloadPoolMeta(ReloadPoolMetaRequest) returns (ReloadPoolMetaResponse) {}; // auth-policy: body-bound
|
||||
|
||||
Reference in New Issue
Block a user