gsadmin/rustfs
1
0
Fork 0
mirror of https://github.com/rustfs/rustfs.git synced 2026-08-18 10:43:15 +00:00
Code Issues Packages Projects Releases Wiki Activity

feat(heal): aggregate replacement recovery status (#5916)

Browse Source 
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:
houseme
2026-08-10 13:03:27 +08:00
committed by GitHub
parent a8c15e90ec
commit 785ee719e7
9 changed files with 639 additions and 43 deletions
Download Patch File Download Diff File Expand all files Collapse all files
crates/ecstore/src/cluster/rpc/peer_rest_client.rs
+36 -3
View File
@@ -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