refactor(ecstore): name replication boundary contracts (#4134)

crates/ecstore/src/bucket/replication/README.md

@@ -23,15 +23,18 @@ and lifecycle/heal scheduling paths.
 |---|---|---|
 | `ReplicationObjectIO` | Object read/write primitives used by config, MRF, resync status, and multipart replication paths. | ECStore object API reader/writer types and storage-api object IO contracts are concentrated in `replication_storage_boundary.rs`. |
 | `ReplicationStorage` | Object read/write/delete, object walk, metadata update, and target object IO. | ECStore object API, storage-api contracts, and read option types are concentrated in `replication_storage_boundary.rs`. |
-| `ReplicationMetadataStore` | Replication config, MRF/resync state, target reset headers, and status persistence. | Metadata sys access and replication metadata path constants are concentrated in `replication_metadata_boundary.rs`; versioning sys and config storage imports remain. |
-| `ReplicationConfigStore` | Replication config persistence and config-derived labels used by target options. | Config read/save helpers and storage class labels are concentrated in `replication_config_store.rs`. |
+| `ReplicationMetadataStore` | Replication config, MRF/resync state, target reset headers, and status persistence. | Metadata sys access and replication metadata path constants are exposed through the contract type in `replication_metadata_boundary.rs`; versioning sys and config storage imports remain separate contracts. |
+| `ReplicationConfigStore` | Replication config persistence and config-derived labels used by target options. | Config read/save helpers and storage class labels are exposed through the contract type in `replication_config_store.rs`. |
 | `ReplicationFileMeta` | Replication status, decisions, MRF entries, resync decisions, and target reset helpers. | `rustfs_filemeta` replication contracts are concentrated in `replication_filemeta_boundary.rs`; `FileInfo` remains in the storage boundary for storage trait bindings and walk options. |
 | `ReplicationErrorBoundary` | ECStore error/result contracts and replication-specific error classifiers. | `crate::error` imports are concentrated in `replication_error_boundary.rs`. |
-| `ReplicationTargetStore` | Bucket target listing, target client lookup, target offline checks, target config types, and target operation option types. | Bucket target sys access, `BucketTargets`, and target operation types are concentrated in `replication_target_boundary.rs`. |
+| `ReplicationTargetStore` | Bucket target listing, target client lookup, target offline checks, target config types, and target operation option types. | Bucket target sys access, `BucketTargets`, and target operation types are exposed through the contract type in `replication_target_boundary.rs`. |
 | `ReplicationRuntime` | Worker pool, queue sizing, stats, bucket monitor, local node identity, cancellation, and admission state. | Direct runtime source/global access and shared replication pool/stat state; ECStore object store and bucket monitor implementation types stay behind local storage/bandwidth boundaries. |
 | `ReplicationBandwidthLimiter` | Target reader wrapping for replication bandwidth accounting and throttling. | Direct bucket bandwidth reader imports from resyncer paths. |
 | `ReplicationEventSink` | Notification and audit events for skipped, failed, pending, and completed replication operations. | Event notification service calls and local event host selection are concentrated in `replication_event_sink.rs`. |
-| `ReplicationTagFilter` | Decode object tag strings for rule and metadata replication decisions. | Direct bucket tagging helper imports from replication workers. |
+| `ReplicationVersioningStore` | Versioning state checks for object and delete replication decisions. | Bucket versioning sys access is exposed through the contract type in `replication_versioning_boundary.rs`. |
+| `ReplicationLockTiming` | Namespace lock timing for replication resync, object replication, and delete replication locks. | SetDisks lock timeout access is exposed through the contract type in `replication_lock_boundary.rs`. |
+| `ReplicationMsgpCodec` | MessagePack time encode/decode and unknown value skipping for persisted resync/MRF state. | Bucket MessagePack helpers are exposed through the contract type in `replication_msgp_boundary.rs`. |
+| `ReplicationTagFilter` | Decode object tag strings for rule and metadata replication decisions. | Bucket tagging helper access is exposed through the contract type in `replication_tagging_boundary.rs`. |
 | `ReplicationLifecycleBridge` | Lifecycle-originated delete and version-purge scheduling. | Direct coupling between lifecycle worker paths and replication delete scheduling. |
 
 ## Migration Rules
@@ -50,8 +53,10 @@ and lifecycle/heal scheduling paths.
    replication module, not `crate::bucket::replication::*` self paths.
 6. Keep runtime source access from importing ECStore object store or bucket
    monitor implementation types directly; use local boundary-owned aliases.
-7. Move at most one contract boundary per code-bearing PR and verify it with
-   focused replication tests before broad gates.
+7. Move at most one owner boundary per code-bearing PR and verify it with
+   focused replication tests before broad gates. Non-behavioral contract-shape
+   cleanup may batch already-established boundary wrappers when the owner and
+   call semantics do not change.
 
 ## First Code-Bearing Step
 

crates/ecstore/src/bucket/replication/config.rs

@@ -13,7 +13,7 @@
 // limitations under the License.
 
 use super::replication_filemeta_boundary::ReplicationType;
-use super::replication_tagging_boundary as tagging_boundary;
+use super::replication_tagging_boundary::ReplicationTagFilter;
 use super::rule::ReplicationRuleExt as _;
 use s3s::dto::DeleteMarkerReplicationStatus;
 use s3s::dto::DeleteReplicationStatus;
@@ -102,7 +102,7 @@ impl ReplicationConfigurationExt for ReplicationConfiguration {
             }
 
             if let Some(filter) = &rule.filter {
-                let object_tags = tagging_boundary::decode_tags_to_map(&obj.user_tags);
+                let object_tags = ReplicationTagFilter::decode_tags_to_map(&obj.user_tags);
                 if filter.test_tags(&object_tags) {
                     rules.push(rule.clone());
                 }

crates/ecstore/src/bucket/replication/replication_config_store.rs

@@ -14,20 +14,26 @@
 
 use super::replication_error_boundary::Result;
 use super::replication_storage_boundary::ReplicationObjectIO;
-use crate::config::com;
-pub(crate) use crate::config::storageclass::{RRS, STANDARD};
+use crate::config::{com, storageclass};
 use std::sync::Arc;
 
-pub(crate) async fn read<S>(api: Arc<S>, file: &str) -> Result<Vec<u8>>
-where
-    S: ReplicationObjectIO,
-{
-    com::read_config(api, file).await
-}
+pub(crate) struct ReplicationConfigStore;
 
-pub(crate) async fn save<S>(api: Arc<S>, file: &str, data: Vec<u8>) -> Result<()>
-where
-    S: ReplicationObjectIO,
-{
-    com::save_config(api, file, data).await
+impl ReplicationConfigStore {
+    pub(crate) const RRS: &'static str = storageclass::RRS;
+    pub(crate) const STANDARD: &'static str = storageclass::STANDARD;
+
+    pub(crate) async fn read<S>(api: Arc<S>, file: &str) -> Result<Vec<u8>>
+    where
+        S: ReplicationObjectIO,
+    {
+        com::read_config(api, file).await
+    }
+
+    pub(crate) async fn save<S>(api: Arc<S>, file: &str, data: Vec<u8>) -> Result<()>
+    where
+        S: ReplicationObjectIO,
+    {
+        com::save_config(api, file, data).await
+    }
 }

crates/ecstore/src/bucket/replication/replication_lock_boundary.rs

@@ -14,6 +14,10 @@
 
 use std::time::Duration;
 
-pub(crate) fn acquire_timeout() -> Duration {
-    crate::set_disk::get_lock_acquire_timeout()
+pub(crate) struct ReplicationLockTiming;
+
+impl ReplicationLockTiming {
+    pub(crate) fn acquire_timeout() -> Duration {
+        crate::set_disk::get_lock_acquire_timeout()
+    }
 }

crates/ecstore/src/bucket/replication/replication_metadata_boundary.rs

@@ -20,42 +20,47 @@ use time::OffsetDateTime;
 
 use super::replication_error_boundary::{Error, Result};
 
-pub(crate) const REPLICATION_DIR: &str = ".replication";
-pub(crate) const RESYNC_FILE_NAME: &str = "resync.bin";
-pub(crate) const MRF_REPLICATION_FILE: &str = "config/replication/mrf.bin";
+const REPLICATION_DIR: &str = ".replication";
+const RESYNC_FILE_NAME: &str = "resync.bin";
 
-pub(crate) async fn replication_config(bucket: &str) -> Result<(ReplicationConfiguration, OffsetDateTime)> {
-    metadata_sys::get_replication_config(bucket).await
-}
+pub(crate) struct ReplicationMetadataStore;
 
-pub(crate) async fn optional_replication_config(bucket: &str) -> Result<Option<ReplicationConfiguration>> {
-    let config = match replication_config(bucket).await {
-        Ok((config, _)) => Some(config),
-        Err(err) => {
-            if err != Error::ConfigNotFound {
-                return Err(err);
+impl ReplicationMetadataStore {
+    pub(crate) const MRF_REPLICATION_FILE: &'static str = "config/replication/mrf.bin";
+
+    pub(crate) async fn replication_config(bucket: &str) -> Result<(ReplicationConfiguration, OffsetDateTime)> {
+        metadata_sys::get_replication_config(bucket).await
+    }
+
+    pub(crate) async fn optional_replication_config(bucket: &str) -> Result<Option<ReplicationConfiguration>> {
+        let config = match Self::replication_config(bucket).await {
+            Ok((config, _)) => Some(config),
+            Err(err) => {
+                if err != Error::ConfigNotFound {
+                    return Err(err);
+                }
+                None
             }
-            None
-        }
-    };
-    Ok(config)
-}
+        };
+        Ok(config)
+    }
 
-pub(crate) fn rustfs_meta_bucket() -> &'static str {
-    RUSTFS_META_BUCKET
-}
+    pub(crate) fn rustfs_meta_bucket() -> &'static str {
+        RUSTFS_META_BUCKET
+    }
 
-pub(crate) fn resync_lock_key(bucket: &str, arn: &str) -> String {
-    format!("{REPLICATION_DIR}/{bucket}/{arn}")
-}
+    pub(crate) fn resync_lock_key(bucket: &str, arn: &str) -> String {
+        format!("{REPLICATION_DIR}/{bucket}/{arn}")
+    }
 
-pub(crate) fn bucket_resync_dir_path(bucket: &str) -> String {
-    path_join_buf(&[BUCKET_META_PREFIX, bucket, REPLICATION_DIR])
-}
+    pub(crate) fn bucket_resync_dir_path(bucket: &str) -> String {
+        path_join_buf(&[BUCKET_META_PREFIX, bucket, REPLICATION_DIR])
+    }
 
-pub(crate) fn bucket_resync_file_path(bucket: &str) -> String {
-    let resync_dir_path = bucket_resync_dir_path(bucket);
-    path_join_buf(&[&resync_dir_path, RESYNC_FILE_NAME])
+    pub(crate) fn bucket_resync_file_path(bucket: &str) -> String {
+        let resync_dir_path = Self::bucket_resync_dir_path(bucket);
+        path_join_buf(&[&resync_dir_path, RESYNC_FILE_NAME])
+    }
 }
 
 #[cfg(test)]
@@ -64,9 +69,18 @@ mod tests {
 
     #[test]
     fn replication_metadata_paths_match_existing_layout() {
-        assert_eq!(resync_lock_key("bucket-a", "arn-a"), ".replication/bucket-a/arn-a");
-        assert_eq!(bucket_resync_dir_path("bucket-a"), "buckets/bucket-a/.replication");
-        assert_eq!(bucket_resync_file_path("bucket-a"), "buckets/bucket-a/.replication/resync.bin");
-        assert_eq!(MRF_REPLICATION_FILE, "config/replication/mrf.bin");
+        assert_eq!(
+            ReplicationMetadataStore::resync_lock_key("bucket-a", "arn-a"),
+            ".replication/bucket-a/arn-a"
+        );
+        assert_eq!(
+            ReplicationMetadataStore::bucket_resync_dir_path("bucket-a"),
+            "buckets/bucket-a/.replication"
+        );
+        assert_eq!(
+            ReplicationMetadataStore::bucket_resync_file_path("bucket-a"),
+            "buckets/bucket-a/.replication/resync.bin"
+        );
+        assert_eq!(ReplicationMetadataStore::MRF_REPLICATION_FILE, "config/replication/mrf.bin");
     }
 }

crates/ecstore/src/bucket/replication/replication_msgp_boundary.rs

@@ -12,4 +12,25 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-pub(crate) use crate::bucket::msgp_decode::{read_msgp_ext8_time, skip_msgp_value, write_msgp_time};
+use std::io::{Read, Write};
+
+use time::OffsetDateTime;
+
+use super::replication_error_boundary::Result;
+use crate::bucket::msgp_decode;
+
+pub(crate) struct ReplicationMsgpCodec;
+
+impl ReplicationMsgpCodec {
+    pub(crate) fn read_ext8_time<R: Read>(rd: &mut R) -> Result<OffsetDateTime> {
+        msgp_decode::read_msgp_ext8_time(rd)
+    }
+
+    pub(crate) fn skip_value<R: Read>(rd: &mut R) -> Result<()> {
+        msgp_decode::skip_msgp_value(rd)
+    }
+
+    pub(crate) fn write_time<W: Write>(wr: &mut W, time: OffsetDateTime) -> Result<()> {
+        msgp_decode::write_msgp_time(wr, time)
+    }
+}

crates/ecstore/src/bucket/replication/replication_pool.rs

@@ -13,14 +13,14 @@
 // limitations under the License.
 
 use super::datatypes::ResyncStatusType;
-use super::replication_config_store as config_store;
+use super::replication_config_store::ReplicationConfigStore;
 use super::replication_error_boundary::Error as EcstoreError;
 use super::replication_filemeta_boundary::{
     MrfOpKind, MrfReplicateEntry, REPLICATE_EXISTING, REPLICATE_HEAL, REPLICATE_HEAL_DELETE, ReplicateDecision,
     ReplicateObjectInfo, ReplicatedTargetInfo, ReplicationStatusType, ReplicationType, ReplicationWorkerOperation,
     ResyncDecision, VersionPurgeStatusType, replication_statuses_map, version_purge_statuses_map,
 };
-use super::replication_metadata_boundary as metadata_boundary;
+use super::replication_metadata_boundary::ReplicationMetadataStore;
 use super::replication_resyncer::{
     BucketReplicationResyncStatus, DeletedObjectReplicationInfo, ReplicationConfig, ReplicationResyncer, ResyncOpts,
     TargetReplicationResyncStatus, decode_mrf_file, decode_resync_file, encode_mrf_file, get_heal_replicate_object_info,
@@ -28,7 +28,7 @@ use super::replication_resyncer::{
 };
 use super::replication_state::ReplicationStats;
 use super::replication_storage_boundary::{DeletedObject, ObjectInfo, ObjectOptions, ReplicationObjectIO, ReplicationStorage};
-use super::replication_target_boundary as target_boundary;
+use super::replication_target_boundary::ReplicationTargetStore;
 use super::runtime_boundary as runtime_sources;
 use lazy_static::lazy_static;
 use rustfs_utils::http::{SUFFIX_REPLICATION_TIMESTAMP, get_str};
@@ -788,7 +788,7 @@ impl<S: ReplicationStorage> ReplicationPool<S> {
         let storage = self.storage.clone();
 
         let handle = tokio::spawn(async move {
-            let data = match config_store::read(storage.clone(), metadata_boundary::MRF_REPLICATION_FILE).await {
+            let data = match ReplicationConfigStore::read(storage.clone(), ReplicationMetadataStore::MRF_REPLICATION_FILE).await {
                 Ok(d) => d,
                 Err(EcstoreError::ConfigNotFound) => return, // no file yet — normal on first start
                 Err(e) => {
@@ -812,9 +812,9 @@ impl<S: ReplicationStorage> ReplicationPool<S> {
                         "Failed to decode MRF recovery file — discarding corrupt data"
                     );
                     // Overwrite the corrupt file so we don't fail again on next restart.
-                    let _ = config_store::save(
+                    let _ = ReplicationConfigStore::save(
                         storage,
-                        metadata_boundary::MRF_REPLICATION_FILE,
+                        ReplicationMetadataStore::MRF_REPLICATION_FILE,
                         encode_mrf_file(&[]).unwrap_or_default(),
                     )
                     .await;
@@ -876,9 +876,12 @@ impl<S: ReplicationStorage> ReplicationPool<S> {
 
             // Clear AFTER all entries are processed so a crash mid-replay causes at-most-twice
             // delivery (idempotent) rather than entry loss.
-            if let Err(e) =
-                config_store::save(storage, metadata_boundary::MRF_REPLICATION_FILE, encode_mrf_file(&[]).unwrap_or_default())
-                    .await
+            if let Err(e) = ReplicationConfigStore::save(
+                storage,
+                ReplicationMetadataStore::MRF_REPLICATION_FILE,
+                encode_mrf_file(&[]).unwrap_or_default(),
+            )
+            .await
             {
                 warn!(
                     component = LOG_COMPONENT_ECSTORE,
@@ -1251,7 +1254,9 @@ impl<S: ReplicationStorage> ReplicationPool<S> {
 async fn flush_mrf_to_disk<S: ReplicationObjectIO>(entries: &[MrfReplicateEntry], storage: &Arc<S>) -> bool {
     match encode_mrf_file(entries) {
         Ok(data) => {
-            if let Err(e) = config_store::save(storage.clone(), metadata_boundary::MRF_REPLICATION_FILE, data).await {
+            if let Err(e) =
+                ReplicationConfigStore::save(storage.clone(), ReplicationMetadataStore::MRF_REPLICATION_FILE, data).await
+            {
                 warn!(
                     component = LOG_COMPONENT_ECSTORE,
                     subsystem = LOG_SUBSYSTEM_REPLICATION,
@@ -1283,9 +1288,9 @@ async fn load_bucket_resync_metadata<S: ReplicationObjectIO>(
 ) -> Result<BucketReplicationResyncStatus, EcstoreError> {
     let mut brs = BucketReplicationResyncStatus::new();
 
-    let resync_file_path = metadata_boundary::bucket_resync_file_path(bucket);
+    let resync_file_path = ReplicationMetadataStore::bucket_resync_file_path(bucket);
 
-    let data = match config_store::read(obj_api, &resync_file_path).await {
+    let data = match ReplicationConfigStore::read(obj_api, &resync_file_path).await {
         Ok(data) => data,
         Err(EcstoreError::ConfigNotFound) => return Ok(brs),
         Err(err) => return Err(err),
@@ -1484,7 +1489,7 @@ pub async fn queue_replication_heal(bucket: &str, oi: ObjectInfo, retry_count: u
         return;
     }
 
-    let rcfg = match metadata_boundary::replication_config(bucket).await {
+    let rcfg = match ReplicationMetadataStore::replication_config(bucket).await {
         Ok((config, _)) => config,
         Err(err) => {
             debug!(
@@ -1501,7 +1506,7 @@ pub async fn queue_replication_heal(bucket: &str, oi: ObjectInfo, retry_count: u
         }
     };
 
-    let tgts = match target_boundary::list_bucket_targets(bucket).await {
+    let tgts = match ReplicationTargetStore::list_bucket_targets(bucket).await {
         Ok(targets) => Some(targets),
         Err(err) => {
             debug!(

crates/ecstore/src/bucket/replication/replication_resyncer.rs

@@ -15,7 +15,7 @@
 use super::config::{ObjectOpts, ReplicationConfigurationExt as _};
 use super::datatypes::ResyncStatusType;
 use super::replication_bandwidth_boundary;
-use super::replication_config_store as config_store;
+use super::replication_config_store::ReplicationConfigStore;
 use super::replication_error_boundary::{Error, Result, is_err_object_not_found, is_err_version_not_found};
 use super::replication_event_sink::{EventArgs, send_event, send_local_event};
 use super::replication_filemeta_boundary::{
@@ -24,19 +24,19 @@ use super::replication_filemeta_boundary::{
     ReplicationType, ReplicationWorkerOperation, ResyncDecision, ResyncTargetDecision, VersionPurgeStatusType,
     get_replication_state, parse_replicate_decision, replication_statuses_map, target_reset_header, version_purge_statuses_map,
 };
-use super::replication_lock_boundary as lock_boundary;
-use super::replication_metadata_boundary as metadata_boundary;
-use super::replication_msgp_boundary::{read_msgp_ext8_time, skip_msgp_value, write_msgp_time};
+use super::replication_lock_boundary::ReplicationLockTiming;
+use super::replication_metadata_boundary::ReplicationMetadataStore;
+use super::replication_msgp_boundary::ReplicationMsgpCodec;
 use super::replication_storage_boundary::{
     AdvancedGetOptions, DeletedObject, EcstoreObjectOperations, HTTPRangeSpec, ObjectInfo, ObjectOptions, ObjectToDelete,
     ReplicationObjectIO, ReplicationStorage, StatObjectOptions, WalkOptions,
 };
-use super::replication_tagging_boundary as tagging_boundary;
-use super::replication_target_boundary;
+use super::replication_tagging_boundary::ReplicationTagFilter;
 use super::replication_target_boundary::{
-    AdvancedPutOptions, BucketTargets, PutObjectOptions, PutObjectPartOptions, RemoveObjectOptions, TargetClient,
+    AdvancedPutOptions, BucketTargets, PutObjectOptions, PutObjectPartOptions, RemoveObjectOptions, ReplicationTargetStore,
+    TargetClient,
 };
-use super::replication_versioning_boundary as versioning_boundary;
+use super::replication_versioning_boundary::ReplicationVersioningStore;
 use super::runtime_boundary as runtime_sources;
 use aws_sdk_s3::error::{ProvideErrorMetadata, SdkError};
 use aws_sdk_s3::operation::head_object::{HeadObjectError, HeadObjectOutput};
@@ -267,7 +267,7 @@ impl BucketReplicationResyncStatus {
         rmp::encode::write_str(&mut wr, "id")?;
         rmp::encode::write_i32(&mut wr, self.id)?;
         rmp::encode::write_str(&mut wr, "lu")?;
-        write_msgp_time(&mut wr, wire_time_or_default(self.last_update))?;
+        ReplicationMsgpCodec::write_time(&mut wr, wire_time_or_default(self.last_update))?;
         Ok(wr)
     }
 
@@ -300,7 +300,7 @@ impl BucketReplicationResyncStatus {
                 "lu" => {
                     out.last_update = normalize_wire_time(read_msgp_time_or_nil(&mut rd)?);
                 }
-                _ => skip_msgp_value(&mut rd)?,
+                _ => ReplicationMsgpCodec::skip_value(&mut rd)?,
             }
         }
         Ok(out)
@@ -385,13 +385,13 @@ impl TargetReplicationResyncStatus {
     fn marshal_wire_msg(&self, wr: &mut Vec<u8>) -> Result<()> {
         rmp::encode::write_map_len(wr, 11)?;
         rmp::encode::write_str(wr, "st")?;
-        write_msgp_time(wr, wire_time_or_default(self.start_time))?;
+        ReplicationMsgpCodec::write_time(wr, wire_time_or_default(self.start_time))?;
         rmp::encode::write_str(wr, "lst")?;
-        write_msgp_time(wr, wire_time_or_default(self.last_update))?;
+        ReplicationMsgpCodec::write_time(wr, wire_time_or_default(self.last_update))?;
         rmp::encode::write_str(wr, "id")?;
         rmp::encode::write_str(wr, &self.resync_id)?;
         rmp::encode::write_str(wr, "rdt")?;
-        write_msgp_time(wr, wire_time_or_default(self.resync_before_date))?;
+        ReplicationMsgpCodec::write_time(wr, wire_time_or_default(self.resync_before_date))?;
         rmp::encode::write_str(wr, "rst")?;
         rmp::encode::write_i32(wr, resync_status_to_i32(self.resync_status))?;
         rmp::encode::write_str(wr, "fs")?;
@@ -431,7 +431,7 @@ impl TargetReplicationResyncStatus {
                 "rrc" => out.replicated_count = rmp::decode::read_int(rd)?,
                 "bkt" => out.bucket = read_msgp_str(rd)?,
                 "obj" => out.object = read_msgp_str(rd)?,
-                _ => skip_msgp_value(rd)?,
+                _ => ReplicationMsgpCodec::skip_value(rd)?,
             }
         }
         Ok(out)
@@ -449,7 +449,7 @@ fn read_msgp_time_or_nil<R: Read>(rd: &mut R) -> Result<Option<OffsetDateTime>>
     let marker = rmp::decode::read_marker(rd).map_err(|e| Error::other(format!("{e:?}")))?;
     match marker {
         rmp::Marker::Null => Ok(None),
-        rmp::Marker::Ext8 => Ok(Some(read_msgp_ext8_time(rd)?)),
+        rmp::Marker::Ext8 => Ok(Some(ReplicationMsgpCodec::read_ext8_time(rd)?)),
         other => Err(Error::other(format!("expected time ext or nil, got marker: {other:?}"))),
     }
 }
@@ -737,9 +737,9 @@ impl ReplicationResyncer {
         // Acquire a cluster-wide leader lock for this (bucket, ARN) pair so that only
         // one node runs the resync scan at a time. Without this, every cluster node would
         // scan and replicate every object independently, causing N-fold duplicate traffic.
-        let resync_lock_key = metadata_boundary::resync_lock_key(&opts.bucket, &opts.arn);
+        let resync_lock_key = ReplicationMetadataStore::resync_lock_key(&opts.bucket, &opts.arn);
         let resync_ns_lock = match storage
-            .new_ns_lock(metadata_boundary::rustfs_meta_bucket(), &resync_lock_key)
+            .new_ns_lock(ReplicationMetadataStore::rustfs_meta_bucket(), &resync_lock_key)
             .await
         {
             Ok(l) => l,
@@ -757,7 +757,7 @@ impl ReplicationResyncer {
                 return;
             }
         };
-        let _resync_leader_guard = match resync_ns_lock.get_write_lock(lock_boundary::acquire_timeout()).await {
+        let _resync_leader_guard = match resync_ns_lock.get_write_lock(ReplicationLockTiming::acquire_timeout()).await {
             Ok(g) => g,
             Err(_) => {
                 debug!(
@@ -792,7 +792,7 @@ impl ReplicationResyncer {
             }
         };
 
-        let targets = match replication_target_boundary::list_bucket_targets(&opts.bucket).await {
+        let targets = match ReplicationTargetStore::list_bucket_targets(&opts.bucket).await {
             Ok(targets) => targets,
             Err(err) => {
                 debug!(
@@ -837,7 +837,7 @@ impl ReplicationResyncer {
             return;
         }
 
-        let Some(target_client) = replication_target_boundary::remote_target_client(&opts.bucket, &target_arns[0]).await else {
+        let Some(target_client) = ReplicationTargetStore::remote_target_client(&opts.bucket, &target_arns[0]).await else {
             error!(
                 event = EVENT_RESYNC_RUNTIME_SKIPPED,
                 component = LOG_COMPONENT_ECSTORE,
@@ -1164,8 +1164,8 @@ pub async fn get_heal_replicate_object_info(oi: &ObjectInfo, rcfg: &ReplicationC
             },
             &oi,
             &ObjectOptions {
-                versioned: versioning_boundary::prefix_enabled(&oi.bucket, &oi.name).await,
-                version_suspended: versioning_boundary::prefix_suspended(&oi.bucket, &oi.name).await,
+                versioned: ReplicationVersioningStore::prefix_enabled(&oi.bucket, &oi.name).await,
+                version_suspended: ReplicationVersioningStore::prefix_suspended(&oi.bucket, &oi.name).await,
                 ..Default::default()
             },
             None,
@@ -1228,14 +1228,14 @@ pub(crate) async fn save_resync_status<S: ReplicationObjectIO>(
 ) -> Result<()> {
     let data = encode_resync_file(status)?;
 
-    let config_file = metadata_boundary::bucket_resync_file_path(bucket);
-    config_store::save(api, &config_file, data).await?;
+    let config_file = ReplicationMetadataStore::bucket_resync_file_path(bucket);
+    ReplicationConfigStore::save(api, &config_file, data).await?;
 
     Ok(())
 }
 
 async fn get_replication_config(bucket: &str) -> Result<Option<ReplicationConfiguration>> {
-    metadata_boundary::optional_replication_config(bucket).await
+    ReplicationMetadataStore::optional_replication_config(bucket).await
 }
 
 #[derive(Debug, Clone, Default)]
@@ -1592,7 +1592,7 @@ pub async fn check_replicate_delete(
             continue;
         }
 
-        let tgt = replication_target_boundary::remote_target_client(bucket, &tgt_arn).await;
+        let tgt = ReplicationTargetStore::remote_target_client(bucket, &tgt_arn).await;
         // The target online status should not be used here while deciding
         // whether to replicate deletes as the target could be temporarily down
         let tgt_dsc = if let Some(tgt) = tgt {
@@ -1673,7 +1673,7 @@ pub async fn must_replicate(bucket: &str, object: &str, mopts: MustReplicateOpti
         return ReplicateDecision::default();
     }
 
-    if !versioning_boundary::prefix_enabled(bucket, object).await {
+    if !ReplicationVersioningStore::prefix_enabled(bucket, object).await {
         return ReplicateDecision::default();
     }
 
@@ -1717,7 +1717,7 @@ pub async fn must_replicate(bucket: &str, object: &str, mopts: MustReplicateOpti
     let mut dsc = ReplicateDecision::default();
 
     for arn in arns {
-        let cli = replication_target_boundary::remote_target_client(bucket, &arn).await;
+        let cli = ReplicationTargetStore::remote_target_client(bucket, &arn).await;
 
         let mut sopts = opts.clone();
         sopts.target_arn = arn.clone();
@@ -1807,8 +1807,9 @@ pub async fn replicate_delete<S: ReplicationStorage>(dobj: DeletedObjectReplicat
                 &dobj.delete_object.object_name,
                 &ObjectOptions {
                     version_id: Some(delete_marker_version_id.to_string()),
-                    versioned: versioning_boundary::prefix_enabled(&bucket, &dobj.delete_object.object_name).await,
-                    version_suspended: versioning_boundary::prefix_suspended(&bucket, &dobj.delete_object.object_name).await,
+                    versioned: ReplicationVersioningStore::prefix_enabled(&bucket, &dobj.delete_object.object_name).await,
+                    version_suspended: ReplicationVersioningStore::prefix_suspended(&bucket, &dobj.delete_object.object_name)
+                        .await,
                     ..Default::default()
                 },
             )
@@ -1928,7 +1929,7 @@ pub async fn replicate_delete<S: ReplicationStorage>(dobj: DeletedObjectReplicat
         }
     };
 
-    let _lock_guard = match ns_lock.get_write_lock(lock_boundary::acquire_timeout()).await {
+    let _lock_guard = match ns_lock.get_write_lock(ReplicationLockTiming::acquire_timeout()).await {
         Ok(lock_guard) => lock_guard,
         Err(e) => {
             debug!(
@@ -1979,7 +1980,7 @@ pub async fn replicate_delete<S: ReplicationStorage>(dobj: DeletedObjectReplicat
         }
 
         // Get the remote target client
-        let Some(tgt_client) = replication_target_boundary::remote_target_client(&bucket, &tgt_entry.arn).await else {
+        let Some(tgt_client) = ReplicationTargetStore::remote_target_client(&bucket, &tgt_entry.arn).await else {
             debug!(
                 event = EVENT_REPLICATION_DELETE_SKIPPED,
                 component = LOG_COMPONENT_ECSTORE,
@@ -2125,8 +2126,8 @@ pub async fn replicate_delete<S: ReplicationStorage>(dobj: DeletedObjectReplicat
                 version_id: version_id.map(|v| v.to_string()),
                 mod_time: dobj.delete_object.delete_marker_mtime,
                 delete_replication: Some(drs),
-                versioned: versioning_boundary::prefix_enabled(&bucket, &dobj.delete_object.object_name).await,
-                version_suspended: versioning_boundary::prefix_suspended(&bucket, &dobj.delete_object.object_name).await,
+                versioned: ReplicationVersioningStore::prefix_enabled(&bucket, &dobj.delete_object.object_name).await,
+                version_suspended: ReplicationVersioningStore::prefix_suspended(&bucket, &dobj.delete_object.object_name).await,
                 ..Default::default()
             },
         )
@@ -2180,8 +2181,8 @@ async fn source_delete_marker_missing<S: EcstoreObjectOperations>(
             object_name,
             &ObjectOptions {
                 version_id: Some(delete_marker_version_id.to_string()),
-                versioned: versioning_boundary::prefix_enabled(bucket, object_name).await,
-                version_suspended: versioning_boundary::prefix_suspended(bucket, object_name).await,
+                versioned: ReplicationVersioningStore::prefix_enabled(bucket, object_name).await,
+                version_suspended: ReplicationVersioningStore::prefix_suspended(bucket, object_name).await,
                 ..Default::default()
             },
         )
@@ -2204,7 +2205,7 @@ async fn replicate_delete_marker_purge_to_targets(bucket: &str, dobj: &DeletedOb
         if !dobj.target_arn.is_empty() && dobj.target_arn != tgt_entry.arn {
             continue;
         }
-        let Some(tgt_client) = replication_target_boundary::remote_target_client(bucket, &tgt_entry.arn).await else {
+        let Some(tgt_client) = ReplicationTargetStore::remote_target_client(bucket, &tgt_entry.arn).await else {
             continue;
         };
 
@@ -2311,7 +2312,7 @@ async fn replicate_force_delete_to_targets<S: ReplicationStorage>(dobj: &Deleted
         }
     };
 
-    let _lock_guard = match ns_lock.get_write_lock(lock_boundary::acquire_timeout()).await {
+    let _lock_guard = match ns_lock.get_write_lock(ReplicationLockTiming::acquire_timeout()).await {
         Ok(guard) => guard,
         Err(e) => {
             warn!(
@@ -2351,7 +2352,7 @@ async fn replicate_force_delete_to_targets<S: ReplicationStorage>(dobj: &Deleted
     let mut join_set = JoinSet::new();
 
     for arn in tgt_arns {
-        let Some(tgt_client) = replication_target_boundary::remote_target_client(bucket, &arn).await else {
+        let Some(tgt_client) = ReplicationTargetStore::remote_target_client(bucket, &arn).await else {
             debug!(
                 event = EVENT_REPLICATION_FORCE_DELETE_SKIPPED,
                 component = LOG_COMPONENT_ECSTORE,
@@ -2379,7 +2380,7 @@ async fn replicate_force_delete_to_targets<S: ReplicationStorage>(dobj: &Deleted
         let object_name = object_name.clone();
 
         join_set.spawn(async move {
-            if replication_target_boundary::target_is_offline(&tgt_client).await {
+            if ReplicationTargetStore::target_is_offline(&tgt_client).await {
                 error!(
                     event = EVENT_REPLICATION_FORCE_DELETE_SKIPPED,
                     component = LOG_COMPONENT_ECSTORE,
@@ -2505,7 +2506,7 @@ async fn replicate_delete_to_target(dobj: &DeletedObjectReplicationInfo, tgt_cli
         return rinfo;
     }
 
-    if replication_target_boundary::target_is_offline(&tgt_client).await {
+    if ReplicationTargetStore::target_is_offline(&tgt_client).await {
         if !is_version_purge {
             rinfo.replication_status = ReplicationStatusType::Failed;
         } else {
@@ -2703,7 +2704,7 @@ pub async fn replicate_object<S: ReplicationStorage>(roi: ReplicateObjectInfo, s
             return;
         }
     };
-    let _obj_lock_guard = match obj_ns_lock.get_write_lock(lock_boundary::acquire_timeout()).await {
+    let _obj_lock_guard = match obj_ns_lock.get_write_lock(ReplicationLockTiming::acquire_timeout()).await {
         Ok(g) => g,
         Err(e) => {
             debug!(
@@ -2730,7 +2731,7 @@ pub async fn replicate_object<S: ReplicationStorage>(roi: ReplicateObjectInfo, s
     let mut join_set = JoinSet::new();
 
     for arn in tgt_arns {
-        let Some(tgt_client) = replication_target_boundary::remote_target_client(&bucket, &arn).await else {
+        let Some(tgt_client) = ReplicationTargetStore::remote_target_client(&bucket, &arn).await else {
             debug!(
                 event = EVENT_RESYNC_RUNTIME_SKIPPED,
                 component = LOG_COMPONENT_ECSTORE,
@@ -2894,7 +2895,7 @@ impl ReplicateObjectInfoExt for ReplicateObjectInfo {
             return rinfo;
         }
 
-        if replication_target_boundary::target_is_offline(&tgt_client).await {
+        if ReplicationTargetStore::target_is_offline(&tgt_client).await {
             debug!(
                 event = EVENT_RESYNC_RUNTIME_SKIPPED,
                 component = LOG_COMPONENT_ECSTORE,
@@ -2915,8 +2916,8 @@ impl ReplicateObjectInfoExt for ReplicateObjectInfo {
             return rinfo;
         }
 
-        let versioned = versioning_boundary::prefix_enabled(&bucket, &object).await;
-        let version_suspended = versioning_boundary::prefix_suspended(&bucket, &object).await;
+        let versioned = ReplicationVersioningStore::prefix_enabled(&bucket, &object).await;
+        let version_suspended = ReplicationVersioningStore::prefix_suspended(&bucket, &object).await;
 
         let obj_opts = ObjectOptions {
             version_id: self.version_id.map(|v| v.to_string()),
@@ -3180,7 +3181,7 @@ impl ReplicateObjectInfoExt for ReplicateObjectInfo {
             ..Default::default()
         };
 
-        if replication_target_boundary::target_is_offline(&tgt_client).await {
+        if ReplicationTargetStore::target_is_offline(&tgt_client).await {
             debug!(
                 event = EVENT_RESYNC_RUNTIME_SKIPPED,
                 component = LOG_COMPONENT_ECSTORE,
@@ -3201,8 +3202,8 @@ impl ReplicateObjectInfoExt for ReplicateObjectInfo {
             return rinfo;
         }
 
-        let versioned = versioning_boundary::prefix_enabled(&bucket, &object).await;
-        let version_suspended = versioning_boundary::prefix_suspended(&bucket, &object).await;
+        let versioned = ReplicationVersioningStore::prefix_enabled(&bucket, &object).await;
+        let version_suspended = ReplicationVersioningStore::prefix_suspended(&bucket, &object).await;
 
         let obj_opts = ObjectOptions {
             version_id: self.version_id.map(|v| v.to_string()),
@@ -3726,7 +3727,7 @@ fn put_replication_opts(sc: &str, object_info: &ObjectInfo) -> Result<(PutObject
     // Handle storage class default
     let storage_class = if sc.is_empty() {
         let obj_sc = object_info.storage_class.as_deref().unwrap_or_default();
-        if obj_sc == config_store::STANDARD || obj_sc == config_store::RRS {
+        if obj_sc == ReplicationConfigStore::STANDARD || obj_sc == ReplicationConfigStore::RRS {
             obj_sc.to_string()
         } else {
             sc.to_string()
@@ -3753,7 +3754,7 @@ fn put_replication_opts(sc: &str, object_info: &ObjectInfo) -> Result<(PutObject
     };
 
     if !object_info.user_tags.is_empty() {
-        let tags = tagging_boundary::decode_tags_to_map(&object_info.user_tags);
+        let tags = ReplicationTagFilter::decode_tags_to_map(&object_info.user_tags);
 
         if !tags.is_empty() {
             put_op.user_tags = tags;
@@ -4029,8 +4030,9 @@ fn get_replication_action(oi1: &ObjectInfo, oi2: &HeadObjectOutput, op_type: Rep
         }
     }
 
-    let oi1_tags = tagging_boundary::decode_tags_to_map(&oi1.user_tags);
-    let oi2_tags = tagging_boundary::decode_tags_to_map(metadata.get(AMZ_OBJECT_TAGGING).cloned().unwrap_or_default().as_str());
+    let oi1_tags = ReplicationTagFilter::decode_tags_to_map(&oi1.user_tags);
+    let oi2_tags =
+        ReplicationTagFilter::decode_tags_to_map(metadata.get(AMZ_OBJECT_TAGGING).cloned().unwrap_or_default().as_str());
 
     if (oi2.tag_count.unwrap_or_default() > 0 && oi1_tags != oi2_tags)
         || oi2.tag_count.unwrap_or_default() != oi1_tags.len() as i32
@@ -4125,13 +4127,13 @@ mod tests {
         rmp::encode::write_str(&mut payload, "arn:replication::1:dest").expect("write arn");
         rmp::encode::write_map_len(&mut payload, 11).expect("write target");
         rmp::encode::write_str(&mut payload, "st").expect("write key");
-        write_msgp_time(&mut payload, start).expect("write time");
+        ReplicationMsgpCodec::write_time(&mut payload, start).expect("write time");
         rmp::encode::write_str(&mut payload, "lst").expect("write key");
-        write_msgp_time(&mut payload, last).expect("write time");
+        ReplicationMsgpCodec::write_time(&mut payload, last).expect("write time");
         rmp::encode::write_str(&mut payload, "id").expect("write key");
         rmp::encode::write_str(&mut payload, "resync-1").expect("write id");
         rmp::encode::write_str(&mut payload, "rdt").expect("write key");
-        write_msgp_time(&mut payload, before).expect("write time");
+        ReplicationMsgpCodec::write_time(&mut payload, before).expect("write time");
         rmp::encode::write_str(&mut payload, "rst").expect("write key");
         rmp::encode::write_i32(&mut payload, 3).expect("write status");
         rmp::encode::write_str(&mut payload, "fs").expect("write key");
@@ -4149,7 +4151,7 @@ mod tests {
         rmp::encode::write_str(&mut payload, "id").expect("write key");
         rmp::encode::write_i32(&mut payload, 42).expect("write id");
         rmp::encode::write_str(&mut payload, "lu").expect("write key");
-        write_msgp_time(&mut payload, bucket_last).expect("write lu");
+        ReplicationMsgpCodec::write_time(&mut payload, bucket_last).expect("write lu");
 
         let got = BucketReplicationResyncStatus::unmarshal_msg(&payload).expect("decode");
         assert_eq!(got.version, 1);

crates/ecstore/src/bucket/replication/replication_tagging_boundary.rs

@@ -14,17 +14,21 @@
 
 use std::collections::HashMap;
 
-pub(crate) fn decode_tags_to_map(tags: &str) -> HashMap<String, String> {
-    crate::bucket::tagging::decode_tags_to_map(tags)
+pub(crate) struct ReplicationTagFilter;
+
+impl ReplicationTagFilter {
+    pub(crate) fn decode_tags_to_map(tags: &str) -> HashMap<String, String> {
+        crate::bucket::tagging::decode_tags_to_map(tags)
+    }
 }
 
 #[cfg(test)]
 mod tests {
-    use super::decode_tags_to_map;
+    use super::ReplicationTagFilter;
 
     #[test]
     fn decode_tags_to_map_preserves_bucket_tagging_parser_behavior() {
-        let tags = decode_tags_to_map("env=prod&encoded=a%2Fb&=ignored");
+        let tags = ReplicationTagFilter::decode_tags_to_map("env=prod&encoded=a%2Fb&=ignored");
 
         assert_eq!(tags.get("env").map(String::as_str), Some("prod"));
         assert_eq!(tags.get("encoded").map(String::as_str), Some("a/b"));

crates/ecstore/src/bucket/replication/replication_target_boundary.rs

@@ -21,14 +21,18 @@ pub(crate) use crate::bucket::bucket_target_sys::{
 };
 pub(crate) use crate::bucket::target::BucketTargets;
 
-pub(crate) async fn list_bucket_targets(bucket: &str) -> Result<BucketTargets, BucketTargetError> {
-    BucketTargetSys::get().list_bucket_targets(bucket).await
-}
+pub(crate) struct ReplicationTargetStore;
 
-pub(crate) async fn remote_target_client(bucket: &str, arn: &str) -> Option<Arc<TargetClient>> {
-    BucketTargetSys::get().get_remote_target_client(bucket, arn).await
-}
+impl ReplicationTargetStore {
+    pub(crate) async fn list_bucket_targets(bucket: &str) -> Result<BucketTargets, BucketTargetError> {
+        BucketTargetSys::get().list_bucket_targets(bucket).await
+    }
 
-pub(crate) async fn target_is_offline(target_client: &TargetClient) -> bool {
-    BucketTargetSys::get().is_offline(&target_client.to_url()).await
+    pub(crate) async fn remote_target_client(bucket: &str, arn: &str) -> Option<Arc<TargetClient>> {
+        BucketTargetSys::get().get_remote_target_client(bucket, arn).await
+    }
+
+    pub(crate) async fn target_is_offline(target_client: &TargetClient) -> bool {
+        BucketTargetSys::get().is_offline(&target_client.to_url()).await
+    }
 }

crates/ecstore/src/bucket/replication/replication_versioning_boundary.rs

@@ -14,10 +14,14 @@
 
 use crate::bucket::versioning_sys::BucketVersioningSys;
 
-pub(crate) async fn prefix_enabled(bucket: &str, prefix: &str) -> bool {
-    BucketVersioningSys::prefix_enabled(bucket, prefix).await
-}
+pub(crate) struct ReplicationVersioningStore;
 
-pub(crate) async fn prefix_suspended(bucket: &str, prefix: &str) -> bool {
-    BucketVersioningSys::prefix_suspended(bucket, prefix).await
+impl ReplicationVersioningStore {
+    pub(crate) async fn prefix_enabled(bucket: &str, prefix: &str) -> bool {
+        BucketVersioningSys::prefix_enabled(bucket, prefix).await
+    }
+
+    pub(crate) async fn prefix_suspended(bucket: &str, prefix: &str) -> bool {
+        BucketVersioningSys::prefix_suspended(bucket, prefix).await
+    }
 }

docs/architecture/ecstore-module-split-plan.md

@@ -134,11 +134,11 @@ Required contracts before crate movement:
   `crates/ecstore/src/bucket/replication/replication_storage_boundary.rs`.
 - `ReplicationMetadataStore`: replication config, target reset headers,
   MRF/resync state, and status persistence. Metadata sys access and replication
-  metadata path constants are concentrated in
+  metadata path constants are exposed through the contract type in
   `crates/ecstore/src/bucket/replication/replication_metadata_boundary.rs`.
 - `ReplicationConfigStore`: replication config persistence and config-derived
   labels used by target options. Config read/save helpers and storage class
-  labels are concentrated in
+  labels are exposed through the contract type in
   `crates/ecstore/src/bucket/replication/replication_config_store.rs`.
 - `ReplicationFileMeta`: replication status, decisions, MRF entries, resync
   decisions, and target reset helpers. `rustfs_filemeta` replication contracts
@@ -153,13 +153,17 @@ Required contracts before crate movement:
 - `ReplicationTargetStore`: bucket target listing, target client lookup,
   target offline checks, target config types, and target operation option
   types. Bucket target sys access, `BucketTargets`, and target operation types
-  are concentrated in
+  are exposed through the contract type in
   `crates/ecstore/src/bucket/replication/replication_target_boundary.rs`.
 - `ReplicationRuntime`: pool, stats, worker admission, bucket monitor, local
   node identity, cancellation, and queue sizing. Concrete ECStore object store
   and bucket monitor types stay behind local storage/bandwidth boundaries.
 - `ReplicationBandwidthLimiter`: target reader wrapping for replication
   bandwidth accounting and throttling.
+- `ReplicationVersioningStore`, `ReplicationLockTiming`, `ReplicationMsgpCodec`,
+  and `ReplicationTagFilter`: smaller state/codec/filter contracts that keep
+  bucket versioning, SetDisks lock timing, MessagePack helpers, and bucket
+  tagging helper access behind local replication boundary types.
 - `ReplicationEventSink`: notification/audit events for skipped, failed, and
   completed replication operations, including local event host selection.
 - `ReplicationLifecycleBridge`: lifecycle-originated delete and version-purge