Merge branch 'main' into fix/b5-t1-a1-replication-deny-edit

The review fix made the bucket usecase read the site-replication peer set
through the admin handlers, an app->interface import the layer guard
rejects. The S3 handlers (interface) now read the peer set and pass it in,
so the usecase stays a pure function of its inputs; a state-read failure
still fails the edit closed, just one layer up.

crates/ecstore/src/api/mod.rs

@@ -196,15 +196,16 @@ pub mod bucket {
             ReplicationOperation, ReplicationPoolTrait, ReplicationPriority, ReplicationQueueAdmission, ReplicationScannerBridge,
             ReplicationState, ReplicationStats, ReplicationStatusType, ReplicationStorage, ReplicationTargetValidationError,
             ReplicationType, ResyncOpts, ResyncStatusType, RuntimeReplicationTargetBacklog, TargetReplicationResyncStatus,
-            VersionPurgeStatusType, XferStats, commit_force_delete_intent, complete_force_delete_intent,
-            delete_replication_state_from_config, delete_replication_version_id, get_global_replication_pool,
-            get_global_replication_stats, get_proxy_targets, init_background_replication,
-            invalid_replication_config_status_field, persist_force_delete_intent, read_durable_mrf_backlog,
-            replication_state_to_filemeta, replication_status_to_filemeta, replication_statuses_map, replication_target_arns,
-            resync_start_conflict_id, should_remove_replication_target, should_schedule_delete_replication,
-            should_use_existing_delete_replication_info, should_use_existing_delete_replication_source,
-            unsupported_replication_config_field, validate_replication_config_structure, validate_replication_config_target_arns,
-            version_purge_status_to_filemeta,
+            VersionPurgeStatusType, XferStats, assign_site_replication_rule_priorities, commit_force_delete_intent,
+            complete_force_delete_intent, delete_replication_state_from_config, delete_replication_version_id,
+            get_global_replication_pool, get_global_replication_stats, get_proxy_targets, init_background_replication,
+            invalid_replication_config_status_field, is_site_replication_rule, merge_incoming_replication_config,
+            merge_user_replication_config, persist_force_delete_intent, read_durable_mrf_backlog, replication_state_to_filemeta,
+            replication_status_to_filemeta, replication_statuses_map, replication_target_arn_deployment_id,
+            replication_target_arns, resync_start_conflict_id, should_remove_replication_target,
+            should_schedule_delete_replication, should_use_existing_delete_replication_info,
+            should_use_existing_delete_replication_source, unsupported_replication_config_field,
+            validate_replication_config_structure, validate_replication_config_target_arns, version_purge_status_to_filemeta,
         };
     }
 

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

@@ -47,8 +47,10 @@ pub use replication_config_boundary::{
     ObjectOpts, REMOTE_TARGET_CAPABILITY_CONTRACT_VERSION, REMOTE_TARGET_UNSUPPORTED_FIELDS, REMOTE_TARGET_WRITABLE_FIELDS,
     REPLICATION_CAPABILITY_CONTRACT_VERSION, REPLICATION_READ_ONLY_HISTORICAL_FIELDS, REPLICATION_WRITABLE_FIELDS,
     ReplicationConfigStructureError, ReplicationConfigurationExt, ReplicationTargetValidationError,
-    invalid_replication_config_status_field, replication_target_arns, should_remove_replication_target,
-    unsupported_replication_config_field, validate_replication_config_structure, validate_replication_config_target_arns,
+    assign_site_replication_rule_priorities, invalid_replication_config_status_field, is_site_replication_rule,
+    merge_incoming_replication_config, merge_user_replication_config, replication_target_arn_deployment_id,
+    replication_target_arns, should_remove_replication_target, unsupported_replication_config_field,
+    validate_replication_config_structure, validate_replication_config_target_arns,
 };
 pub(crate) use replication_filemeta_boundary::version_purge_statuses_map;
 pub use replication_filemeta_boundary::{

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

@@ -16,6 +16,8 @@ pub use rustfs_replication::{
     ObjectOpts, REMOTE_TARGET_CAPABILITY_CONTRACT_VERSION, REMOTE_TARGET_UNSUPPORTED_FIELDS, REMOTE_TARGET_WRITABLE_FIELDS,
     REPLICATION_CAPABILITY_CONTRACT_VERSION, REPLICATION_READ_ONLY_HISTORICAL_FIELDS, REPLICATION_WRITABLE_FIELDS,
     ReplicationConfigStructureError, ReplicationConfigurationExt, ReplicationRuleExt, ReplicationTargetValidationError,
-    invalid_replication_config_status_field, replication_target_arns, should_remove_replication_target,
-    unsupported_replication_config_field, validate_replication_config_structure, validate_replication_config_target_arns,
+    assign_site_replication_rule_priorities, invalid_replication_config_status_field, is_site_replication_rule,
+    merge_incoming_replication_config, merge_user_replication_config, replication_target_arn_deployment_id,
+    replication_target_arns, should_remove_replication_target, unsupported_replication_config_field,
+    validate_replication_config_structure, validate_replication_config_target_arns,
 };

crates/ecstore/src/cluster/rpc/peer_rest_client.rs

@@ -1089,9 +1089,7 @@ impl PeerRestClient {
                     .await?
                     .max_decoding_message_size(BACKGROUND_HEAL_STATUS_MAX_MESSAGE_SIZE);
                 let response = match client
-                    .background_heal_status(Request::new(BackgroundHealStatusRequest {
-                        protocol_version: rustfs_protos::BACKGROUND_HEAL_STATUS_PROTOCOL_VERSION,
-                    }))
+                    .background_heal_status(Request::new(BackgroundHealStatusRequest::default()))
                     .await
                 {
                     Ok(response) => response.into_inner(),

crates/ecstore/src/io_support/bitrot.rs

@@ -784,24 +784,6 @@ pub(crate) fn create_deferred_bitrot_reader_with_stripe_handle(
 ///
 /// # Returns
 /// A Result containing the BitrotWriterWrapper or an error
-/// Size hint handed to `DiskAPI::create_file` for a bitrot-wrapped shard.
-///
-/// A known length is grown by one checksum per shard so the on-disk file size
-/// matches what the bitrot writer emits. A negative length is the
-/// unknown-size sentinel (`HashReader::SIZE_PRESERVE_LAYER`, used by SSE and
-/// compression) and must be preserved: `RemoteDisk::create_file` forwards it
-/// in the `put_file_stream` query, and the receiver only treats `size > 0` as
-/// a fixed body length when locating the authenticated trailer. Clamping it
-/// to `0` would claim an empty body and misframe the stream. `0` stays `0`
-/// because a genuinely empty object still means an empty body.
-fn bitrot_create_file_size(length: i64, shard_size: usize, checksum_algo: &HashAlgorithm) -> i64 {
-    if length <= 0 {
-        return length;
-    }
-    let length = length as usize;
-    (length.div_ceil(shard_size) * checksum_algo.size() + length) as i64
-}
-
 pub async fn create_bitrot_writer(
     is_inline_buffer: bool,
     disk: Option<&DiskStore>,
@@ -814,7 +796,12 @@ pub async fn create_bitrot_writer(
     let writer = if is_inline_buffer {
         CustomWriter::new_inline_buffer()
     } else if let Some(disk) = disk {
-        let length = bitrot_create_file_size(length, shard_size, &checksum_algo);
+        let length = if length > 0 {
+            let length = length as usize;
+            (length.div_ceil(shard_size) * checksum_algo.size() + length) as i64
+        } else {
+            0
+        };
 
         let file = disk.create_file("", volume, path, length).await?;
         #[cfg(feature = "hotpath")]
@@ -833,25 +820,6 @@ mod tests {
     use rustfs_rio::ChunkReader;
     use std::collections::VecDeque;
 
-    #[test]
-    fn bitrot_create_file_size_grows_known_length_by_checksums() {
-        // 10 bytes over 4-byte shards = 3 shards, each followed by a 32-byte hash.
-        assert_eq!(bitrot_create_file_size(10, 4, &HashAlgorithm::HighwayHash256), 10 + 3 * 32);
-        assert_eq!(bitrot_create_file_size(10, 4, &HashAlgorithm::None), 10);
-    }
-
-    #[test]
-    fn bitrot_create_file_size_keeps_empty_and_unknown_distinct() {
-        assert_eq!(bitrot_create_file_size(0, 4, &HashAlgorithm::HighwayHash256), 0);
-        // SSE/compression streams advertise SIZE_PRESERVE_LAYER (-1); the remote
-        // put_file_stream receiver relies on a non-positive size to parse the auth
-        // trailer from the stream tail, so the sentinel must survive untouched.
-        assert_eq!(
-            bitrot_create_file_size(rustfs_rio::HashReader::SIZE_PRESERVE_LAYER, 4, &HashAlgorithm::HighwayHash256),
-            rustfs_rio::HashReader::SIZE_PRESERVE_LAYER
-        );
-    }
-
     struct TestChunkReader {
         chunks: VecDeque<Bytes>,
     }

crates/heal/src/heal/erasure_healer.rs

@@ -13,10 +13,10 @@
 // limitations under the License.
 
 use crate::heal::{
-    progress::{HealProgress, add_bytes, increment_counter},
+    progress::HealProgress,
     resume::{
-        CheckpointManager, CheckpointObjectOutcome, CheckpointObjectOutcomeRecord, ReplacementTargetIdentity, ResumeManager,
-        ResumeUtils, compose_key, replacement_target_identities_match,
+        CheckpointManager, ReplacementTargetIdentity, ResumeManager, ResumeUtils, compose_key,
+        replacement_target_identities_match,
     },
     storage::{HealStorageAPI, next_heal_listing_token},
     task::{demote_to_debug_when, is_missing_object_dir_heal_result, take_failure_log_sample},
@@ -410,9 +410,6 @@ impl ErasureSetHealer {
                 && state.successful_objects == 0
                 && state.failed_objects == 0
                 && state.skipped_objects == 0
-                && state.skipped_new_versions == 0
-                && state.skipped_ilm_expired == 0
-                && state.processed_bytes == 0
             {
                 // schedule_retry persists the authoritative resume reset before
                 // resetting the checkpoint. Reapply the checkpoint reset after
@@ -477,23 +474,6 @@ impl ErasureSetHealer {
 
         // 2. initialize progress
         self.initialize_progress(buckets, &state).await;
-        let (baseline_known, baseline_count, baseline_size, baseline_generation) = {
-            let baseline = self.progress.read().await;
-            (
-                baseline.baseline_known,
-                baseline.objects_total_count,
-                baseline.objects_total_size,
-                baseline.baseline_generation,
-            )
-        };
-        if baseline_known {
-            resume_manager
-                .set_progress_baseline(baseline_count, baseline_size, baseline_generation)
-                .await?;
-            checkpoint_manager
-                .set_progress_baseline(baseline_count, baseline_size, baseline_generation)
-                .await?;
-        }
 
         // 3. continue from checkpoint
         let current_bucket_index = checkpoint.current_bucket_index;
@@ -503,66 +483,12 @@ impl ErasureSetHealer {
         let mut successful_objects = state.successful_objects;
         let mut failed_objects = state.failed_objects;
         let mut skipped_objects = state.skipped_objects;
-        let checkpoint_has_progress = checkpoint.baseline_known
-            || checkpoint.successful_objects > 0
-            || checkpoint.failed_object_count > 0
-            || checkpoint.skipped_object_count > 0
-            || checkpoint.skipped_new_versions > 0
-            || checkpoint.skipped_ilm_expired > 0
-            || checkpoint.processed_bytes > 0
-            || checkpoint.total_objects > 0
-            || checkpoint.total_bytes > 0
-            || checkpoint.baseline_generation.is_some()
-            || checkpoint.counter_unknown;
-        let checkpoint_generation_mismatch = checkpoint.baseline_known && checkpoint.baseline_generation != baseline_generation;
-        let mut restored_counter_unknown = state.counter_unknown || checkpoint.counter_unknown;
-        if checkpoint_has_progress {
-            successful_objects = checkpoint.successful_objects;
-            failed_objects = checkpoint.failed_object_count;
-            skipped_objects = checkpoint.skipped_object_count;
-            let restored_processed_objects = successful_objects
-                .checked_add(failed_objects)
-                .and_then(|value| value.checked_add(skipped_objects))
-                .and_then(|value| value.checked_add(checkpoint.skipped_new_versions))
-                .and_then(|value| value.checked_add(checkpoint.skipped_ilm_expired));
-            let checkpoint_counter_overflow = restored_processed_objects.is_none();
-            restored_counter_unknown |= checkpoint_counter_overflow;
-            processed_objects = restored_processed_objects.unwrap_or(u64::MAX);
-            let mut progress = self.progress.write().await;
-            progress.objects_scanned = processed_objects;
-            progress.objects_healed = successful_objects;
-            progress.objects_failed = failed_objects;
-            progress.skipped_objects = skipped_objects;
-            progress.skipped_new_versions = checkpoint.skipped_new_versions;
-            progress.skipped_ilm_expired = checkpoint.skipped_ilm_expired;
-            if checkpoint.baseline_known && !checkpoint_generation_mismatch {
-                progress.objects_total_count = checkpoint.total_objects;
-                progress.objects_total_size = checkpoint.total_bytes;
-                progress.baseline_generation = checkpoint.baseline_generation;
-                progress.baseline_known = true;
-            }
-            progress.bytes_processed = checkpoint.processed_bytes;
-            progress.counter_unknown = state.counter_unknown || checkpoint.counter_unknown;
-            progress.refresh_progress_percentage();
-            if checkpoint_generation_mismatch || checkpoint_counter_overflow || progress.counter_unknown {
-                progress.mark_unknown();
-            }
-        }
-        if checkpoint_generation_mismatch {
-            restored_counter_unknown = true;
-        }
-        if restored_counter_unknown {
-            checkpoint_manager.mark_counter_unknown().await?;
-            resume_manager.mark_counter_unknown().await?;
-        }
         let mut failed_buckets = 0u64;
 
         // 4. process remaining buckets
         for (bucket_idx, bucket) in buckets.iter().enumerate().skip(current_bucket_index) {
             // check if completed
             if state.completed_buckets.contains(bucket) {
-                checkpoint_manager.complete_bucket(bucket_idx.saturating_add(1)).await?;
-                current_object_index = 0;
                 continue;
             }
 
@@ -590,42 +516,13 @@ impl ErasureSetHealer {
                 return bucket_result;
             }
 
-            // update progress
-            let progress_snapshot = self.progress.read().await;
-            let bytes_processed = progress_snapshot.bytes_processed;
-            let skipped_new_versions = progress_snapshot.skipped_new_versions;
-            let skipped_ilm_expired = progress_snapshot.skipped_ilm_expired;
-            let counter_unknown = progress_snapshot.counter_unknown;
-            drop(progress_snapshot);
-            // The checkpoint is the recovery authority for object progress.
-            // Publish its counters and fence before the resume summary so a
-            // crash between the two stores cannot make recovery select newer
-            // summary bytes with an older checkpoint ledger.
-            if counter_unknown {
-                checkpoint_manager.mark_counter_unknown().await?;
-            }
-            checkpoint_manager
-                .update_progress(successful_objects, failed_objects, skipped_objects, bytes_processed)
-                .await?;
-            checkpoint_manager
-                .set_skipped_version_counts(skipped_new_versions, skipped_ilm_expired)
-                .await?;
+            // update checkpoint position
             checkpoint_manager.update_position(bucket_idx, current_object_index).await?;
+
+            // update progress
             resume_manager
-                .update_progress_with_bytes(
-                    processed_objects,
-                    successful_objects,
-                    failed_objects,
-                    skipped_objects,
-                    bytes_processed,
-                )
+                .update_progress(processed_objects, successful_objects, failed_objects, skipped_objects)
                 .await?;
-            resume_manager
-                .set_skipped_version_counts(skipped_new_versions, skipped_ilm_expired)
-                .await?;
-            if counter_unknown {
-                resume_manager.mark_counter_unknown().await?;
-            }
 
             // check cancel status
             if self.cancel_token.is_cancelled() {
@@ -645,7 +542,6 @@ impl ErasureSetHealer {
             match bucket_result {
                 Ok(_) => {
                     resume_manager.complete_bucket(bucket).await?;
-                    checkpoint_manager.complete_bucket(bucket_idx.saturating_add(1)).await?;
                     debug!(
                         target: "rustfs::heal::erasure_healer",
                         event = EVENT_HEAL_ERASURE_BUCKET_STATE,
@@ -671,9 +567,7 @@ impl ErasureSetHealer {
                         error = %e,
                         "Erasure set bucket heal failed"
                     );
-                    // A single durable cursor and ledger cannot safely preserve
-                    // this bucket while processing a later one.
-                    break;
+                    // continue to next bucket, do not interrupt the whole process
                 }
             }
 
@@ -881,49 +775,20 @@ impl ErasureSetHealer {
 
                 // Per-version dedup identity — the single canonical key.
                 let key = compose_key(&item.name, item.version_id.as_deref());
-                if checkpoint.processed_objects.contains(&key)
-                    || checkpoint.failed_objects.contains(&key)
-                    || checkpoint.skipped_objects.contains(&key)
-                {
+                if checkpoint.processed_objects.contains(&key) || checkpoint.skipped_objects.contains(&key) {
                     continue;
                 }
 
                 if should_skip_new_version(item.mod_time_unix_nanos, started_at_secs) {
-                    let counter_ok = increment_counter(processed_objects);
+                    checkpoint_manager.add_processed_object(key).await?;
+                    *processed_objects = processed_objects.saturating_add(1);
                     completed_in_page = completed_in_page.saturating_add(1);
                     counter!("rustfs_heal_skipped_new_versions_total").increment(1);
-                    let (outcome_record, counter_unknown) = {
+                    {
                         let mut progress = self.progress.write().await;
                         progress.record_skipped_new_version();
                         progress.set_current_object(Some(format!("skipped_new: {bucket}/{}", item.name)));
-                        progress.update_object_progress(
-                            *processed_objects,
-                            *successful_objects,
-                            *failed_objects,
-                            *skipped_objects,
-                            bytes_processed,
-                        );
-                        if !counter_ok {
-                            progress.mark_unknown();
-                        }
-                        (
-                            CheckpointObjectOutcomeRecord {
-                                object: key,
-                                outcome: CheckpointObjectOutcome::Processed,
-                                successful: progress.objects_healed,
-                                failed: progress.objects_failed,
-                                skipped: progress.skipped_objects,
-                                bytes: progress.bytes_processed,
-                                skipped_new_versions: progress.skipped_new_versions,
-                                skipped_ilm_expired: progress.skipped_ilm_expired,
-                                counter_unknown: progress.counter_unknown,
-                            },
-                            progress.counter_unknown,
-                        )
-                    };
-                    checkpoint_manager.record_object_outcome(outcome_record).await?;
-                    if counter_unknown {
-                        resume_manager.mark_counter_unknown().await?;
+                        progress.update_progress(*processed_objects, *successful_objects, *failed_objects, bytes_processed);
                     }
                     debug!(
                         target: "rustfs::heal::erasure_healer",
@@ -955,41 +820,15 @@ impl ErasureSetHealer {
                         )
                         .await?
                 {
-                    let counter_ok = increment_counter(processed_objects);
+                    checkpoint_manager.add_processed_object(key).await?;
+                    *processed_objects = processed_objects.saturating_add(1);
                     completed_in_page = completed_in_page.saturating_add(1);
                     counter!("rustfs_heal_skipped_ilm_expired_total").increment(1);
-                    let (outcome_record, counter_unknown) = {
+                    {
                         let mut progress = self.progress.write().await;
                         progress.record_skipped_ilm_expired();
                         progress.set_current_object(Some(format!("skipped_ilm: {bucket}/{}", item.name)));
-                        progress.update_object_progress(
-                            *processed_objects,
-                            *successful_objects,
-                            *failed_objects,
-                            *skipped_objects,
-                            bytes_processed,
-                        );
-                        if !counter_ok {
-                            progress.mark_unknown();
-                        }
-                        (
-                            CheckpointObjectOutcomeRecord {
-                                object: key,
-                                outcome: CheckpointObjectOutcome::Processed,
-                                successful: progress.objects_healed,
-                                failed: progress.objects_failed,
-                                skipped: progress.skipped_objects,
-                                bytes: progress.bytes_processed,
-                                skipped_new_versions: progress.skipped_new_versions,
-                                skipped_ilm_expired: progress.skipped_ilm_expired,
-                                counter_unknown: progress.counter_unknown,
-                            },
-                            progress.counter_unknown,
-                        )
-                    };
-                    checkpoint_manager.record_object_outcome(outcome_record).await?;
-                    if counter_unknown {
-                        resume_manager.mark_counter_unknown().await?;
+                        progress.update_progress(*processed_objects, *successful_objects, *failed_objects, bytes_processed);
                     }
                     debug!(
                         target: "rustfs::heal::erasure_healer",
@@ -1115,11 +954,11 @@ impl ErasureSetHealer {
 
             while let Some((key, object, version_id, result)) = page_tasks.next().await {
                 let (object_size, result) = result;
-                let mut telemetry_unknown = false;
-                let checkpoint_outcome = match result {
+                match result {
                     Ok(true) => {
-                        telemetry_unknown |= !increment_counter(successful_objects);
-                        telemetry_unknown |= !add_bytes(&mut bytes_processed, object_size);
+                        *successful_objects += 1;
+                        bytes_processed = bytes_processed.saturating_add(object_size);
+                        checkpoint_manager.add_processed_object(key).await?;
                         debug!(
                             target: "rustfs::heal::erasure_healer",
                             event = EVENT_HEAL_ERASURE_OBJECT_STATE,
@@ -1132,11 +971,11 @@ impl ErasureSetHealer {
                             state = "healed",
                             "Erasure set object healed"
                         );
-                        CheckpointObjectOutcome::Processed
                     }
                     Ok(false) => {
-                        telemetry_unknown |= !increment_counter(successful_objects);
-                        telemetry_unknown |= !add_bytes(&mut bytes_processed, object_size);
+                        checkpoint_manager.add_processed_object(key).await?;
+                        *successful_objects += 1;
+                        bytes_processed = bytes_processed.saturating_add(object_size);
                         debug!(
                             target: "rustfs::heal::erasure_healer",
                             event = EVENT_HEAL_ERASURE_OBJECT_STATE,
@@ -1149,12 +988,12 @@ impl ErasureSetHealer {
                             state = "missing_treated_as_ok",
                             "Erasure set missing object treated as ok"
                         );
-                        CheckpointObjectOutcome::Processed
                     }
                     Err(err @ Error::TaskCancelled) | Err(err @ Error::TaskTimeout) => return Err(err),
                     Err(Error::TransientSkip { message }) => {
-                        telemetry_unknown |= !increment_counter(skipped_objects);
-                        telemetry_unknown |= !add_bytes(&mut bytes_processed, object_size);
+                        *skipped_objects += 1;
+                        bytes_processed = bytes_processed.saturating_add(object_size);
+                        checkpoint_manager.add_skipped_object(key).await?;
                         demote_to_debug_when!(!take_failure_log_sample(&mut transient_skip_samples_logged), warn, target: "rustfs::heal::erasure_healer", {
                             event = EVENT_HEAL_ERASURE_OBJECT_STATE,
                             component = LOG_COMPONENT_HEAL,
@@ -1167,11 +1006,11 @@ impl ErasureSetHealer {
                             error = %message,
                             "Erasure set object heal skipped due to transient error"
                         });
-                        CheckpointObjectOutcome::Skipped
                     }
                     Err(err) => {
-                        telemetry_unknown |= !increment_counter(failed_objects);
-                        telemetry_unknown |= !add_bytes(&mut bytes_processed, object_size);
+                        *failed_objects += 1;
+                        bytes_processed = bytes_processed.saturating_add(object_size);
+                        checkpoint_manager.add_failed_object(key).await?;
                         demote_to_debug_when!(!take_failure_log_sample(&mut failure_samples_logged), warn, target: "rustfs::heal::erasure_healer", {
                             event = EVENT_HEAL_ERASURE_OBJECT_STATE,
                             component = LOG_COMPONENT_HEAL,
@@ -1184,43 +1023,15 @@ impl ErasureSetHealer {
                             error = %err,
                             "Erasure set object heal failed"
                         });
-                        CheckpointObjectOutcome::Failed
                     }
-                };
+                }
 
-                telemetry_unknown |= !increment_counter(processed_objects);
+                *processed_objects += 1;
                 completed_in_page += 1;
-                let (outcome_record, counter_unknown) = {
+                {
                     let mut progress = self.progress.write().await;
                     progress.set_current_object(Some(format!("{bucket}/{object}")));
-                    progress.update_object_progress(
-                        *processed_objects,
-                        *successful_objects,
-                        *failed_objects,
-                        *skipped_objects,
-                        bytes_processed,
-                    );
-                    if telemetry_unknown {
-                        progress.mark_unknown();
-                    }
-                    (
-                        CheckpointObjectOutcomeRecord {
-                            object: key,
-                            outcome: checkpoint_outcome,
-                            successful: progress.objects_healed,
-                            failed: progress.objects_failed,
-                            skipped: progress.skipped_objects,
-                            bytes: progress.bytes_processed,
-                            skipped_new_versions: progress.skipped_new_versions,
-                            skipped_ilm_expired: progress.skipped_ilm_expired,
-                            counter_unknown: progress.counter_unknown,
-                        },
-                        progress.counter_unknown,
-                    )
-                };
-                checkpoint_manager.record_object_outcome(outcome_record).await?;
-                if counter_unknown {
-                    resume_manager.mark_counter_unknown().await?;
+                    progress.update_progress(*processed_objects, *successful_objects, *failed_objects, bytes_processed);
                 }
 
                 if completed_in_page.is_multiple_of(100) {
@@ -1230,22 +1041,16 @@ impl ErasureSetHealer {
 
             *current_object_index = global_obj_idx;
 
-            // Persist the checkpoint ledger and page position before exposing
-            // the next resume cursor. A crash before cursor publication keeps
-            // the page identities available for exact-once replay.
-            checkpoint_manager.advance_page(bucket_index, *current_object_index).await?;
+            // Persist the authoritative cursor FIRST (points at the next page
+            // boundary), then prune the per-version dedup sets. Both are
+            // idempotent under crash: heal_object re-heals safely.
+            let next_cursor = if is_truncated { next_token.clone() } else { None };
+            resume_manager.set_resume_cursor(next_cursor.clone()).await?;
+            checkpoint_manager.complete_page(bucket_index, *current_object_index).await?;
             // Check if there are more pages
             if !is_truncated {
                 break;
             }
-            continuation_token = next_heal_listing_token(bucket, "", next_token, is_truncated)?;
-            if continuation_token.is_none() {
-                // A truncated page without a continuation token is terminal.
-                // Retain its ledger until bucket completion is durable.
-                break;
-            }
-            resume_manager.set_resume_cursor(continuation_token.clone()).await?;
-            checkpoint_manager.prune_completed_page().await?;
 
             // Anti-loop guard: an empty page reported as truncated cannot advance
             // the cursor (there is no last identity to move past), so treat it as a
@@ -1264,6 +1069,12 @@ impl ErasureSetHealer {
                 )));
             }
             previous_page_last = page_last;
+
+            continuation_token = next_heal_listing_token(bucket, "", next_token, is_truncated)?;
+            if continuation_token.is_none() {
+                // Truncated but no continuation token: treat as end of listing.
+                break;
+            }
         }
 
         Ok(())
@@ -1272,66 +1083,10 @@ impl ErasureSetHealer {
     /// initialize progress tracking
     async fn initialize_progress(&self, _buckets: &[String], state: &crate::heal::resume::ResumeState) {
         let mut progress = self.progress.write().await;
-        let existing_baseline = (
-            progress.objects_total_count,
-            progress.objects_total_size,
-            progress.baseline_generation,
-            progress.progress_state,
-            progress.baseline_known,
-        );
-        let baseline_generation_mismatch =
-            state.baseline_known && existing_baseline.4 && state.baseline_generation != existing_baseline.2;
-        let use_persisted_baseline = state.baseline_known && !baseline_generation_mismatch;
-        progress.objects_scanned = state.processed_objects;
+        progress.objects_scanned = state.total_objects;
         progress.objects_healed = state.successful_objects;
         progress.objects_failed = state.failed_objects;
-        progress.skipped_objects = state.skipped_objects;
-        progress.skipped_new_versions = state.skipped_new_versions;
-        progress.skipped_ilm_expired = state.skipped_ilm_expired;
-        progress.bytes_processed = state.processed_bytes;
-        progress.counter_unknown = state.counter_unknown;
-        if use_persisted_baseline
-            || existing_baseline.0 > 0
-            || existing_baseline.1 > 0
-            || existing_baseline.2.is_some()
-            || existing_baseline.4
-        {
-            progress.objects_total_count = if use_persisted_baseline {
-                state.total_objects
-            } else {
-                existing_baseline.0
-            };
-            progress.objects_total_size = if use_persisted_baseline {
-                state.total_bytes
-            } else {
-                existing_baseline.1
-            };
-            progress.baseline_generation = if use_persisted_baseline {
-                state.baseline_generation
-            } else {
-                existing_baseline.2
-            };
-            progress.baseline_known = use_persisted_baseline
-                || existing_baseline.0 > 0
-                || existing_baseline.1 > 0
-                || existing_baseline.2.is_some()
-                || existing_baseline.4;
-        }
-        progress.progress_state = if use_persisted_baseline
-            || existing_baseline.0 > 0
-            || existing_baseline.1 > 0
-            || existing_baseline.2.is_some()
-            || existing_baseline.4
-        {
-            crate::heal::progress::HealProgressState::Running
-        } else {
-            crate::heal::progress::HealProgressState::Indeterminate
-        };
-        if baseline_generation_mismatch || state.counter_unknown {
-            progress.mark_unknown();
-        }
-        progress.ledger_complete = false;
-        progress.refresh_progress_percentage();
+        progress.bytes_processed = 0; // Resume state tracks object counts, not byte counters.
         progress.start_time = UNIX_EPOCH.checked_add(Duration::from_secs(state.start_time));
         progress.last_update_time = UNIX_EPOCH.checked_add(Duration::from_secs(state.last_update));
         progress.set_current_object(state.current_object.clone());
@@ -1509,8 +1264,8 @@ mod resume_loop_tests {
     };
     use crate::heal::progress::HealProgress;
     use crate::heal::resume::{
-        CheckpointManager, CheckpointObjectOutcome, CheckpointObjectOutcomeRecord, RESUME_CHECKPOINT_FILE,
-        ReplacementTargetIdentity, ResumeDeleteFailure, ResumeManager, ResumeUtils, compose_key,
+        CheckpointManager, RESUME_CHECKPOINT_FILE, ReplacementTargetIdentity, ResumeDeleteFailure, ResumeManager, ResumeUtils,
+        compose_key,
     };
     use crate::heal::storage::{HealLifecycleExpiryContext, HealListItem, HealObjectInfo, HealStorageAPI};
     use crate::heal::storage_api::status::BucketInfo;
@@ -1650,7 +1405,6 @@ mod resume_loop_tests {
         list_include_lifecycle_object_info: Mutex<Vec<bool>>,
         replacement_target_identity_sequences: Mutex<VecDeque<Vec<ReplacementTargetIdentity>>>,
         fail_listing: AtomicBool,
-        fail_listing_buckets: Mutex<HashSet<String>>,
     }
 
     impl FakeStorage {
@@ -1687,9 +1441,6 @@ mod resume_loop_tests {
         fn fail_listing(&self) {
             self.fail_listing.store(true, Ordering::SeqCst);
         }
-        fn fail_bucket_listing(&self, bucket: &str) {
-            self.fail_listing_buckets.lock().unwrap().insert(bucket.to_string());
-        }
     }
 
     #[async_trait::async_trait]
@@ -1780,7 +1531,7 @@ mod resume_loop_tests {
         }
         async fn list_objects_for_heal_page(
             &self,
-            bucket: &str,
+            _bucket: &str,
             _prefix: &str,
             continuation_token: Option<&str>,
             include_lifecycle_object_info: bool,
@@ -1789,7 +1540,7 @@ mod resume_loop_tests {
                 .lock()
                 .unwrap()
                 .push(include_lifecycle_object_info);
-            if self.fail_listing.load(Ordering::SeqCst) || self.fail_listing_buckets.lock().unwrap().contains(bucket) {
+            if self.fail_listing.load(Ordering::SeqCst) {
                 return Err(Error::other("injected listing failure"));
             }
             let key = continuation_token.map(str::to_string);
@@ -2167,49 +1918,6 @@ mod resume_loop_tests {
         assert!(state.completed_buckets.is_empty(), "the failed bucket must remain resumable");
     }
 
-    #[tokio::test]
-    async fn bucket_failure_stops_before_a_later_bucket_checkpoint() {
-        let env = make_env().await;
-        let task_id = ResumeUtils::generate_task_id();
-        let buckets = vec!["a".to_string(), "b".to_string()];
-        let resume = ResumeManager::new(
-            env.healer.disk.clone(),
-            task_id.clone(),
-            "erasure_set".to_string(),
-            "pool_0_set_0".to_string(),
-            buckets.clone(),
-        )
-        .await
-        .unwrap();
-        let checkpoint = CheckpointManager::new(env.healer.disk.clone(), task_id.clone())
-            .await
-            .unwrap();
-        env.storage.fail_bucket_listing("a");
-        for _ in 0..3 {
-            assert!(resume.schedule_retry().await.unwrap());
-        }
-
-        env.healer
-            .execute_heal_with_resume(&buckets, "pool_0_set_0", &resume, &checkpoint)
-            .await
-            .expect_err("the first bucket failure must keep the pass incomplete");
-        let persisted = checkpoint.get_checkpoint().await;
-        assert_eq!(persisted.current_bucket_index, 0);
-        assert!(resume.get_state().await.completed_buckets.is_empty());
-
-        let resumed = ResumeManager::load_from_disk(env.healer.disk.clone(), &task_id)
-            .await
-            .unwrap();
-        let checkpoint = CheckpointManager::load_from_disk(env.healer.disk.clone(), &task_id)
-            .await
-            .unwrap();
-        env.healer
-            .execute_heal_with_resume(&buckets, "pool_0_set_0", &resumed, &checkpoint)
-            .await
-            .expect_err("recovery must retry the earlier failed bucket");
-        assert!(!resumed.get_state().await.completed);
-    }
-
     #[tokio::test]
     async fn completed_resume_state_is_not_selected_for_a_new_heal() {
         let env = make_env().await;
@@ -2399,175 +2107,8 @@ mod resume_loop_tests {
         let mut names: Vec<String> = env.storage.calls().into_iter().map(|(n, _)| n).collect();
         names.sort();
         assert_eq!(names, vec!["a", "b", "c", "d"], "every object exactly once, none dropped/doubled");
-        // Keep the final page cursor until the outer loop durably completes the
-        // bucket, so a crash can replay only this page against its identities.
-        assert_eq!(env.resume.resume_cursor().await, Some("t1".to_string()));
-    }
-
-    #[tokio::test]
-    async fn persisted_failure_waits_for_the_bounded_retry_after_page_replay() {
-        let env = make_env().await;
-        env.storage.set_page(
-            None,
-            Page {
-                items: vec![item("object", Some("v1"), false)],
-                next: None,
-                truncated: false,
-            },
-        );
-        env.checkpoint
-            .record_object_outcome(CheckpointObjectOutcomeRecord {
-                object: compose_key("object", Some("v1")),
-                outcome: CheckpointObjectOutcome::Failed,
-                successful: 0,
-                failed: 1,
-                skipped: 0,
-                bytes: 0,
-                skipped_new_versions: 0,
-                skipped_ilm_expired: 0,
-                counter_unknown: false,
-            })
-            .await
-            .unwrap();
-        env.checkpoint.advance_page(0, 1).await.unwrap();
-
-        let resumed = ResumeManager::load_from_disk(env.healer.disk.clone(), &env.task_id)
-            .await
-            .unwrap();
-        let checkpoint = CheckpointManager::load_from_disk(env.healer.disk.clone(), &env.task_id)
-            .await
-            .unwrap();
-
-        env.healer
-            .execute_heal_with_resume(&["b".to_string()], "pool_0_set_0", &resumed, &checkpoint)
-            .await
-            .expect_err("the persisted failure must schedule a bounded retry");
-        assert!(
-            env.storage.calls().is_empty(),
-            "the failed identity must not be repeated in the same pass"
-        );
-
-        env.healer
-            .execute_heal_with_resume(&["b".to_string()], "pool_0_set_0", &resumed, &checkpoint)
-            .await
-            .expect("the bounded retry must heal the object");
-        assert_eq!(env.storage.calls(), vec![("object".to_string(), Some("v1".to_string()))]);
-        let state = resumed.get_state().await;
-        assert_eq!(state.successful_objects, 1);
-        assert_eq!(state.failed_objects, 0);
-    }
-
-    #[tokio::test]
-    async fn final_page_crash_replays_only_the_retained_page_identities() {
-        let env = make_env().await;
-        env.storage.set_page(
-            None,
-            Page {
-                items: vec![item("first", Some("v1"), false)],
-                next: Some("final-page".to_string()),
-                truncated: true,
-            },
-        );
-        env.storage.set_page(
-            Some("final-page"),
-            Page {
-                items: vec![item("last", Some("v1"), false)],
-                next: None,
-                truncated: false,
-            },
-        );
-
-        let (processed, successful, failed, skipped, result) = run(&env).await;
-        result.expect("the bucket pass must finish before the simulated crash");
-        assert_eq!((processed, successful, failed, skipped), (2, 2, 0, 0));
-
-        let resumed = ResumeManager::load_from_disk(env.healer.disk.clone(), &env.task_id)
-            .await
-            .unwrap();
-        let checkpoint = CheckpointManager::load_from_disk(env.healer.disk.clone(), &env.task_id)
-            .await
-            .unwrap();
-        env.healer
-            .execute_heal_with_resume(&["b".to_string()], "pool_0_set_0", &resumed, &checkpoint)
-            .await
-            .expect("the retained final-page ledger must make recovery exact");
-
-        assert_eq!(
-            env.storage.calls(),
-            vec![
-                ("first".to_string(), Some("v1".to_string())),
-                ("last".to_string(), Some("v1".to_string()))
-            ]
-        );
-        let state = resumed.get_state().await;
-        assert_eq!(state.successful_objects, 2);
-        assert_eq!(state.processed_objects, 2);
-    }
-
-    #[tokio::test]
-    async fn truncated_page_without_token_retains_its_replay_ledger() {
-        let env = make_env().await;
-        env.storage.set_page(
-            None,
-            Page {
-                items: vec![item("object", Some("v1"), false)],
-                next: None,
-                truncated: true,
-            },
-        );
-
-        let (processed, successful, failed, skipped, result) = run(&env).await;
-        result.expect("the tokenless truncated page is a terminal page");
-        assert_eq!((processed, successful, failed, skipped), (1, 1, 0, 0));
-
-        let resumed = ResumeManager::load_from_disk(env.healer.disk.clone(), &env.task_id)
-            .await
-            .unwrap();
-        let checkpoint = CheckpointManager::load_from_disk(env.healer.disk.clone(), &env.task_id)
-            .await
-            .unwrap();
-        env.healer
-            .execute_heal_with_resume(&["b".to_string()], "pool_0_set_0", &resumed, &checkpoint)
-            .await
-            .expect("terminal-page recovery must not replay a durable identity");
-
-        assert_eq!(env.storage.calls(), vec![("object".to_string(), Some("v1".to_string()))]);
-        let state = resumed.get_state().await;
-        assert_eq!(state.successful_objects, 1);
-        assert_eq!(state.processed_objects, 1);
-    }
-
-    #[tokio::test]
-    async fn completed_bucket_reconciles_its_final_page_checkpoint_after_crash() {
-        let env = make_env().await;
-        env.storage.set_page(
-            None,
-            Page {
-                items: vec![item("object", Some("v1"), false)],
-                next: None,
-                truncated: false,
-            },
-        );
-
-        let (_, _, _, _, result) = run(&env).await;
-        result.expect("the bucket pass must finish before the simulated crash");
-        env.resume.complete_bucket("b").await.unwrap();
-
-        let resumed = ResumeManager::load_from_disk(env.healer.disk.clone(), &env.task_id)
-            .await
-            .unwrap();
-        let checkpoint = CheckpointManager::load_from_disk(env.healer.disk.clone(), &env.task_id)
-            .await
-            .unwrap();
-        env.healer
-            .execute_heal_with_resume(&["b".to_string()], "pool_0_set_0", &resumed, &checkpoint)
-            .await
-            .expect("recovery must finish the checkpoint transition without replaying the bucket");
-
-        assert_eq!(env.storage.calls(), vec![("object".to_string(), Some("v1".to_string()))]);
-        let checkpoint = checkpoint.get_checkpoint().await;
-        assert_eq!(checkpoint.current_bucket_index, 1);
-        assert!(checkpoint.processed_objects.is_empty());
+        // Final page not truncated => cursor cleared.
+        assert_eq!(env.resume.resume_cursor().await, None);
     }
 
     #[tokio::test]

crates/heal/src/heal/manager.rs

@@ -2011,11 +2011,34 @@ impl HealManager {
             return None;
         }
 
-        let mut progresses = Vec::with_capacity(active_tasks.len());
+        let mut snapshot = HealProgress::default();
         for task in active_tasks {
-            progresses.push(task.get_progress().await);
+            let progress = task.get_progress().await;
+            snapshot.objects_scanned = snapshot.objects_scanned.saturating_add(progress.objects_scanned);
+            snapshot.objects_healed = snapshot.objects_healed.saturating_add(progress.objects_healed);
+            snapshot.objects_failed = snapshot.objects_failed.saturating_add(progress.objects_failed);
+            snapshot.skipped_new_versions = snapshot.skipped_new_versions.saturating_add(progress.skipped_new_versions);
+            snapshot.skipped_ilm_expired = snapshot.skipped_ilm_expired.saturating_add(progress.skipped_ilm_expired);
+            snapshot.objects_total_count = snapshot.objects_total_count.saturating_add(progress.objects_total_count);
+            snapshot.objects_total_size = snapshot.objects_total_size.saturating_add(progress.objects_total_size);
+            snapshot.bytes_processed = snapshot.bytes_processed.saturating_add(progress.bytes_processed);
+            snapshot.start_time = match (snapshot.start_time, progress.start_time) {
+                (Some(current), Some(next)) => Some(current.min(next)),
+                (None, next) => next,
+                (current, None) => current,
+            };
+            snapshot.last_update_time = match (snapshot.last_update_time, progress.last_update_time) {
+                (Some(current), Some(next)) => Some(current.max(next)),
+                (None, next) => next,
+                (current, None) => current,
+            };
+            if progress.current_object.is_some() {
+                snapshot.current_object = progress.current_object;
+            }
         }
-        crate::heal::progress::aggregate_heal_progress(progresses)
+        snapshot.refresh_progress_percentage();
+        snapshot.refresh_estimated_completion_time();
+        Some(snapshot)
     }
 }
 

crates/heal/src/heal/progress.rs

@@ -15,91 +15,15 @@
 use serde::{Deserialize, Serialize};
 use std::time::{Duration, SystemTime};
 
-pub(crate) fn stable_generation(parts: &[&[u8]]) -> u64 {
-    let mut hash = 0xcbf29ce484222325u64;
-    for part in parts {
-        for byte in (part.len() as u64).to_be_bytes().into_iter().chain(part.iter().copied()) {
-            hash ^= u64::from(byte);
-            hash = hash.wrapping_mul(0x100000001b3);
-        }
-    }
-    hash
-}
-
-#[cfg(test)]
-mod stable_generation_tests {
-    use super::stable_generation;
-
-    #[test]
-    fn stable_generation_has_a_fixed_vector() {
-        assert_eq!(stable_generation(&[b"rustfs", b"heal", b"42"]), 11_007_672_338_488_385_056);
-    }
-}
-
-pub(crate) fn increment_counter(counter: &mut u64) -> bool {
-    match counter.checked_add(1) {
-        Some(next) => {
-            *counter = next;
-            true
-        }
-        None => {
-            *counter = u64::MAX;
-            false
-        }
-    }
-}
-
-pub(crate) fn add_bytes(total: &mut u64, amount: u64) -> bool {
-    match total.checked_add(amount) {
-        Some(next) => {
-            *total = next;
-            true
-        }
-        None => {
-            *total = u64::MAX;
-            false
-        }
-    }
-}
-
-#[derive(Debug, Default, Clone, Copy, PartialEq, Eq, Serialize, Deserialize)]
+#[derive(Debug, Default, Clone, Serialize, Deserialize)]
 #[serde(rename_all = "camelCase")]
-pub enum HealProgressKind {
-    #[default]
-    Unknown,
-    Stage,
-    ObjectSweep,
-}
-
-/// Whether the object ledger can produce a meaningful percentage.
-///
-/// A zero-valued baseline is not a completed scan: it means that no complete
-/// usage snapshot was available.  Keep this state explicit so callers do not
-/// mistake the legacy `0.0` wire value for a measured zero-percent result.
-#[derive(Debug, Default, Clone, Copy, PartialEq, Eq, Serialize, Deserialize)]
-#[serde(rename_all = "camelCase")]
-pub enum HealProgressState {
-    #[default]
-    Unknown,
-    Indeterminate,
-    Running,
-    Completed,
-}
-
-#[derive(Debug, Default, Clone, PartialEq, Serialize, Deserialize)]
-#[serde(default, rename_all = "camelCase")]
 pub struct HealProgress {
-    #[serde(default)]
-    pub kind: HealProgressKind,
     /// Objects scanned
     pub objects_scanned: u64,
     /// Objects healed
     pub objects_healed: u64,
     /// Objects failed
     pub objects_failed: u64,
-    /// Versions deferred for a later retry pass.
-    #[serde(default)]
-    pub skipped_objects: u64,
     /// Versions skipped because they were written after this heal started
     pub skipped_new_versions: u64,
     /// Versions skipped because lifecycle already selected them for expiry
@@ -120,38 +44,11 @@ pub struct HealProgress {
     pub last_update_time: Option<SystemTime>,
     /// Estimated completion time
     pub estimated_completion_time: Option<SystemTime>,
-    /// Current stage number. Stage updates are intentionally independent from
-    /// the object ledger below.
-    #[serde(default)]
-    pub stage_current: u64,
-    /// Number of stages in the current task.
-    #[serde(default)]
-    pub stage_total: u64,
-    /// Explicitly distinguishes a missing usage baseline from measured 0%.
-    #[serde(default)]
-    pub progress_state: HealProgressState,
-    /// True only after the task's durable completion ledger was committed.
-    #[serde(default)]
-    pub ledger_complete: bool,
-    /// Generation of the usage snapshot used for the baseline, if available.
-    #[serde(default)]
-    pub baseline_generation: Option<u64>,
-    /// Whether the baseline was explicitly observed.  This is separate from
-    /// the counters so a known empty scope (0 objects, 0 bytes) is not
-    /// confused with a legacy snapshot that omitted the baseline fields.
-    #[serde(default)]
-    pub baseline_known: bool,
-    /// Internal telemetry fence set when an aggregate counter overflows or
-    /// becomes inconsistent.  It prevents a later refresh from fabricating a
-    /// percentage from the poisoned values.
-    #[serde(default)]
-    pub counter_unknown: bool,
 }
 
 impl HealProgress {
     pub fn new() -> Self {
         Self {
-            kind: HealProgressKind::Unknown,
             start_time: Some(SystemTime::now()),
             last_update_time: Some(SystemTime::now()),
             ..Default::default()
@@ -159,87 +56,12 @@ impl HealProgress {
     }
 
     pub fn update_progress(&mut self, scanned: u64, healed: u64, failed: u64, bytes: u64) {
-        self.update_object_sweep_progress(scanned, healed, failed, bytes);
-    }
-
-    pub fn update_object_sweep_progress(&mut self, scanned: u64, healed: u64, failed: u64, bytes: u64) {
-        self.kind = HealProgressKind::ObjectSweep;
         self.objects_scanned = scanned;
         self.objects_healed = healed;
         self.objects_failed = failed;
         self.bytes_processed = bytes;
         self.last_update_time = Some(SystemTime::now());
 
-        let explicit_skipped = match self.skipped_new_versions.checked_add(self.skipped_ilm_expired) {
-            Some(value) => value,
-            None => {
-                self.mark_unknown();
-                0
-            }
-        };
-        let skipped = healed
-            .checked_add(failed)
-            .and_then(|value| value.checked_add(explicit_skipped))
-            .and_then(|value| scanned.checked_sub(value))
-            .unwrap_or(0);
-        self.update_object_progress(scanned, healed, failed, skipped, bytes);
-    }
-
-    /// Update task stage progress without modifying object counters.
-    pub fn update_stage(&mut self, current: u64, total: u64) {
-        let object_sweep_active = matches!(self.kind, HealProgressKind::ObjectSweep);
-        if !object_sweep_active {
-            self.kind = HealProgressKind::Stage;
-        }
-        self.ledger_complete = false;
-        self.stage_current = current.min(total);
-        self.stage_total = total;
-        if object_sweep_active {
-            self.last_update_time = Some(SystemTime::now());
-            self.refresh_progress_percentage();
-            return;
-        }
-        self.progress_state = if total == 0 {
-            HealProgressState::Indeterminate
-        } else {
-            HealProgressState::Running
-        };
-        self.progress_percentage = if total == 0 {
-            0.0
-        } else {
-            (current as f64 / total as f64 * 100.0).min(100.0)
-        };
-        self.last_update_time = Some(SystemTime::now());
-    }
-
-    /// Update the disjoint object ledger. `scanned` is the number of terminal
-    /// object outcomes and must equal healed + failed + deferred skipped plus
-    /// the two terminal skip classes. Overflow is a corrupt/unknown counter
-    /// state, not a reason to abort a completed heal.
-    pub fn update_object_progress(&mut self, scanned: u64, healed: u64, failed: u64, skipped: u64, bytes: u64) {
-        self.kind = HealProgressKind::ObjectSweep;
-        // `skipped` is the transient/deferred class.  The two explicit skip
-        // counters are terminal classifications too, so include them in the
-        // same ledger without making callers maintain a second aggregate.
-        let outcomes = healed
-            .checked_add(failed)
-            .and_then(|value| value.checked_add(skipped))
-            .and_then(|value| value.checked_add(self.skipped_new_versions))
-            .and_then(|value| value.checked_add(self.skipped_ilm_expired));
-        self.objects_scanned = scanned;
-        self.objects_healed = healed;
-        self.objects_failed = failed;
-        self.skipped_objects = skipped;
-        self.bytes_processed = bytes;
-        self.last_update_time = Some(SystemTime::now());
-        self.ledger_complete = false;
-        if outcomes != Some(scanned) {
-            // Telemetry corruption must not abort a heal.  Preserve the
-            // counters for diagnostics, but do not derive a percentage from a
-            // double-counted or overflowing ledger.
-            self.mark_unknown();
-            return;
-        }
         self.refresh_progress_percentage();
         self.refresh_estimated_completion_time();
     }
@@ -247,88 +69,50 @@ impl HealProgress {
     pub fn set_total_baseline(&mut self, objects_total_count: u64, objects_total_size: u64) {
         self.objects_total_count = objects_total_count;
         self.objects_total_size = objects_total_size;
-        self.baseline_known = true;
         self.last_update_time = Some(SystemTime::now());
         self.refresh_progress_percentage();
         self.refresh_estimated_completion_time();
     }
 
-    pub fn set_total_baseline_with_generation(&mut self, objects_total_count: u64, objects_total_size: u64, generation: u64) {
-        self.baseline_generation = Some(generation);
-        self.set_total_baseline(objects_total_count, objects_total_size);
-    }
-
     pub fn record_skipped_new_version(&mut self) {
-        let Some(next) = self.skipped_new_versions.checked_add(1) else {
-            self.mark_unknown();
-            return;
-        };
-        self.skipped_new_versions = next;
+        self.skipped_new_versions = self.skipped_new_versions.saturating_add(1);
         self.last_update_time = Some(SystemTime::now());
         self.refresh_progress_percentage();
         self.refresh_estimated_completion_time();
     }
 
     pub fn record_skipped_ilm_expired(&mut self) {
-        let Some(next) = self.skipped_ilm_expired.checked_add(1) else {
-            self.mark_unknown();
-            return;
-        };
-        self.skipped_ilm_expired = next;
+        self.skipped_ilm_expired = self.skipped_ilm_expired.saturating_add(1);
         self.last_update_time = Some(SystemTime::now());
         self.refresh_progress_percentage();
         self.refresh_estimated_completion_time();
     }
 
-    fn completed_for_baseline(&self) -> Option<u64> {
+    fn completed_for_baseline(&self) -> u64 {
         self.objects_healed
-            .checked_add(self.objects_failed)?
-            .checked_add(self.skipped_objects)?
-            .checked_add(self.skipped_new_versions)?
-            .checked_add(self.skipped_ilm_expired)
+            .saturating_add(self.objects_failed)
+            .saturating_add(self.skipped_new_versions)
+            .saturating_add(self.skipped_ilm_expired)
     }
 
     pub(crate) fn refresh_progress_percentage(&mut self) {
-        if self.ledger_complete {
-            self.progress_state = HealProgressState::Completed;
-            self.progress_percentage = 100.0;
-            return;
-        }
-        if self.counter_unknown {
-            self.progress_state = HealProgressState::Unknown;
-            self.progress_percentage = 0.0;
-            return;
-        }
-        if !self.baseline_known {
-            self.progress_state = HealProgressState::Indeterminate;
-            self.progress_percentage = 0.0;
-            self.estimated_completion_time = None;
-            return;
-        }
         if self.objects_total_size > 0 {
             self.progress_percentage = ((self.bytes_processed as f64 / self.objects_total_size as f64) * 100.0).min(100.0);
-            self.progress_percentage = self.progress_percentage.min(99.999);
-            self.progress_state = HealProgressState::Running;
             return;
         }
         if self.objects_total_count > 0 {
-            let Some(completed) = self.completed_for_baseline() else {
-                self.progress_state = HealProgressState::Unknown;
-                self.progress_percentage = 0.0;
-                return;
-            };
+            let completed = self.completed_for_baseline();
             self.progress_percentage = ((completed as f64 / self.objects_total_count as f64) * 100.0).min(100.0);
-            self.progress_percentage = self.progress_percentage.min(99.999);
-            self.progress_state = HealProgressState::Running;
             return;
         }
-        if self.baseline_known {
-            self.progress_state = HealProgressState::Running;
-            self.progress_percentage = 0.0;
-            return;
+
+        let total = self
+            .objects_scanned
+            .saturating_add(self.objects_healed)
+            .saturating_add(self.objects_failed);
+        if total > 0 {
+            self.progress_percentage = (self.objects_healed as f64 / total as f64) * 100.0;
         }
-        self.progress_state = HealProgressState::Indeterminate;
-        self.progress_percentage = 0.0;
     }
 
     pub fn set_current_object(&mut self, object: Option<String>) {
@@ -341,11 +125,7 @@ impl HealProgress {
             self.estimated_completion_time = None;
             return;
         };
-        if self.is_completed()
-            || self.progress_percentage <= 0.0
-            || self.progress_percentage >= 100.0
-            || self.bytes_processed == 0
-        {
+        if self.is_completed() || !(0.0..100.0).contains(&self.progress_percentage) || self.bytes_processed == 0 {
             self.estimated_completion_time = None;
             return;
         }
@@ -362,39 +142,18 @@ impl HealProgress {
     }
 
     pub fn is_completed(&self) -> bool {
-        self.ledger_complete
-    }
-
-    /// Mark telemetry unknown while allowing the underlying heal operation to
-    /// continue.  This is used for corrupt/overflowing counters at the
-    /// observability boundary; it must never turn a successful heal into an
-    /// execution error.
-    pub fn mark_unknown(&mut self) {
-        self.counter_unknown = true;
-        self.progress_state = HealProgressState::Unknown;
-        self.ledger_complete = false;
-        self.progress_percentage = 0.0;
-        self.estimated_completion_time = None;
-        self.last_update_time = Some(SystemTime::now());
-    }
-
-    /// Mark the object ledger terminal only after the enclosing task has
-    /// committed all durable resume state and cleanup fences.
-    pub fn mark_completed(&mut self) {
-        let telemetry_unknown = self.counter_unknown || self.progress_state == HealProgressState::Unknown;
-        self.ledger_complete = true;
-        if !telemetry_unknown {
-            self.progress_state = HealProgressState::Completed;
+        if self.progress_percentage >= 100.0 {
+            return true;
         }
-        self.progress_percentage = 100.0;
-        self.last_update_time = Some(SystemTime::now());
-        self.estimated_completion_time = None;
+        if self.objects_total_count > 0 || self.objects_total_size > 0 {
+            return false;
+        }
+
+        self.objects_scanned > 0 && self.objects_healed.saturating_add(self.objects_failed) >= self.objects_scanned
     }
 
     pub fn get_success_rate(&self) -> f64 {
-        let Some(total) = self.objects_healed.checked_add(self.objects_failed) else {
-            return 0.0;
-        };
+        let total = self.objects_healed + self.objects_failed;
         if total > 0 {
             (self.objects_healed as f64 / total as f64) * 100.0
         } else {
@@ -403,101 +162,6 @@ impl HealProgress {
     }
 }
 
-pub fn aggregate_heal_progress(progresses: impl IntoIterator<Item = HealProgress>) -> Option<HealProgress> {
-    let mut snapshot = HealProgress::default();
-    let mut found = false;
-    let mut has_object_sweep = false;
-    let mut all_object_baselines_known = true;
-    let mut baseline_generation = None;
-    let mut baseline_generation_consistent = true;
-    let mut all_ledgers_complete = true;
-    let mut counter_overflow = false;
-
-    for progress in progresses {
-        found = true;
-        let object_sweep = matches!(progress.kind, HealProgressKind::ObjectSweep);
-        has_object_sweep |= object_sweep;
-        all_ledgers_complete &= progress.ledger_complete;
-        if object_sweep {
-            all_object_baselines_known &= progress.baseline_known;
-            match baseline_generation {
-                None => baseline_generation = Some(progress.baseline_generation),
-                Some(generation) => baseline_generation_consistent &= generation == progress.baseline_generation,
-            }
-        }
-        counter_overflow |= progress.counter_unknown || matches!(progress.progress_state, HealProgressState::Unknown);
-        for (target, value) in [
-            (&mut snapshot.objects_scanned, progress.objects_scanned),
-            (&mut snapshot.objects_healed, progress.objects_healed),
-            (&mut snapshot.objects_failed, progress.objects_failed),
-            (&mut snapshot.skipped_objects, progress.skipped_objects),
-            (&mut snapshot.skipped_new_versions, progress.skipped_new_versions),
-            (&mut snapshot.skipped_ilm_expired, progress.skipped_ilm_expired),
-            (&mut snapshot.objects_total_count, progress.objects_total_count),
-            (&mut snapshot.objects_total_size, progress.objects_total_size),
-            (&mut snapshot.bytes_processed, progress.bytes_processed),
-            (&mut snapshot.stage_current, progress.stage_current),
-            (&mut snapshot.stage_total, progress.stage_total),
-        ] {
-            match target.checked_add(value) {
-                Some(sum) => *target = sum,
-                None => {
-                    *target = u64::MAX;
-                    counter_overflow = true;
-                }
-            }
-        }
-        snapshot.start_time = match (snapshot.start_time, progress.start_time) {
-            (Some(current), Some(next)) => Some(current.min(next)),
-            (None, next) => next,
-            (current, None) => current,
-        };
-        snapshot.last_update_time = match (snapshot.last_update_time, progress.last_update_time) {
-            (Some(current), Some(next)) => Some(current.max(next)),
-            (None, next) => next,
-            (current, None) => current,
-        };
-        if progress.current_object.is_some() {
-            snapshot.current_object = progress.current_object;
-        }
-    }
-
-    if !found {
-        return None;
-    }
-
-    snapshot.kind = if has_object_sweep {
-        HealProgressKind::ObjectSweep
-    } else {
-        HealProgressKind::Stage
-    };
-    snapshot.baseline_known = has_object_sweep && all_object_baselines_known && baseline_generation_consistent;
-    snapshot.baseline_generation = if snapshot.baseline_known && baseline_generation_consistent {
-        baseline_generation.flatten()
-    } else {
-        None
-    };
-    snapshot.ledger_complete = all_ledgers_complete;
-    snapshot.counter_unknown = counter_overflow;
-    if counter_overflow {
-        snapshot.progress_state = HealProgressState::Unknown;
-        snapshot.progress_percentage = if snapshot.ledger_complete { 100.0 } else { 0.0 };
-    } else if snapshot.ledger_complete {
-        snapshot.progress_state = HealProgressState::Completed;
-        snapshot.progress_percentage = 100.0;
-    } else if has_object_sweep {
-        snapshot.refresh_progress_percentage();
-    } else if snapshot.stage_total == 0 {
-        snapshot.progress_state = HealProgressState::Indeterminate;
-        snapshot.progress_percentage = 0.0;
-    } else {
-        snapshot.progress_state = HealProgressState::Running;
-        snapshot.progress_percentage = ((snapshot.stage_current as f64 / snapshot.stage_total as f64) * 100.0).min(99.999);
-    }
-    snapshot.refresh_estimated_completion_time();
-    Some(snapshot)
-}
-
 #[derive(Debug, Clone, Serialize, Deserialize)]
 pub struct HealStatistics {
     /// Total heal tasks
@@ -566,7 +230,6 @@ mod tests {
         assert_eq!(progress.objects_scanned, 0);
         assert_eq!(progress.objects_healed, 0);
         assert_eq!(progress.objects_failed, 0);
-        assert_eq!(progress.skipped_objects, 0);
         assert_eq!(progress.skipped_new_versions, 0);
         assert_eq!(progress.skipped_ilm_expired, 0);
         assert_eq!(progress.objects_total_count, 0);
@@ -587,8 +250,10 @@ mod tests {
         assert_eq!(progress.objects_healed, 8);
         assert_eq!(progress.objects_failed, 2);
         assert_eq!(progress.bytes_processed, 1024);
-        assert_eq!(progress.progress_state, HealProgressState::Indeterminate);
-        assert_eq!(progress.progress_percentage, 0.0);
+        // Progress percentage should be calculated based on healed/total
+        // total = scanned + healed + failed = 10 + 8 + 2 = 20
+        // healed/total = 8/20 = 0.4 = 40%
+        assert!((progress.progress_percentage - 40.0).abs() < 0.001);
         assert!(progress.last_update_time.is_some());
     }
 
@@ -597,8 +262,7 @@ mod tests {
         let mut progress = HealProgress::new();
         progress.start_time = Some(SystemTime::now() - Duration::from_secs(10));
 
-        progress.set_total_baseline(100, 16384);
-        progress.update_progress(25, 25, 0, 4096);
+        progress.update_progress(100, 25, 0, 4096);
 
         let eta = progress
             .estimated_completion_time
@@ -611,7 +275,7 @@ mod tests {
         let mut progress = HealProgress::new();
         progress.set_total_baseline(10, 8192);
 
-        progress.update_progress(25, 25, 0, 4096);
+        progress.update_progress(100, 25, 0, 4096);
 
         assert!((progress.progress_percentage - 50.0).abs() < 0.001);
     }
@@ -621,7 +285,7 @@ mod tests {
         let mut progress = HealProgress::new();
         progress.set_total_baseline(10, 0);
 
-        progress.update_progress(5, 3, 2, 0);
+        progress.update_progress(100, 3, 2, 0);
 
         assert!((progress.progress_percentage - 50.0).abs() < 0.001);
     }
@@ -631,7 +295,7 @@ mod tests {
         let mut progress = HealProgress::new();
         progress.set_total_baseline(10, 0);
 
-        progress.update_progress(5, 3, 2, 0);
+        progress.update_progress(100, 3, 2, 0);
         progress.record_skipped_new_version();
 
         assert_eq!(progress.skipped_new_versions, 1);
@@ -672,8 +336,7 @@ mod tests {
     fn test_heal_progress_update_progress_all_healed() {
         let mut progress = HealProgress::new();
         // When scanned=0, healed=10, failed=0: total=10, progress = 10/10 = 100%
-        progress.update_progress(10, 10, 0, 2048);
-        progress.mark_completed();
+        progress.update_progress(0, 10, 0, 2048);
 
         // All healed, should be 100%
         assert!((progress.progress_percentage - 100.0).abs() < 0.001);
@@ -731,7 +394,6 @@ mod tests {
         assert_eq!(json["objectsScanned"], 10);
         assert_eq!(json["objectsHealed"], 8);
         assert_eq!(json["objectsFailed"], 2);
-        assert_eq!(json["skippedObjects"], 0);
         assert_eq!(json["skippedNewVersions"], 0);
         assert_eq!(json["skippedIlmExpired"], 0);
         assert_eq!(json["bytesProcessed"], 1024);
@@ -743,7 +405,6 @@ mod tests {
     fn test_heal_progress_is_completed_by_percentage() {
         let mut progress = HealProgress::new();
         progress.update_progress(10, 10, 0, 1024);
-        progress.mark_completed();
 
         assert!(progress.is_completed());
     }
@@ -754,7 +415,7 @@ mod tests {
         progress.objects_scanned = 10;
         progress.objects_healed = 8;
         progress.objects_failed = 2;
-        progress.mark_completed();
+        // healed + failed = 8 + 2 = 10 >= scanned = 10
         assert!(progress.is_completed());
     }
 
@@ -794,108 +455,6 @@ mod tests {
         assert!((progress.get_success_rate() - 100.0).abs() < 0.001);
     }
 
-    #[test]
-    fn single_object_progress_reaches_terminal_100() {
-        let mut progress = HealProgress::new();
-        progress.update_object_progress(1, 1, 0, 0, 128);
-        assert!(!progress.is_completed());
-        progress.mark_completed();
-        assert!(progress.is_completed());
-        assert_eq!(progress.progress_percentage, 100.0);
-    }
-
-    #[test]
-    fn progress_without_baseline_is_indeterminate() {
-        let mut progress = HealProgress::new();
-        progress.update_object_progress(1, 1, 0, 0, 128);
-        assert_eq!(progress.progress_state, HealProgressState::Indeterminate);
-        assert_eq!(progress.progress_percentage, 0.0);
-        assert!(progress.estimated_completion_time.is_none());
-    }
-
-    #[test]
-    fn progress_retry_is_exactly_once() {
-        let mut progress = HealProgress::new();
-        progress.set_total_baseline(1, 128);
-        progress.update_object_progress(1, 1, 0, 0, 128);
-        progress.update_object_progress(1, 1, 0, 0, 128);
-        assert_eq!(progress.objects_scanned, 1);
-        assert_eq!(progress.objects_healed, 1);
-        assert_eq!(progress.bytes_processed, 128);
-    }
-
-    #[test]
-    fn progress_never_triggers_cleanup_before_terminal_ledger_empty() {
-        let mut progress = HealProgress::new();
-        progress.progress_percentage = 100.0;
-        assert!(!progress.is_completed());
-        progress.mark_completed();
-        assert!(progress.is_completed());
-    }
-
-    #[test]
-    fn progress_counter_overflow_is_marked_unknown_without_aborting_completed_heal() {
-        let mut progress = HealProgress::new();
-        progress.update_object_progress(u64::MAX, u64::MAX, 1, 0, 0);
-        assert_eq!(progress.progress_state, HealProgressState::Unknown);
-        progress.mark_completed();
-        assert!(progress.is_completed());
-        assert_eq!(progress.progress_state, HealProgressState::Unknown);
-
-        let aggregate = aggregate_heal_progress([progress]).expect("progress should aggregate");
-        assert!(aggregate.ledger_complete);
-        assert!(aggregate.counter_unknown);
-        assert_eq!(aggregate.progress_state, HealProgressState::Unknown);
-        assert_eq!(aggregate.progress_percentage, 100.0);
-    }
-
-    #[test]
-    fn aggregate_rejects_mixed_baseline_generations() {
-        let progress = |generation| HealProgress {
-            kind: HealProgressKind::ObjectSweep,
-            objects_scanned: 5,
-            objects_total_count: 10,
-            progress_state: HealProgressState::Running,
-            baseline_generation: Some(generation),
-            baseline_known: true,
-            ..Default::default()
-        };
-
-        let aggregate = aggregate_heal_progress([progress(1), progress(2)]).expect("progress should aggregate");
-        assert!(!aggregate.baseline_known);
-        assert_eq!(aggregate.baseline_generation, None);
-        assert_eq!(aggregate.progress_state, HealProgressState::Indeterminate);
-        assert_eq!(aggregate.progress_percentage, 0.0);
-    }
-
-    #[test]
-    fn aggregate_accepts_multiple_sets_from_one_snapshot_generation() {
-        let progress = |objects_scanned| HealProgress {
-            kind: HealProgressKind::ObjectSweep,
-            objects_scanned,
-            objects_total_count: 10,
-            progress_state: HealProgressState::Running,
-            baseline_generation: Some(7),
-            baseline_known: true,
-            ..Default::default()
-        };
-
-        let aggregate = aggregate_heal_progress([progress(5), progress(3)]).expect("progress should aggregate");
-        assert!(aggregate.baseline_known);
-        assert_eq!(aggregate.baseline_generation, Some(7));
-    }
-
-    #[test]
-    fn stage_updates_do_not_double_count_object_outcomes() {
-        let mut progress = HealProgress::new();
-        progress.update_object_progress(2, 1, 0, 1, 256);
-        progress.update_stage(3, 4);
-        assert_eq!(progress.kind, HealProgressKind::ObjectSweep);
-        assert_eq!(progress.objects_scanned, 2);
-        assert_eq!(progress.objects_healed, 1);
-        assert_eq!(progress.skipped_objects, 1);
-    }
-
     #[test]
     fn test_heal_statistics_new() {
         let stats = HealStatistics::new();

crates/heal/src/heal/resume.rs

@@ -31,7 +31,7 @@ mod checkpoint;
 mod replacement;
 mod utils;
 
-pub use checkpoint::{CheckpointManager, CheckpointObjectOutcome, CheckpointObjectOutcomeRecord, ResumeCheckpoint};
+pub use checkpoint::{CheckpointManager, ResumeCheckpoint};
 pub(crate) use replacement::replacement_target_identities_match;
 use replacement::replacement_targets_match_identities;
 pub use replacement::{
@@ -340,12 +340,6 @@ pub struct ResumeState {
     pub failed_objects: u64,
     /// skipped objects
     pub skipped_objects: u64,
-    /// Terminal versions skipped because they were newer than the heal start.
-    #[serde(default)]
-    pub skipped_new_versions: u64,
-    /// Terminal versions handed to lifecycle expiry.
-    #[serde(default)]
-    pub skipped_ilm_expired: u64,
     /// current bucket
     pub current_bucket: Option<String>,
     /// current object
@@ -360,24 +354,6 @@ pub struct ResumeState {
     pub retry_count: u32,
     /// max retries
     pub max_retries: u32,
-    /// Bytes accounted by the object ledger; additive for old snapshots.
-    #[serde(default)]
-    pub processed_bytes: u64,
-    /// Total bytes from a complete usage snapshot, when available.
-    #[serde(default)]
-    pub total_bytes: u64,
-    /// Generation of the usage snapshot used for the baseline.
-    #[serde(default)]
-    pub baseline_generation: Option<u64>,
-    /// Whether the usage baseline is known. Missing in old snapshots means
-    /// indeterminate rather than a measured zero baseline.
-    #[serde(default)]
-    pub baseline_known: bool,
-    /// Persistent telemetry fence for counter/byte overflow or corruption.
-    /// It must survive a restart so a saturated snapshot is never presented as
-    /// a measured percentage on the next resume.
-    #[serde(default)]
-    pub counter_unknown: bool,
 }
 
 impl ResumeState {
@@ -401,8 +377,6 @@ impl ResumeState {
             successful_objects: 0,
             failed_objects: 0,
             skipped_objects: 0,
-            skipped_new_versions: 0,
-            skipped_ilm_expired: 0,
             current_bucket: None,
             current_object: None,
             completed_buckets: Vec::new(),
@@ -410,11 +384,6 @@ impl ResumeState {
             error_message: None,
             retry_count: 0,
             max_retries: 3,
-            processed_bytes: 0,
-            total_bytes: 0,
-            baseline_generation: None,
-            baseline_known: false,
-            counter_unknown: false,
         }
     }
 
@@ -443,39 +412,6 @@ impl ResumeState {
         self.last_update = SystemTime::now().duration_since(UNIX_EPOCH).unwrap_or_default().as_secs();
     }
 
-    pub fn update_progress_with_bytes(
-        &mut self,
-        processed: u64,
-        successful: u64,
-        failed: u64,
-        skipped: u64,
-        processed_bytes: u64,
-    ) {
-        self.update_progress(processed, successful, failed, skipped);
-        self.processed_bytes = processed_bytes;
-    }
-
-    pub fn set_skipped_version_counts(&mut self, new_versions: u64, ilm_expired: u64) {
-        self.skipped_new_versions = new_versions;
-        self.skipped_ilm_expired = ilm_expired;
-        self.last_update = SystemTime::now().duration_since(UNIX_EPOCH).unwrap_or_default().as_secs();
-    }
-
-    pub fn set_progress_baseline(&mut self, total_objects: u64, total_bytes: u64, generation: Option<u64>) {
-        self.total_objects = total_objects;
-        self.total_bytes = total_bytes;
-        self.baseline_generation = generation;
-        // This method is called only after a complete usage snapshot has been
-        // validated.  A complete but empty snapshot is still a known baseline.
-        self.baseline_known = true;
-        self.last_update = SystemTime::now().duration_since(UNIX_EPOCH).unwrap_or_default().as_secs();
-    }
-
-    pub fn mark_counter_unknown(&mut self) {
-        self.counter_unknown = true;
-        self.last_update = SystemTime::now().duration_since(UNIX_EPOCH).unwrap_or_default().as_secs();
-    }
-
     pub fn set_current_item(&mut self, bucket: Option<String>, object: Option<String>) {
         self.current_bucket = bucket;
         self.current_object = object;
@@ -501,7 +437,6 @@ impl ResumeState {
         if let Some(pos) = self.pending_buckets.iter().position(|b| b == bucket) {
             self.pending_buckets.remove(pos);
         }
-        self.resume_cursor = None;
         self.last_update = SystemTime::now().duration_since(UNIX_EPOCH).unwrap_or_default().as_secs();
     }
 
@@ -519,10 +454,6 @@ impl ResumeState {
         self.successful_objects = 0;
         self.failed_objects = 0;
         self.skipped_objects = 0;
-        self.skipped_new_versions = 0;
-        self.skipped_ilm_expired = 0;
-        self.processed_bytes = 0;
-        self.counter_unknown = false;
         self.completed = false;
         // A retry re-scans every bucket from the beginning, so the version
         // cursor must be cleared too — otherwise the retry would resume mid-scan.
@@ -545,28 +476,14 @@ impl ResumeState {
     }
 
     pub fn get_progress_percentage(&self) -> f64 {
-        if self.completed {
-            return 100.0;
-        }
-        if self.counter_unknown {
-            return 0.0;
-        }
-        if !self.baseline_known {
-            return 0.0;
-        }
-        if self.total_bytes > 0 {
-            return ((self.processed_bytes as f64 / self.total_bytes as f64) * 100.0).min(99.999);
-        }
         if self.total_objects == 0 {
             return 0.0;
         }
-        ((self.processed_objects as f64 / self.total_objects as f64) * 100.0).min(99.999)
+        (self.processed_objects as f64 / self.total_objects as f64) * 100.0
     }
 
     pub fn get_success_rate(&self) -> f64 {
-        let Some(total) = self.successful_objects.checked_add(self.failed_objects) else {
-            return 0.0;
-        };
+        let total = self.successful_objects + self.failed_objects;
         if total == 0 {
             return 0.0;
         }
@@ -837,14 +754,6 @@ impl ResumeManager {
             state.successful_objects = 0;
             state.failed_objects = 0;
             state.skipped_objects = 0;
-            state.skipped_new_versions = 0;
-            state.skipped_ilm_expired = 0;
-            state.processed_bytes = 0;
-            state.total_objects = 0;
-            state.total_bytes = 0;
-            state.baseline_generation = None;
-            state.baseline_known = false;
-            state.counter_unknown = false;
             state.completed = false;
             state.completed_buckets.clear();
             state.schema_version = CURRENT_RESUME_SCHEMA;
@@ -929,41 +838,6 @@ impl ResumeManager {
         self.save_state_throttled().await
     }
 
-    pub async fn update_progress_with_bytes(
-        &self,
-        processed: u64,
-        successful: u64,
-        failed: u64,
-        skipped: u64,
-        processed_bytes: u64,
-    ) -> Result<()> {
-        let mut state = self.state.write().await;
-        state.update_progress_with_bytes(processed, successful, failed, skipped, processed_bytes);
-        drop(state);
-        self.save_state_throttled().await
-    }
-
-    pub async fn set_progress_baseline(&self, total_objects: u64, total_bytes: u64, generation: Option<u64>) -> Result<()> {
-        let mut state = self.state.write().await;
-        state.set_progress_baseline(total_objects, total_bytes, generation);
-        drop(state);
-        self.save_state_throttled().await
-    }
-
-    pub async fn mark_counter_unknown(&self) -> Result<()> {
-        let mut state = self.state.write().await;
-        state.mark_counter_unknown();
-        drop(state);
-        self.save_state().await
-    }
-
-    pub async fn set_skipped_version_counts(&self, new_versions: u64, ilm_expired: u64) -> Result<()> {
-        let mut state = self.state.write().await;
-        state.set_skipped_version_counts(new_versions, ilm_expired);
-        drop(state);
-        self.save_state_throttled().await
-    }
-
     /// Set current item. Called once per healed object, so persistence is
     /// throttled: the in-memory state always updates, but the snapshot is only
     /// written every `PERSIST_EVERY_MUTATIONS` calls or `PERSIST_INTERVAL`.
@@ -1008,7 +882,7 @@ impl ResumeManager {
         let mut state = self.state.write().await;
         state.complete_bucket(bucket);
         drop(state);
-        self.save_state().await
+        self.save_state_throttled().await
     }
 
     /// mark task completed

crates/heal/src/heal/resume/checkpoint.rs

@@ -29,30 +29,10 @@ use super::{
 
 const EVENT_HEAL_CHECKPOINT_STATE: &str = "heal_checkpoint_state";
 
-/// Current on-disk schema version for `ResumeCheckpoint`. Schema 5 could
-/// persist dedup identities without the aggregate counters needed to restore
-/// them safely, so stale checkpoints are discarded and replayed.
-pub(super) const CURRENT_CHECKPOINT_SCHEMA: u32 = 6;
-
-#[derive(Debug, Clone, Copy)]
-pub enum CheckpointObjectOutcome {
-    Processed,
-    Failed,
-    Skipped,
-}
-
-#[derive(Debug)]
-pub struct CheckpointObjectOutcomeRecord {
-    pub object: String,
-    pub outcome: CheckpointObjectOutcome,
-    pub successful: u64,
-    pub failed: u64,
-    pub skipped: u64,
-    pub bytes: u64,
-    pub skipped_new_versions: u64,
-    pub skipped_ilm_expired: u64,
-    pub counter_unknown: bool,
-}
+/// Current on-disk schema version for `ResumeCheckpoint`. Same rationale as
+/// `CURRENT_RESUME_SCHEMA`: pre-per-version dedup identities are not comparable
+/// to the new `compose_key` identities, so a stale checkpoint is discarded.
+pub(super) const CURRENT_CHECKPOINT_SCHEMA: u32 = 5;
 
 /// resume checkpoint
 #[derive(Debug, Clone, Serialize, Deserialize)]
@@ -77,30 +57,6 @@ pub struct ResumeCheckpoint {
     pub failed_objects: HashSet<String>,
     /// skipped objects
     pub skipped_objects: HashSet<String>,
-    /// Aggregate object ledger counters restored alongside the dedup sets.
-    #[serde(default)]
-    pub successful_objects: u64,
-    #[serde(default)]
-    pub failed_object_count: u64,
-    #[serde(default)]
-    pub skipped_object_count: u64,
-    #[serde(default)]
-    pub skipped_new_versions: u64,
-    #[serde(default)]
-    pub skipped_ilm_expired: u64,
-    #[serde(default)]
-    pub processed_bytes: u64,
-    #[serde(default)]
-    pub total_objects: u64,
-    #[serde(default)]
-    pub total_bytes: u64,
-    #[serde(default)]
-    pub baseline_generation: Option<u64>,
-    #[serde(default)]
-    pub baseline_known: bool,
-    /// Persistent telemetry fence for counter/byte overflow or corruption.
-    #[serde(default)]
-    pub counter_unknown: bool,
 }
 
 impl ResumeCheckpoint {
@@ -114,17 +70,6 @@ impl ResumeCheckpoint {
             processed_objects: HashSet::new(),
             failed_objects: HashSet::new(),
             skipped_objects: HashSet::new(),
-            successful_objects: 0,
-            failed_object_count: 0,
-            skipped_object_count: 0,
-            skipped_new_versions: 0,
-            skipped_ilm_expired: 0,
-            processed_bytes: 0,
-            total_objects: 0,
-            total_bytes: 0,
-            baseline_generation: None,
-            baseline_known: false,
-            counter_unknown: false,
         }
     }
 
@@ -146,34 +91,6 @@ impl ResumeCheckpoint {
         self.skipped_objects.insert(object);
     }
 
-    pub fn update_progress(&mut self, successful: u64, failed: u64, skipped: u64, bytes: u64) {
-        self.successful_objects = successful;
-        self.failed_object_count = failed;
-        self.skipped_object_count = skipped;
-        self.processed_bytes = bytes;
-    }
-
-    pub fn set_progress_baseline(&mut self, total_objects: u64, total_bytes: u64, generation: Option<u64>) {
-        self.total_objects = total_objects;
-        self.total_bytes = total_bytes;
-        self.baseline_generation = generation;
-        // The caller has already validated that this is a complete snapshot;
-        // preserve the distinction between a known empty scope and an old
-        // checkpoint that omitted all baseline fields.
-        self.baseline_known = true;
-    }
-
-    pub fn mark_counter_unknown(&mut self) {
-        self.counter_unknown = true;
-        self.checkpoint_time = SystemTime::now().duration_since(UNIX_EPOCH).unwrap_or_default().as_secs();
-    }
-
-    pub fn set_skipped_version_counts(&mut self, new_versions: u64, ilm_expired: u64) {
-        self.skipped_new_versions = new_versions;
-        self.skipped_ilm_expired = ilm_expired;
-        self.checkpoint_time = SystemTime::now().duration_since(UNIX_EPOCH).unwrap_or_default().as_secs();
-    }
-
     /// Advance past a fully-processed page: objects below `object_index` are
     /// skipped by position on resume, so the per-object sets no longer need
     /// their entries and would otherwise grow with the whole bucket.
@@ -190,17 +107,6 @@ impl ResumeCheckpoint {
         self.update_position(0, 0);
         self.processed_objects.clear();
         self.skipped_objects.clear();
-        self.successful_objects = 0;
-        self.failed_object_count = 0;
-        self.skipped_object_count = 0;
-        self.skipped_new_versions = 0;
-        self.skipped_ilm_expired = 0;
-        self.processed_bytes = 0;
-        self.total_objects = 0;
-        self.total_bytes = 0;
-        self.baseline_generation = None;
-        self.baseline_known = false;
-        self.counter_unknown = false;
         self.failed_objects.clear();
     }
 }
@@ -252,9 +158,10 @@ impl CheckpointManager {
             });
         }
 
-        // Older checkpoints can contain identities that are not comparable to
-        // the current keys or lack their corresponding aggregate counters.
-        // Discard the stale sets and position so the scan restarts cleanly.
+        // A checkpoint from an older schema stored latest-only dedup identities
+        // that are not comparable to the new per-version `compose_key`
+        // identities. Discard the stale sets and position, then stamp the
+        // current schema so the scan restarts cleanly.
         if checkpoint.schema_version > CURRENT_CHECKPOINT_SCHEMA {
             return Err(Error::TaskExecutionFailed {
                 message: format!(
@@ -278,17 +185,6 @@ impl CheckpointManager {
             checkpoint.processed_objects.clear();
             checkpoint.failed_objects.clear();
             checkpoint.skipped_objects.clear();
-            checkpoint.successful_objects = 0;
-            checkpoint.failed_object_count = 0;
-            checkpoint.skipped_object_count = 0;
-            checkpoint.skipped_new_versions = 0;
-            checkpoint.skipped_ilm_expired = 0;
-            checkpoint.processed_bytes = 0;
-            checkpoint.total_objects = 0;
-            checkpoint.total_bytes = 0;
-            checkpoint.baseline_generation = None;
-            checkpoint.baseline_known = false;
-            checkpoint.counter_unknown = false;
             checkpoint.current_bucket_index = 0;
             checkpoint.current_object_index = 0;
             checkpoint.schema_version = CURRENT_CHECKPOINT_SCHEMA;
@@ -329,7 +225,7 @@ impl CheckpointManager {
         self.save_checkpoint_throttled().await
     }
 
-    /// Persist a completed page position while retaining its identities.
+    /// Advance past a completed page and prune the per-object sets, then persist.
     pub async fn complete_page(&self, bucket_index: usize, object_index: usize) -> Result<()> {
         let mut checkpoint = self.checkpoint.write().await;
         checkpoint.complete_page(bucket_index, object_index);
@@ -337,35 +233,6 @@ impl CheckpointManager {
         self.save_checkpoint_throttled().await
     }
 
-    /// Persist the page position while retaining identities until the resume
-    /// cursor is durable.
-    pub async fn advance_page(&self, bucket_index: usize, object_index: usize) -> Result<()> {
-        let mut checkpoint = self.checkpoint.write().await;
-        checkpoint.update_position(bucket_index, object_index);
-        drop(checkpoint);
-        self.save_checkpoint().await
-    }
-
-    /// Remove the previous page's dedup identities only after its resume cursor
-    /// has been durably exposed.
-    pub async fn prune_completed_page(&self) -> Result<()> {
-        let mut checkpoint = self.checkpoint.write().await;
-        checkpoint.processed_objects.clear();
-        checkpoint.skipped_objects.clear();
-        checkpoint.failed_objects.clear();
-        drop(checkpoint);
-        self.save_checkpoint().await
-    }
-
-    /// Advance to the next bucket and clear the final page identities after the
-    /// resume state has durably recorded the completed bucket.
-    pub async fn complete_bucket(&self, next_bucket_index: usize) -> Result<()> {
-        let mut checkpoint = self.checkpoint.write().await;
-        checkpoint.complete_page(next_bucket_index, 0);
-        drop(checkpoint);
-        self.save_checkpoint().await
-    }
-
     /// Reset the checkpoint to the start of the scan for a retry, then persist.
     pub async fn reset_for_retry(&self) -> Result<()> {
         let mut checkpoint = self.checkpoint.write().await;
@@ -400,62 +267,6 @@ impl CheckpointManager {
         self.save_checkpoint_if_due().await
     }
 
-    /// Atomically persist an object's dedup identity with its aggregate result.
-    pub async fn record_object_outcome(&self, record: CheckpointObjectOutcomeRecord) -> Result<()> {
-        let CheckpointObjectOutcomeRecord {
-            object,
-            outcome,
-            successful,
-            failed,
-            skipped,
-            bytes,
-            skipped_new_versions,
-            skipped_ilm_expired,
-            counter_unknown,
-        } = record;
-        let mut checkpoint = self.checkpoint.write().await;
-        match outcome {
-            CheckpointObjectOutcome::Processed => checkpoint.add_processed_object(object),
-            CheckpointObjectOutcome::Failed => checkpoint.add_failed_object(object),
-            CheckpointObjectOutcome::Skipped => checkpoint.add_skipped_object(object),
-        }
-        checkpoint.update_progress(successful, failed, skipped, bytes);
-        checkpoint.set_skipped_version_counts(skipped_new_versions, skipped_ilm_expired);
-        if counter_unknown {
-            checkpoint.mark_counter_unknown();
-        }
-        drop(checkpoint);
-        self.save_checkpoint_if_due().await
-    }
-
-    pub async fn update_progress(&self, successful: u64, failed: u64, skipped: u64, bytes: u64) -> Result<()> {
-        let mut checkpoint = self.checkpoint.write().await;
-        checkpoint.update_progress(successful, failed, skipped, bytes);
-        drop(checkpoint);
-        self.save_checkpoint_if_due().await
-    }
-
-    pub async fn set_progress_baseline(&self, total_objects: u64, total_bytes: u64, generation: Option<u64>) -> Result<()> {
-        let mut checkpoint = self.checkpoint.write().await;
-        checkpoint.set_progress_baseline(total_objects, total_bytes, generation);
-        drop(checkpoint);
-        self.save_checkpoint_throttled().await
-    }
-
-    pub async fn mark_counter_unknown(&self) -> Result<()> {
-        let mut checkpoint = self.checkpoint.write().await;
-        checkpoint.mark_counter_unknown();
-        drop(checkpoint);
-        self.save_checkpoint().await
-    }
-
-    pub async fn set_skipped_version_counts(&self, new_versions: u64, ilm_expired: u64) -> Result<()> {
-        let mut checkpoint = self.checkpoint.write().await;
-        checkpoint.set_skipped_version_counts(new_versions, ilm_expired);
-        drop(checkpoint);
-        self.save_checkpoint_throttled().await
-    }
-
     async fn save_checkpoint_if_due(&self) -> Result<()> {
         let should_save = self.throttle.lock().map(|mut throttle| throttle.record()).unwrap_or(true);
         if !should_save {

crates/heal/src/heal/resume/tests.rs

@@ -1296,7 +1296,6 @@ async fn test_resume_state_progress() {
     assert_eq!(progress, 0.0); // total_objects is 0
 
     state.total_objects = 100;
-    state.baseline_known = true;
     let progress = state.get_progress_percentage();
     assert_eq!(progress, 10.0);
 }
@@ -1476,40 +1475,6 @@ fn test_checkpoint_object_sets_dedupe_and_prune() {
     assert!(checkpoint.failed_objects.is_empty());
 }
 
-#[tokio::test]
-async fn checkpoint_page_commit_keeps_ledger_until_cursor_is_durable() {
-    let (_temp_dir, disk) = schema_test_disk().await;
-    let task_id = ResumeUtils::generate_task_id();
-    let checkpoint = CheckpointManager::new(disk.clone(), task_id.clone()).await.unwrap();
-
-    checkpoint
-        .record_object_outcome(CheckpointObjectOutcomeRecord {
-            object: "bucket/object:v1".to_string(),
-            outcome: CheckpointObjectOutcome::Processed,
-            successful: 1,
-            failed: 0,
-            skipped: 0,
-            bytes: 128,
-            skipped_new_versions: 0,
-            skipped_ilm_expired: 0,
-            counter_unknown: false,
-        })
-        .await
-        .unwrap();
-    checkpoint.advance_page(0, 1).await.unwrap();
-
-    let reloaded = CheckpointManager::load_from_disk(disk.clone(), &task_id).await.unwrap();
-    let snapshot = reloaded.get_checkpoint().await;
-    assert_eq!(snapshot.current_object_index, 1);
-    assert_eq!(snapshot.successful_objects, 1);
-    assert_eq!(snapshot.processed_bytes, 128);
-    assert!(snapshot.processed_objects.contains("bucket/object:v1"));
-
-    checkpoint.prune_completed_page().await.unwrap();
-    let reloaded = CheckpointManager::load_from_disk(disk, &task_id).await.unwrap();
-    assert!(reloaded.get_checkpoint().await.processed_objects.is_empty());
-}
-
 #[test]
 fn test_checkpoint_loads_legacy_vec_format() {
     // Checkpoints written before the HashSet migration stored the object
@@ -1603,14 +1568,14 @@ async fn test_resumestate_schema_v0_discarded_on_load() {
 }
 
 #[tokio::test]
-async fn test_checkpoint_schema_v5_discarded_on_load() {
+async fn test_checkpoint_schema_v4_discarded_on_load() {
     let (temp_dir, disk) = schema_test_disk().await;
 
-    // Schema v5 can persist failed identities without the aggregate counters
-    // that make those identities safe to deduplicate after an upgrade.
+    // The previous checkpoint schema is unsafe once its paired resume
+    // state is discarded: retaining either position would skip work.
     let task_id = "00000000-0000-4000-8000-000000000002";
     let legacy = r#"{
-            "schema_version": 5,
+            "schema_version": 4,
             "task_id": "00000000-0000-4000-8000-000000000002",
             "checkpoint_time": 1700000000,
             "current_bucket_index": 2,
@@ -1674,120 +1639,6 @@ async fn current_normal_resume_schema_preserves_progress() {
     temp_dir.close().expect("remove schema test directory");
 }
 
-#[test]
-fn progress_checkpoint_restores_bytes_and_generation() {
-    let mut checkpoint = ResumeCheckpoint::new("progress-checkpoint".to_string());
-    checkpoint.set_progress_baseline(9, 4096, Some(77));
-    checkpoint.update_progress(4, 1, 2, 2048);
-    checkpoint.set_skipped_version_counts(3, 1);
-    checkpoint.mark_counter_unknown();
-
-    let restored: ResumeCheckpoint =
-        serde_json::from_slice(&serde_json::to_vec(&checkpoint).expect("serialize checkpoint")).expect("deserialize checkpoint");
-    assert_eq!(restored.processed_bytes, 2048);
-    assert_eq!(restored.total_objects, 9);
-    assert_eq!(restored.total_bytes, 4096);
-    assert_eq!(restored.baseline_generation, Some(77));
-    assert!(restored.baseline_known);
-    assert_eq!(restored.skipped_new_versions, 3);
-    assert_eq!(restored.skipped_ilm_expired, 1);
-    assert!(restored.counter_unknown);
-}
-
-#[test]
-fn old_progress_schema_migrates_missing_fields_to_unknown() {
-    let state = ResumeState::new(
-        "legacy-progress".to_string(),
-        "erasure_set".to_string(),
-        "pool_0_set_0".to_string(),
-        Vec::new(),
-    );
-    let mut value = serde_json::to_value(state).expect("serialize legacy-compatible state");
-    let object = value.as_object_mut().expect("state must be an object");
-    for field in [
-        "processed_bytes",
-        "total_bytes",
-        "baseline_generation",
-        "baseline_known",
-        "skipped_new_versions",
-        "skipped_ilm_expired",
-    ] {
-        object.remove(field);
-    }
-    object.insert("total_objects".to_string(), serde_json::json!(10));
-    object.insert("processed_objects".to_string(), serde_json::json!(5));
-    let restored: ResumeState = serde_json::from_value(value).expect("deserialize old progress state");
-    assert_eq!(restored.processed_bytes, 0);
-    assert_eq!(restored.total_bytes, 0);
-    assert_eq!(restored.baseline_generation, None);
-    assert!(!restored.baseline_known, "missing baseline must remain unknown");
-    assert_eq!(restored.get_progress_percentage(), 0.0);
-    assert_eq!(restored.skipped_new_versions, 0);
-    assert_eq!(restored.skipped_ilm_expired, 0);
-}
-
-#[test]
-fn progress_counter_unknown_survives_resume_round_trip() {
-    let mut state = ResumeState::new(
-        "overflow-progress".to_string(),
-        "erasure_set".to_string(),
-        "pool_0_set_0".to_string(),
-        Vec::new(),
-    );
-    state.mark_counter_unknown();
-
-    let restored: ResumeState =
-        serde_json::from_slice(&serde_json::to_vec(&state).expect("serialize resume state")).expect("deserialize resume state");
-    assert!(restored.counter_unknown);
-}
-
-#[tokio::test]
-async fn checkpoint_progress_survives_a_torn_resume_summary_write() {
-    let (_temp_dir, disk) = schema_test_disk().await;
-    let task_id = ResumeUtils::generate_task_id();
-    let _resume = ResumeManager::new(
-        disk.clone(),
-        task_id.clone(),
-        "erasure_set".to_string(),
-        "pool_0_set_0".to_string(),
-        vec!["bucket".to_string()],
-    )
-    .await
-    .expect("resume state should persist");
-    let checkpoint = CheckpointManager::new(disk.clone(), task_id.clone())
-        .await
-        .expect("checkpoint should persist");
-
-    // This is the ordering used by the erasure-set loop: the checkpoint is
-    // durable before the summary write.  Stop here to model a crash in the
-    // inter-store window and verify that the recovery authority retains the
-    // telemetry fence and bytes.
-    checkpoint
-        .update_progress(3, 0, 0, 1024)
-        .await
-        .expect("checkpoint progress should persist");
-    checkpoint.mark_counter_unknown().await.expect("unknown fence should persist");
-    checkpoint
-        .update_position(0, 3)
-        .await
-        .expect("checkpoint position should persist");
-
-    let restored_checkpoint = CheckpointManager::load_from_disk(disk.clone(), &task_id)
-        .await
-        .expect("checkpoint should reload")
-        .get_checkpoint()
-        .await;
-    let restored_resume = ResumeManager::load_from_disk(disk, &task_id)
-        .await
-        .expect("resume summary should reload")
-        .get_state()
-        .await;
-    assert!(restored_checkpoint.counter_unknown);
-    assert_eq!(restored_checkpoint.processed_bytes, 1024);
-    assert_eq!(restored_checkpoint.current_object_index, 3);
-    assert!(!restored_resume.counter_unknown, "summary is intentionally the torn/older store");
-}
-
 #[tokio::test]
 async fn future_resume_and_checkpoint_schemas_are_rejected() {
     let (temp_dir, disk) = schema_test_disk().await;

crates/heal/src/heal/storage.rs

@@ -22,7 +22,6 @@ use serde::{Deserialize, Serialize};
 use std::sync::Arc;
 use tracing::{debug, error, warn};
 
-use super::progress::stable_generation;
 use super::storage_api::owner::{EcstoreHealLifecycleExpiryContext, ecstore_load_admin_data_usage_from_backend_cached};
 use super::storage_api::storage::{
     BucketInfo, BucketOperations, DiskSetSelector, HealOperations as _, ListOperations as _, ObjectIO as _,
@@ -35,9 +34,6 @@ pub use super::{HealObjectInfo, HealObjectOptions, HealPutObjReader};
 pub struct HealBucketUsageBaseline {
     pub objects_count: u64,
     pub bytes: u64,
-    /// Stable identity of the validated usage snapshot and selected scope.
-    /// `None` is retained for test/legacy providers that cannot expose one.
-    pub generation: Option<u64>,
 }
 
 pub struct HealLifecycleExpiryContext {
@@ -789,52 +785,11 @@ impl HealStorageAPI for ECStoreHealStorage {
         let mut baseline = HealBucketUsageBaseline::default();
         for bucket in buckets {
             if let Some(usage) = info.buckets_usage.get(bucket) {
-                baseline.objects_count = match baseline.objects_count.checked_add(usage.objects_count) {
-                    Some(total) => total,
-                    // A corrupt/overflowing usage snapshot is not a usable
-                    // denominator.  Leave progress indeterminate instead of
-                    // turning saturation into a plausible percentage.
-                    None => return Ok(None),
-                };
-                baseline.bytes = match baseline.bytes.checked_add(usage.size) {
-                    Some(total) => total,
-                    None => return Ok(None),
-                };
+                baseline.objects_count = baseline.objects_count.saturating_add(usage.objects_count);
+                baseline.bytes = baseline.bytes.saturating_add(usage.size);
             }
         }
 
-        let identity = info.snapshot_identity();
-        let mut canonical = Vec::new();
-        match identity.last_update {
-            Some(last_update) => {
-                canonical.push(1);
-                canonical.extend_from_slice(
-                    &last_update
-                        .duration_since(std::time::UNIX_EPOCH)
-                        .unwrap_or_default()
-                        .as_nanos()
-                        .to_be_bytes(),
-                );
-            }
-            None => canonical.push(0),
-        }
-        for value in [identity.scanner_cycle, identity.scanner_epoch] {
-            match value {
-                Some(value) => {
-                    canonical.push(1);
-                    canonical.extend_from_slice(&value.to_be_bytes());
-                }
-                None => canonical.push(0),
-            }
-        }
-        let mut scope = buckets.to_vec();
-        scope.sort_unstable();
-        for bucket in scope {
-            canonical.extend_from_slice(&(bucket.len() as u64).to_be_bytes());
-            canonical.extend_from_slice(bucket.as_bytes());
-        }
-        baseline.generation = Some(stable_generation(&[&canonical]));
-
         Ok(Some(baseline))
     }
 

crates/heal/src/heal/task.rs

@@ -649,7 +649,7 @@ impl HealTask {
 
         let mut progress = self.progress.write().await;
         progress.set_current_object(Some(format!("skipped: {bucket}/{object}")));
-        progress.update_stage(1, 1);
+        progress.update_progress(0, 1, 0, 0);
         Ok(())
     }
 
@@ -733,7 +733,7 @@ impl HealTask {
             "Heal object skipped for data usage cache after transient error"
         );
         let mut progress = self.progress.write().await;
-        progress.update_stage(3, 3);
+        progress.update_progress(3, 3, 0, 0);
         true
     }
 
@@ -757,7 +757,7 @@ impl HealTask {
         );
         let mut progress = self.progress.write().await;
         progress.set_current_object(Some(format!("skipped: {bucket}/{object}")));
-        progress.update_stage(4, 4);
+        progress.update_progress(4, 4, 0, 0);
         true
     }
 
@@ -831,10 +831,6 @@ impl HealTask {
 
         match &result {
             Ok(_) => {
-                // A stage can reach its final step before the durable resume
-                // ledger and cleanup fences commit. Publish terminal 100 only
-                // after the enclosing operation has returned success.
-                self.progress.write().await.mark_completed();
                 let mut status = self.status.write().await;
                 *status = HealTaskStatus::Completed;
                 demote_to_debug_when!(self.heal_type.is_per_object(), info, target: "rustfs::heal::task", {

crates/heal/src/heal/task/heal_bucket.rs

@@ -13,7 +13,6 @@
 // limitations under the License.
 /// bucket/cluster/prefix heal: the recursive bucket-objects sweep and the erasure-set usage baseline
 use super::*;
-use crate::heal::progress::{add_bytes, increment_counter, stable_generation};
 
 impl HealTask {
     pub(super) async fn heal_bucket(&self, bucket: &str) -> Result<()> {
@@ -33,7 +32,7 @@ impl HealTask {
         {
             let mut progress = self.progress.write().await;
             progress.set_current_object(Some(format!("bucket: {bucket}")));
-            progress.update_stage(0, 3);
+            progress.update_progress(0, 3, 0, 0);
         }
 
         // Step 1: Check if bucket exists
@@ -67,7 +66,7 @@ impl HealTask {
 
         {
             let mut progress = self.progress.write().await;
-            progress.update_stage(1, 3);
+            progress.update_progress(1, 3, 0, 0);
         }
 
         // Step 2: Perform bucket heal using ecstore
@@ -123,7 +122,7 @@ impl HealTask {
 
                 if !self.options.recursive {
                     let mut progress = self.progress.write().await;
-                    progress.update_stage(3, 3);
+                    progress.update_progress(3, 3, 0, 0);
                 }
                 Ok(())
             }
@@ -143,7 +142,7 @@ impl HealTask {
                 );
                 {
                     let mut progress = self.progress.write().await;
-                    progress.update_stage(3, 3);
+                    progress.update_progress(3, 3, 0, 0);
                 }
                 Err(Error::TaskExecutionFailed {
                     message: format!("Failed to heal bucket {bucket}: {e}"),
@@ -246,7 +245,6 @@ impl HealTask {
         let mut scanned = 0u64;
         let mut healed = 0u64;
         let mut failed = 0u64;
-        let mut skipped = 0u64;
         let mut retryable_failed = 0u64;
         let mut permanent_failed = 0u64;
         let mut bytes = 0u64;
@@ -288,14 +286,16 @@ impl HealTask {
                 let mut retry = Vec::with_capacity(pending.len());
                 for item in pending {
                     self.check_control_flags().await?;
-                    let mut telemetry_unknown = false;
                     let object = item.name.as_str();
+                    if retry_attempt == 0 {
+                        scanned = scanned.saturating_add(1);
+                    }
                     {
                         let mut progress = self.progress.write().await;
                         progress.set_current_object(Some(format!("{bucket}/{object}")));
+                        progress.update_progress(scanned, healed, failed, bytes);
                     }
 
-                    let mut terminal_outcome = true;
                     let error = match self
                         .await_with_control(
                             self.storage
@@ -304,13 +304,13 @@ impl HealTask {
                         .await
                     {
                         Ok((result, None)) => {
-                            telemetry_unknown |= !increment_counter(&mut healed);
-                            telemetry_unknown |= !add_bytes(&mut bytes, u64::try_from(result.object_size).unwrap_or(u64::MAX));
+                            healed = healed.saturating_add(1);
+                            bytes = bytes.saturating_add(u64::try_from(result.object_size).unwrap_or_default());
                             self.record_result_item(result).await;
                             None
                         }
                         Ok((_, Some(err))) if is_missing_object_dir_heal_result(object, &err) => {
-                            telemetry_unknown |= !increment_counter(&mut healed);
+                            healed = healed.saturating_add(1);
                             debug!(
                                 target: "rustfs::heal::task",
                                 event = EVENT_HEAL_BUCKET_RESULT,
@@ -329,7 +329,6 @@ impl HealTask {
 
                     if let Some(err) = error {
                         if Self::should_skip_data_usage_cache_heal_error(bucket, object, &err) {
-                            telemetry_unknown |= !increment_counter(&mut skipped);
                             warn!(
                                 target: "rustfs::heal::task",
                                 event = EVENT_HEAL_BUCKET_RESULT,
@@ -343,7 +342,6 @@ impl HealTask {
                                 "Heal bucket object repair skipped due to transient metadata error"
                             );
                         } else if err.is_recoverable_heal() && retry_attempt < MAX_BUCKET_OBJECT_HEAL_RETRIES {
-                            terminal_outcome = false;
                             debug!(
                                 target: "rustfs::heal::task",
                                 event = EVENT_HEAL_BUCKET_RESULT,
@@ -359,7 +357,7 @@ impl HealTask {
                             );
                             retry.push(item);
                         } else {
-                            telemetry_unknown |= !increment_counter(&mut failed);
+                            failed = failed.saturating_add(1);
                             if err.is_recoverable_heal() {
                                 retryable_failed = retryable_failed.saturating_add(1);
                             } else {
@@ -385,19 +383,8 @@ impl HealTask {
                         }
                     }
 
-                    if terminal_outcome {
-                        telemetry_unknown |= !increment_counter(&mut scanned);
-                    }
-
-                    if !terminal_outcome {
-                        continue;
-                    }
-
                     let mut progress = self.progress.write().await;
-                    progress.update_object_progress(scanned, healed, failed, skipped, bytes);
-                    if telemetry_unknown {
-                        progress.mark_unknown();
-                    }
+                    progress.update_progress(scanned, healed, failed, bytes);
                 }
                 pending = retry;
                 retry_attempt = retry_attempt.saturating_add(1);
@@ -444,10 +431,7 @@ impl HealTask {
         Ok(())
     }
 
-    pub(super) async fn apply_erasure_set_usage_baseline(&self, buckets: &[String], _set_disk_id: &str) -> Result<()> {
-        if matches!(self.options.scan_mode, HealScanMode::Deep) || matches!(self.source, HealRequestSource::AutoHeal) {
-            return Ok(());
-        }
+    pub(super) async fn apply_erasure_set_usage_baseline(&self, buckets: &[String]) -> Result<()> {
         let baseline = match self
             .await_with_control(self.storage.erasure_set_usage_baseline(buckets))
             .await
@@ -458,18 +442,9 @@ impl HealTask {
             Err(_) => return Ok(()),
         };
 
-        let HealBucketUsageBaseline {
-            objects_count,
-            bytes,
-            generation,
-        } = baseline;
-        let generation = generation.map(|snapshot_generation| stable_generation(&[&snapshot_generation.to_be_bytes()]));
+        let HealBucketUsageBaseline { objects_count, bytes } = baseline;
         let mut progress = self.progress.write().await;
-        if let Some(generation) = generation {
-            progress.set_total_baseline_with_generation(objects_count, bytes, generation);
-        } else {
-            progress.set_total_baseline(objects_count, bytes);
-        }
+        progress.set_total_baseline(objects_count, bytes);
         Ok(())
     }
 }

crates/heal/src/heal/task/heal_erasure_set.rs

@@ -32,7 +32,7 @@ impl HealTask {
         {
             let mut progress = self.progress.write().await;
             progress.set_current_object(Some(format!("erasure_set: {} ({} buckets)", set_disk_id, buckets.len())));
-            progress.update_stage(0, 4);
+            progress.update_progress(0, 4, 0, 0);
         }
 
         let is_auto_replacement = matches!(self.source, HealRequestSource::AutoHeal) && !self.heal_endpoints.is_empty();
@@ -158,7 +158,7 @@ impl HealTask {
             None
         };
 
-        self.apply_erasure_set_usage_baseline(&buckets, &set_disk_id).await?;
+        self.apply_erasure_set_usage_baseline(&buckets).await?;
 
         let healing_marker = format!("{set_disk_id}:{}", self.id);
         if let Some((disk, resume_manager, _)) = replacement_resume.as_ref() {
@@ -248,7 +248,7 @@ impl HealTask {
                         );
                         {
                             let mut progress = self.progress.write().await;
-                            progress.update_stage(4, 4);
+                            progress.update_progress(4, 4, 0, 0);
                         }
                         return Err(Error::TaskExecutionFailed {
                             message: format!("Failed to heal disk format for {set_disk_id}: {error}"),
@@ -304,7 +304,7 @@ impl HealTask {
                 );
                 {
                     let mut progress = self.progress.write().await;
-                    progress.update_stage(4, 4);
+                    progress.update_progress(4, 4, 0, 0);
                 }
                 return Err(Error::TaskExecutionFailed {
                     message: format!("Failed to heal disk format for {set_disk_id}: {e}"),
@@ -314,7 +314,7 @@ impl HealTask {
 
         {
             let mut progress = self.progress.write().await;
-            progress.update_stage(1, 4);
+            progress.update_progress(1, 4, 0, 0);
         }
 
         // The rebuilt disks are formatted now: mark them as healing so
@@ -343,7 +343,7 @@ impl HealTask {
 
         {
             let mut progress = self.progress.write().await;
-            progress.update_stage(2, 4);
+            progress.update_progress(2, 4, 0, 0);
         }
 
         // Step 3: Heal bucket structure
@@ -427,7 +427,7 @@ impl HealTask {
 
         {
             let mut progress = self.progress.write().await;
-            progress.update_stage(3, 4);
+            progress.update_progress(3, 4, 0, 0);
         }
 
         // Step 4: Execute erasure set heal with resume
@@ -470,7 +470,9 @@ impl HealTask {
         };
 
         {
-            self.progress.write().await.update_stage(4, 4);
+            let mut progress = self.progress.write().await;
+            let bytes_processed = progress.bytes_processed;
+            progress.update_progress(4, 4, 0, bytes_processed);
         }
 
         match result {

crates/heal/src/heal/task/heal_metadata.rs

@@ -32,7 +32,7 @@ impl HealTask {
         {
             let mut progress = self.progress.write().await;
             progress.set_current_object(Some(format!("metadata: {bucket}/{object}")));
-            progress.update_stage(0, 3);
+            progress.update_progress(0, 3, 0, 0);
         }
 
         // Step 1: Check if object exists
@@ -74,7 +74,7 @@ impl HealTask {
 
         {
             let mut progress = self.progress.write().await;
-            progress.update_stage(1, 3);
+            progress.update_progress(1, 3, 0, 0);
         }
 
         // Step 2: Perform metadata heal using ecstore
@@ -122,7 +122,7 @@ impl HealTask {
                     );
                     {
                         let mut progress = self.progress.write().await;
-                        progress.update_stage(3, 3);
+                        progress.update_progress(3, 3, 0, 0);
                     }
                     return Err(Error::TaskExecutionFailed {
                         message: format!("Failed to heal metadata {bucket}/{object}: {e}"),
@@ -145,7 +145,7 @@ impl HealTask {
 
                 {
                     let mut progress = self.progress.write().await;
-                    progress.update_stage(3, 3);
+                    progress.update_progress(3, 3, 0, 0);
                 }
                 self.record_result_item(result).await;
                 Ok(())
@@ -167,7 +167,7 @@ impl HealTask {
                 );
                 {
                     let mut progress = self.progress.write().await;
-                    progress.update_stage(3, 3);
+                    progress.update_progress(3, 3, 0, 0);
                 }
                 Err(Error::TaskExecutionFailed {
                     message: format!("Failed to heal metadata {bucket}/{object}: {e}"),
@@ -194,7 +194,7 @@ impl HealTask {
         {
             let mut progress = self.progress.write().await;
             progress.set_current_object(Some(format!("ec_decode: {bucket}/{object}")));
-            progress.update_stage(0, 3);
+            progress.update_progress(0, 3, 0, 0);
         }
 
         // Step 1: Check if object exists
@@ -236,7 +236,7 @@ impl HealTask {
 
         {
             let mut progress = self.progress.write().await;
-            progress.update_stage(1, 3);
+            progress.update_progress(1, 3, 0, 0);
         }
 
         // Step 2: Perform EC decode heal using ecstore
@@ -284,7 +284,7 @@ impl HealTask {
                     );
                     {
                         let mut progress = self.progress.write().await;
-                        progress.update_stage(3, 3);
+                        progress.update_progress(3, 3, 0, 0);
                     }
                     return Err(Error::TaskExecutionFailed {
                         message: format!("Failed to heal EC decode {bucket}/{object}: {e}"),
@@ -309,7 +309,7 @@ impl HealTask {
 
                 {
                     let mut progress = self.progress.write().await;
-                    progress.update_object_progress(1, 1, 0, 0, object_size);
+                    progress.update_progress(3, 3, 0, object_size);
                 }
                 self.record_result_item(result).await;
                 Ok(())
@@ -331,7 +331,7 @@ impl HealTask {
                 );
                 {
                     let mut progress = self.progress.write().await;
-                    progress.update_stage(3, 3);
+                    progress.update_progress(3, 3, 0, 0);
                 }
                 Err(Error::TaskExecutionFailed {
                     message: format!("Failed to heal EC decode {bucket}/{object}: {e}"),

crates/heal/src/heal/task/heal_object.rs

@@ -36,7 +36,7 @@ impl HealTask {
         {
             let mut progress = self.progress.write().await;
             progress.set_current_object(Some(format!("{bucket}/{object}")));
-            progress.update_stage(0, 4);
+            progress.update_progress(0, 4, 0, 0);
         }
 
         // Step 1: Check if object exists and get metadata
@@ -132,7 +132,7 @@ impl HealTask {
 
         {
             let mut progress = self.progress.write().await;
-            progress.update_stage(1, 3);
+            progress.update_progress(1, 3, 0, 0);
         }
 
         // Step 2: directly call ecstore to perform heal
@@ -187,7 +187,7 @@ impl HealTask {
                         );
                         {
                             let mut progress = self.progress.write().await;
-                            progress.update_stage(3, 3);
+                            progress.update_progress(3, 3, 0, 0);
                         }
                         return Ok(());
                     }
@@ -207,7 +207,7 @@ impl HealTask {
 
                     {
                         let mut progress = self.progress.write().await;
-                        progress.update_stage(3, 3);
+                        progress.update_progress(3, 3, 0, 0);
                     }
 
                     if Self::should_return_typed_heal_error(&e) {
@@ -249,7 +249,7 @@ impl HealTask {
 
                 {
                     let mut progress = self.progress.write().await;
-                    progress.update_object_progress(1, 1, 0, 0, object_size);
+                    progress.update_progress(3, 3, 0, object_size);
                 }
                 self.record_result_item(result).await;
                 Ok(())
@@ -275,7 +275,7 @@ impl HealTask {
                     );
                     {
                         let mut progress = self.progress.write().await;
-                        progress.update_stage(3, 3);
+                        progress.update_progress(3, 3, 0, 0);
                     }
                     return Ok(());
                 }
@@ -295,7 +295,7 @@ impl HealTask {
 
                 {
                     let mut progress = self.progress.write().await;
-                    progress.update_stage(3, 3);
+                    progress.update_progress(3, 3, 0, 0);
                 }
 
                 if Self::should_return_typed_heal_error(&e) {
@@ -414,7 +414,7 @@ impl HealTask {
 
                 {
                     let mut progress = self.progress.write().await;
-                    progress.update_object_progress(1, 1, 0, 0, object_size);
+                    progress.update_progress(4, 4, 0, object_size);
                 }
                 self.record_result_item(result).await;
                 Ok(())

crates/heal/src/heal/task/tests.rs

@@ -22,7 +22,6 @@ use std::sync::Mutex;
 use tempfile::TempDir;
 
 use super::super::storage_api::status::BucketInfo;
-use crate::heal::progress::HealProgressState;
 
 #[tokio::test]
 async fn retry_request_carries_remaining_timeout_budget() {
@@ -2125,7 +2124,6 @@ async fn erasure_set_heal_applies_usage_baseline_to_progress() {
         usage_baseline: Mutex::new(Some(HealBucketUsageBaseline {
             objects_count: 10,
             bytes: 8,
-            generation: Some(1),
         })),
         ..Default::default()
     });
@@ -2149,55 +2147,10 @@ async fn erasure_set_heal_applies_usage_baseline_to_progress() {
     let progress = task.get_progress().await;
     assert_eq!(progress.objects_total_count, 10);
     assert_eq!(progress.objects_total_size, 8);
-    assert!(progress.baseline_generation.is_some());
-    assert!(progress.baseline_known);
     assert_eq!(progress.bytes_processed, 2);
     assert!((progress.progress_percentage - 25.0).abs() < 0.001);
 }
 
-#[tokio::test]
-async fn erasure_set_disk_walk_keeps_cluster_usage_baseline_indeterminate() {
-    for (scan_mode, source) in [
-        (HealScanMode::Deep, HealRequestSource::Admin),
-        (HealScanMode::Normal, HealRequestSource::AutoHeal),
-    ] {
-        let temp = TempDir::new().expect("temporary directory should be created");
-        let disk = make_resume_disk(&temp).await;
-        let storage = Arc::new(MockStorage {
-            resume_disk: Mutex::new(Some(disk)),
-            usage_baseline: Mutex::new(Some(HealBucketUsageBaseline {
-                objects_count: 10,
-                bytes: 8,
-                generation: Some(1),
-            })),
-            ..Default::default()
-        });
-        let mut request = HealRequest::new(
-            HealType::ErasureSet {
-                buckets: vec!["bucket-a".to_string()],
-                set_disk_id: "pool_0_set_0".to_string(),
-            },
-            HealOptions {
-                scan_mode,
-                timeout: None,
-                ..Default::default()
-            },
-            HealPriority::Normal,
-        );
-        request.source = source;
-        let task = HealTask::from_request(request, storage);
-
-        task.heal_erasure_set(vec!["bucket-a".to_string()], "pool_0_set_0".to_string())
-            .await
-            .expect("erasure set heal should complete");
-
-        let progress = task.get_progress().await;
-        assert!(!progress.baseline_known);
-        assert_eq!(progress.baseline_generation, None);
-        assert_eq!(progress.progress_state, HealProgressState::Indeterminate);
-    }
-}
-
 #[tokio::test]
 async fn erasure_set_heal_ignores_usage_baseline_errors() {
     let temp = TempDir::new().expect("temporary directory should be created");

crates/heal/src/lib.rs

@@ -19,7 +19,7 @@ pub use error::{Error, Result};
 pub use heal::{
     HealManager, HealOperationsSnapshot, HealOptions, HealPriority, HealPriorityCounts, HealRequest, HealSourceCounts, HealType,
     channel::HealChannelProcessor,
-    progress::{HealProgress, aggregate_heal_progress},
+    progress::HealProgress,
     resume::{ReplacementRecoveryRecord, ReplacementRecoveryState, ResumeUtils},
 };
 use rustfs_concurrency::WorkloadAdmissionSnapshotProvider;

crates/protos/src/generated/proto_gen/node_service.rs

@@ -1215,10 +1215,7 @@ pub struct ScannerActivityResponse {
     pub dirty_usage_pending: bool,
 }
 #[derive(Clone, Copy, PartialEq, Eq, Hash, ::prost::Message)]
-pub struct BackgroundHealStatusRequest {
-    #[prost(uint32, tag = "1")]
-    pub protocol_version: u32,
-}
+pub struct BackgroundHealStatusRequest {}
 #[derive(Clone, PartialEq, Eq, Hash, ::prost::Message)]
 pub struct BackgroundHealStatusResponse {
     #[prost(bool, tag = "1")]

crates/protos/src/lib.rs

@@ -170,7 +170,6 @@ pub fn internode_rpc_max_message_size() -> usize {
 pub const HEAL_CONTROL_RPC_MAX_MESSAGE_SIZE: usize = heal_control::RESULT_MAX_SIZE + 1024;
 pub const HEAL_CONTROL_PROTOCOL_VERSION: u32 = 3;
 pub const DYNAMIC_CONFIG_PROTOCOL_VERSION: u32 = 1;
-pub const BACKGROUND_HEAL_STATUS_PROTOCOL_VERSION: u32 = 2;
 pub const HEAL_CONTROL_CAPABILITY_PROBE_PREFIX: &[u8] = b"rustfs-heal-control-capability-v3\0";
 pub const REMOTE_VERSION_STATE_CAPABILITY_PROBE_PREFIX: &[u8] = b"rustfs-tier-remote-version-state-capability-v1\0";
 pub const CROSS_POOL_FENCE_CAPABILITY_PROBE_PREFIX: &[u8] = b"rustfs-cross-pool-fence-capability-v1\0";
@@ -2347,28 +2346,10 @@ pub async fn evict_failed_connection_with_log_level(addr: &str, log_level: Conne
 #[cfg(test)]
 mod tests {
     use super::*;
-    use prost::Message as _;
     use std::sync::Mutex;
 
     static INTERNODE_RPC_MSGPACK_ONLY_ENV_LOCK: Mutex<()> = Mutex::new(());
 
-    #[derive(Clone, PartialEq, prost::Message)]
-    struct BackgroundHealStatusRequestV1 {}
-
-    #[test]
-    fn background_heal_status_request_remains_rolling_upgrade_compatible() {
-        let current = proto_gen::node_service::BackgroundHealStatusRequest {
-            protocol_version: BACKGROUND_HEAL_STATUS_PROTOCOL_VERSION,
-        };
-        let encoded = current.encode_to_vec();
-        BackgroundHealStatusRequestV1::decode(encoded.as_slice()).expect("v1 server should ignore the version field");
-
-        let encoded = BackgroundHealStatusRequestV1 {}.encode_to_vec();
-        let decoded = proto_gen::node_service::BackgroundHealStatusRequest::decode(encoded.as_slice())
-            .expect("v2 server should accept a v1 request");
-        assert_eq!(decoded.protocol_version, 0);
-    }
-
     #[derive(Clone, Copy, Debug, Eq, Ord, PartialEq, PartialOrd)]
     struct CompatPayloadField {
         message: &'static str,

crates/protos/src/node.proto

@@ -846,9 +846,7 @@ message ScannerActivityResponse {
   bool dirty_usage_pending = 9;
 }
 
-message BackgroundHealStatusRequest {
-  uint32 protocol_version = 1;
-}
+message BackgroundHealStatusRequest {}
 
 message BackgroundHealStatusResponse {
   bool success = 1;

crates/replication/src/config.rs

@@ -270,6 +270,160 @@ pub fn active_replication_rule_destination_arns(config: &ReplicationConfiguratio
     arns
 }
 
+/// Deployment id extracted from a site-replication target ARN
+/// (`arn:{rustfs|minio}:replication::<deployment-id>:<bucket>`), or `None`
+/// for an operator-authored ARN.
+pub fn replication_target_arn_deployment_id(arn: &str) -> Option<String> {
+    let parts: Vec<_> = arn.split(':').collect();
+    if parts.len() == 6
+        && parts[0] == "arn"
+        && matches!(parts[1], "rustfs" | "minio")
+        && parts[2] == "replication"
+        && !parts[4].is_empty()
+    {
+        return Some(parts[4].to_string());
+    }
+
+    None
+}
+
+/// Rule id prefix the site-replication reconciler stamps on the rules it
+/// derives (`site-repl-<peer deployment id>`).
+pub const SITE_REPLICATION_RULE_ID_PREFIX: &str = "site-repl-";
+
+/// Whether `rule` carries a site-replication rule id (`site-repl-*`). The
+/// reconciler and the peer ingestion path treat the whole namespace as theirs
+/// on a site-replication bucket; the S3 edit path must not — rule ids are not
+/// reserved, so see [`site_replication_rule_deployment_id`].
+pub fn is_site_replication_rule(rule: &ReplicationRule) -> bool {
+    rule.id
+        .as_deref()
+        .is_some_and(|id| id.starts_with(SITE_REPLICATION_RULE_ID_PREFIX))
+}
+
+/// Deployment id of the peer a reconciler-derived rule replicates to, or
+/// `None` for any other rule. The reconciler builds each rule from one peer:
+/// the id is `site-repl-<deployment id>` and the destination ARN names that
+/// same deployment id — an operator-authored `site-repl-user` rule, or a
+/// `site-repl-<peer>` id pasted onto a foreign ARN, fails the agreement check.
+/// Callers that know the current peer set must also confirm the id is one of
+/// those peers before treating the rule as reconciler-owned.
+pub fn site_replication_rule_deployment_id(rule: &ReplicationRule) -> Option<&str> {
+    let deployment_id = rule.id.as_deref()?.strip_prefix(SITE_REPLICATION_RULE_ID_PREFIX)?;
+    (!deployment_id.is_empty()
+        && replication_target_arn_deployment_id(&rule.destination.bucket).as_deref() == Some(deployment_id))
+    .then_some(deployment_id)
+}
+
+/// Whether `rule` is one the local reconciler derived for a current remote
+/// site-replication peer in `peer_deployment_ids`. With an empty peer set
+/// (site replication disabled) nothing qualifies, so a bucket outside site
+/// replication keeps the verbatim S3 put/delete semantics.
+pub fn is_reconciler_owned_site_replication_rule(rule: &ReplicationRule, peer_deployment_ids: &HashSet<String>) -> bool {
+    site_replication_rule_deployment_id(rule).is_some_and(|deployment_id| peer_deployment_ids.contains(deployment_id))
+}
+
+/// Merge an incoming replication config into the local one.
+///
+/// `site-repl-*` rules encode the *holder's* outbound direction — their
+/// destination ARN names another site — so applying an external rule set
+/// verbatim replaces the local reverse rule with one this site can never
+/// satisfy (no bucket target backs it) and replication silently stops. Only
+/// operator-authored rules travel: the site-replication peer ingestion path
+/// and the S3 put/delete-bucket-replication path both keep the local site's
+/// `site-repl-*` rules through this merge. `incoming == None` models a
+/// delete of the operator-authored rules.
+pub fn merge_incoming_replication_config(
+    incoming: Option<ReplicationConfiguration>,
+    local: Option<ReplicationConfiguration>,
+) -> Option<ReplicationConfiguration> {
+    merge_replication_config_keeping_site_rules(incoming, local, is_site_replication_rule)
+}
+
+/// [`merge_incoming_replication_config`] for the S3 put/delete-bucket-replication
+/// path (issue #1948): only rules the local reconciler derived for a current
+/// peer in `peer_deployment_ids` survive as site rules; every other stored
+/// rule — including an operator-authored `site-repl-*` id — is operator state
+/// that the request replaces or deletes. An incoming rule whose id is a
+/// current peer's `site-repl-<id>` is dropped whatever its ARN: accepting it
+/// would duplicate the reconciler rule's id.
+pub fn merge_user_replication_config(
+    incoming: Option<ReplicationConfiguration>,
+    local: Option<ReplicationConfiguration>,
+    peer_deployment_ids: &HashSet<String>,
+) -> Option<ReplicationConfiguration> {
+    let incoming = incoming.map(|mut config| {
+        config.rules.retain(|rule| {
+            !rule
+                .id
+                .as_deref()
+                .and_then(|id| id.strip_prefix(SITE_REPLICATION_RULE_ID_PREFIX))
+                .is_some_and(|deployment_id| peer_deployment_ids.contains(deployment_id))
+        });
+        config
+    });
+    merge_replication_config_keeping_site_rules(incoming, local, |rule| {
+        is_reconciler_owned_site_replication_rule(rule, peer_deployment_ids)
+    })
+}
+
+fn merge_replication_config_keeping_site_rules(
+    incoming: Option<ReplicationConfiguration>,
+    local: Option<ReplicationConfiguration>,
+    is_site_rule: impl Fn(&ReplicationRule) -> bool,
+) -> Option<ReplicationConfiguration> {
+    let incoming_role = incoming.as_ref().map(|config| config.role.clone()).unwrap_or_default();
+    // Operator rules first, then the local site rules — the same order the
+    // site-replication reconciler produces, so its no-op check matches and
+    // the bucket metadata is written once per broadcast, not twice.
+    let mut rules: Vec<ReplicationRule> = incoming
+        .into_iter()
+        .flat_map(|config| config.rules)
+        .filter(|rule| !is_site_rule(rule))
+        .collect();
+    rules.extend(local.into_iter().flat_map(|config| config.rules).filter(&is_site_rule));
+
+    if rules.is_empty() {
+        return None;
+    }
+
+    assign_site_replication_rule_priorities(&mut rules, &is_site_rule);
+
+    // A site-replication ARN in `role` is the sender's, and the reconciler's
+    // per-peer target lookup reads it — carrying it over would pin the
+    // receiver's targets to the sender's identity.
+    let role = match replication_target_arn_deployment_id(&incoming_role) {
+        Some(_) => String::new(),
+        None => incoming_role,
+    };
+
+    Some(ReplicationConfiguration { role, rules })
+}
+
+/// Give the site rules in `rules` the lowest priorities no operator rule uses,
+/// in rule order, leaving every operator rule's priority untouched. Operator
+/// priorities decide which rule wins per target, so they are part of the
+/// submitted policy; site rules are derived state and only need to be unique
+/// (`validate_replication_config_structure` rejects duplicates). The result
+/// is a pure function of the rule list, so the site-replication reconciler,
+/// the peer ingestion merge and the S3 edit merge all converge on the same
+/// bytes and the reconciler's no-op check holds.
+pub fn assign_site_replication_rule_priorities(rules: &mut [ReplicationRule], is_site_rule: impl Fn(&ReplicationRule) -> bool) {
+    let taken: HashSet<i32> = rules
+        .iter()
+        .filter(|rule| !is_site_rule(rule))
+        .map(|rule| rule.priority.unwrap_or(0))
+        .collect();
+    let mut next = 1;
+    for rule in rules.iter_mut().filter(|rule| is_site_rule(rule)) {
+        while taken.contains(&next) {
+            next += 1;
+        }
+        rule.priority = Some(next);
+        next = next.saturating_add(1);
+    }
+}
+
 pub fn replication_target_arns(config: &ReplicationConfiguration) -> HashSet<String> {
     let role = config.role.trim();
     if !role.is_empty() {
@@ -1544,4 +1698,103 @@ mod tests {
             "the child rule must win for target A while the overlapping child target B remains eligible"
         );
     }
+
+    #[test]
+    fn site_replication_rule_deployment_id_requires_id_and_arn_agreement() {
+        let reconciler_rule = replication_rule("site-repl-peer-dep", "arn:rustfs:replication::peer-dep:bucket");
+        assert_eq!(site_replication_rule_deployment_id(&reconciler_rule), Some("peer-dep"));
+
+        // A remote-target ARN carries the remote's deployment id (or a random
+        // uuid), never the operator's rule id.
+        let operator_named_rule = replication_rule("site-repl-user", "arn:minio:replication:us-east-1:2f1c-remote:bucket");
+        assert_eq!(site_replication_rule_deployment_id(&operator_named_rule), None);
+
+        let foreign_arn = replication_rule("site-repl-peer-dep", "arn:rustfs:replication::other-dep:bucket");
+        assert_eq!(site_replication_rule_deployment_id(&foreign_arn), None);
+
+        let empty_id = replication_rule("site-repl-", "arn:rustfs:replication::peer-dep:bucket");
+        assert_eq!(site_replication_rule_deployment_id(&empty_id), None);
+
+        let peers = HashSet::from(["peer-dep".to_string()]);
+        assert!(is_reconciler_owned_site_replication_rule(&reconciler_rule, &peers));
+        assert!(!is_reconciler_owned_site_replication_rule(&reconciler_rule, &HashSet::new()));
+        let removed_peer = replication_rule("site-repl-gone-dep", "arn:rustfs:replication::gone-dep:bucket");
+        assert!(!is_reconciler_owned_site_replication_rule(&removed_peer, &peers));
+    }
+
+    // The merge must not rewrite the operator's priorities: with the
+    // priority-5 rule listed first and renumbered 1 then 2, the priority-1
+    // delete-marker-disabled rule would win the replication decision.
+    #[test]
+    fn merge_keeps_operator_priorities_and_replication_decision() {
+        let user_arn = "arn:minio:replication:us-east-1:2f1c-remote:bucket";
+        let peer_arn = "arn:rustfs:replication::peer-dep:bucket";
+        let incoming = ReplicationConfiguration {
+            role: String::new(),
+            rules: vec![
+                delete_marker_rule("dm-enabled", user_arn, "logs/", 5, true),
+                delete_marker_rule("dm-disabled", user_arn, "logs/2026/", 1, false),
+            ],
+        };
+        let mut site_rule = delete_marker_rule("site-repl-peer-dep", peer_arn, "", 7, true);
+        site_rule.prefix = None;
+        let local = structure_config(vec![site_rule]);
+        let opts = ObjectOpts {
+            name: "logs/2026/app.log".to_string(),
+            op_type: ReplicationType::Delete,
+            delete_marker: true,
+            version_id: None,
+            ..Default::default()
+        };
+        let submitted: Vec<_> = incoming.filter_target_replication_decisions(&opts);
+
+        let peers = HashSet::from(["peer-dep".to_string()]);
+        let merged = merge_user_replication_config(Some(incoming.clone()), Some(local.clone()), &peers).expect("rules");
+
+        let priorities: Vec<_> = merged
+            .rules
+            .iter()
+            .map(|rule| (rule.id.as_deref().unwrap(), rule.priority))
+            .collect();
+        assert_eq!(
+            priorities,
+            vec![
+                ("dm-enabled", Some(5)),
+                ("dm-disabled", Some(1)),
+                ("site-repl-peer-dep", Some(2))
+            ],
+            "operator priorities are kept verbatim; the site rule takes the lowest free slot"
+        );
+        assert!(validate_replication_config_structure(&merged).is_ok());
+        let mut decisions = merged.filter_target_replication_decisions(&opts);
+        decisions.retain(|(arn, _)| arn == user_arn);
+        assert_eq!(decisions, submitted, "the merged config must replicate exactly as the operator submitted");
+        assert_eq!(decisions, vec![(user_arn.to_string(), true)]);
+
+        // The peer ingestion merge follows the same rule.
+        let merged = merge_incoming_replication_config(Some(incoming), Some(local)).expect("rules");
+        let priorities: Vec<_> = merged.rules.iter().map(|rule| rule.priority).collect();
+        assert_eq!(priorities, vec![Some(5), Some(1), Some(2)]);
+    }
+
+    #[test]
+    fn site_rule_priorities_skip_every_operator_priority() {
+        let mut rules = vec![
+            delete_marker_rule("a", "arn:a", "", 2, true),
+            delete_marker_rule("site-repl-x", "arn:rustfs:replication::x:b", "", 9, true),
+            delete_marker_rule("b", "arn:a", "", 1, true),
+            delete_marker_rule("site-repl-y", "arn:rustfs:replication::y:b", "", 9, true),
+            delete_marker_rule("c", "arn:a", "", 4, true),
+        ];
+        assign_site_replication_rule_priorities(&mut rules, is_site_replication_rule);
+        let priorities: Vec<_> = rules.iter().map(|rule| rule.priority).collect();
+        assert_eq!(priorities, vec![Some(2), Some(3), Some(1), Some(5), Some(4)]);
+        assert!(validate_replication_config_structure(&structure_config(rules.clone())).is_ok());
+
+        // Idempotent, so the reconciler's pass over an already-merged config
+        // is a byte-stable no-op rather than a rewrite every period.
+        let settled = rules.clone();
+        assign_site_replication_rule_priorities(&mut rules, is_site_replication_rule);
+        assert_eq!(rules, settled);
+    }
 }

crates/replication/src/lib.rs

@@ -32,9 +32,11 @@ pub use config::{
     ObjectOpts, REMOTE_TARGET_CAPABILITY_CONTRACT_VERSION, REMOTE_TARGET_UNSUPPORTED_FIELDS, REMOTE_TARGET_WRITABLE_FIELDS,
     REPLICATION_CAPABILITY_CONTRACT_VERSION, REPLICATION_READ_ONLY_HISTORICAL_FIELDS, REPLICATION_WRITABLE_FIELDS,
     ReplicationConfigStructureError, ReplicationConfigurationExt, ReplicationTargetValidationError,
-    active_replication_rule_destination_arns, invalid_replication_config_status_field, replication_target_arns,
-    should_remove_replication_target, unsupported_replication_config_field, validate_replication_config_structure,
-    validate_replication_config_target_arns,
+    active_replication_rule_destination_arns, assign_site_replication_rule_priorities, invalid_replication_config_status_field,
+    is_reconciler_owned_site_replication_rule, is_site_replication_rule, merge_incoming_replication_config,
+    merge_user_replication_config, replication_target_arn_deployment_id, replication_target_arns,
+    should_remove_replication_target, site_replication_rule_deployment_id, unsupported_replication_config_field,
+    validate_replication_config_structure, validate_replication_config_target_arns,
 };
 pub use delete::{
     DeletedObjectReplicationInfo, delete_marker_purge_mrf_entry, delete_marker_purge_version_id,

rustfs/src/admin/handlers/heal.rs

@@ -20,7 +20,7 @@ use crate::admin::storage_api::bucket::utils::is_valid_object_prefix;
 use crate::server::ADMIN_PREFIX;
 use crate::server::RemoteAddr;
 use crate::storage::rpc::node_service::heal::{
-    HealControlCoordinator, NodeHealStatusSnapshot, capture_node_heal_status, decode_node_heal_status,
+    HealControlCoordinator, NodeHealProgress, NodeHealStatusSnapshot, capture_node_heal_status, decode_node_heal_status,
     decode_node_replacement_recovery_status, heal_control_coordinator, heal_topology_fingerprint,
 };
 use bytes::Bytes;
@@ -298,7 +298,14 @@ fn background_heal_runtime_state(
     }
 }
 
-type BackgroundHealProgress = rustfs_heal::HealProgress;
+#[derive(Debug, Serialize)]
+#[serde(rename_all = "camelCase")]
+struct BackgroundHealProgress {
+    objects_scanned: u64,
+    objects_healed: u64,
+    objects_failed: u64,
+    bytes_processed: u64,
+}
 
 #[derive(Debug)]
 struct ClusterHealStatusSnapshot {
@@ -337,10 +344,17 @@ fn add_operations(total: &mut rustfs_heal::HealOperationsSnapshot, next: rustfs_
     add_source_counts(&mut total.retrying_by_source, next.retrying_by_source);
 }
 
+fn add_progress(total: &mut BackgroundHealProgress, next: NodeHealProgress) {
+    total.objects_scanned = total.objects_scanned.saturating_add(next.objects_scanned);
+    total.objects_healed = total.objects_healed.saturating_add(next.objects_healed);
+    total.objects_failed = total.objects_failed.saturating_add(next.objects_failed);
+    total.bytes_processed = total.bytes_processed.saturating_add(next.bytes_processed);
+}
+
 fn aggregate_cluster_heal_status(snapshots: Vec<NodeHealStatusSnapshot>) -> ClusterHealStatusSnapshot {
     let mut info = BackgroundHealInfo::default();
     let mut operations = rustfs_heal::HealOperationsSnapshot::default();
-    let mut progress = Vec::new();
+    let mut progress = None;
     let mut any_services_enabled = false;
     let mut any_initialized = false;
 
@@ -357,12 +371,18 @@ fn aggregate_cluster_heal_status(snapshots: Vec<NodeHealStatusSnapshot>) -> Clus
         }
         add_operations(&mut operations, snapshot.operations);
         if let Some(next) = snapshot.progress {
-            progress.push(next);
+            add_progress(
+                progress.get_or_insert(BackgroundHealProgress {
+                    objects_scanned: 0,
+                    objects_healed: 0,
+                    objects_failed: 0,
+                    bytes_processed: 0,
+                }),
+                next,
+            );
         }
     }
 
-    let progress = rustfs_heal::aggregate_heal_progress(progress);
-
     let state = if operations.queue_length > 0 || operations.active_tasks > 0 || operations.retrying_tasks > 0 {
         HealRuntimeState::Active
     } else if any_initialized {
@@ -2181,19 +2201,10 @@ mod tests {
         };
 
         let progress = BackgroundHealProgress {
-            kind: rustfs_heal::heal::progress::HealProgressKind::ObjectSweep,
             objects_scanned: 7,
             objects_healed: 3,
             objects_failed: 1,
-            skipped_objects: 3,
-            objects_total_count: 10,
-            objects_total_size: 8192,
             bytes_processed: 4096,
-            progress_percentage: 50.0,
-            progress_state: rustfs_heal::heal::progress::HealProgressState::Running,
-            baseline_generation: Some(42),
-            baseline_known: true,
-            ..Default::default()
         };
 
         let encoded = encode_background_heal_status(
@@ -2209,14 +2220,7 @@ mod tests {
         assert_eq!(json["progress"]["objectsScanned"], 7);
         assert_eq!(json["progress"]["objectsHealed"], 3);
         assert_eq!(json["progress"]["objectsFailed"], 1);
-        assert_eq!(json["progress"]["skippedObjects"], 3);
-        assert_eq!(json["progress"]["objectsTotalCount"], 10);
-        assert_eq!(json["progress"]["objectsTotalSize"], 8192);
         assert_eq!(json["progress"]["bytesProcessed"], 4096);
-        assert_eq!(json["progress"]["progressState"], "running");
-        assert_eq!(json["progress"]["baselineGeneration"], 42);
-        assert_eq!(json["progress"]["baselineKnown"], true);
-        assert_eq!(json["progress"]["counterUnknown"], false);
     }
 
     #[test]
@@ -2238,18 +2242,10 @@ mod tests {
                 ..Default::default()
             },
             Some(NodeHealProgress {
-                kind: rustfs_heal::heal::progress::HealProgressKind::ObjectSweep,
                 objects_scanned: 3,
                 objects_healed: 1,
                 objects_failed: 0,
-                skipped_objects: 2,
-                objects_total_count: 6,
-                objects_total_size: 400,
                 bytes_processed: 100,
-                progress_state: rustfs_heal::heal::progress::HealProgressState::Running,
-                baseline_generation: Some(9),
-                baseline_known: true,
-                ..Default::default()
             }),
         );
         let peer = NodeHealStatusSnapshot::for_test(
@@ -2269,17 +2265,10 @@ mod tests {
                 ..Default::default()
             },
             Some(NodeHealProgress {
-                kind: rustfs_heal::heal::progress::HealProgressKind::ObjectSweep,
                 objects_scanned: 5,
                 objects_healed: 4,
                 objects_failed: 1,
-                objects_total_count: 4,
-                objects_total_size: 1600,
                 bytes_processed: 900,
-                progress_state: rustfs_heal::heal::progress::HealProgressState::Running,
-                baseline_generation: Some(9),
-                baseline_known: true,
-                ..Default::default()
             }),
         );
 
@@ -2298,15 +2287,7 @@ mod tests {
         assert_eq!(progress.objects_scanned, 8);
         assert_eq!(progress.objects_healed, 5);
         assert_eq!(progress.objects_failed, 1);
-        assert_eq!(progress.skipped_objects, 2);
-        assert_eq!(progress.objects_total_count, 10);
-        assert_eq!(progress.objects_total_size, 2000);
         assert_eq!(progress.bytes_processed, 1000);
-        assert_eq!(progress.progress_percentage, 50.0);
-        assert_eq!(progress.progress_state, rustfs_heal::heal::progress::HealProgressState::Running);
-        assert_eq!(progress.baseline_generation, Some(9));
-        assert!(progress.baseline_known);
-        assert!(!progress.counter_unknown);
 
         assert_eq!(peer_first.state, HealRuntimeState::Active);
         assert_eq!(peer_first.operations, local_first.operations);
@@ -2346,7 +2327,6 @@ mod tests {
             objects_healed: value,
             objects_failed: value,
             bytes_processed: value,
-            ..Default::default()
         };
         let saturated = NodeHealStatusSnapshot::for_test(
             true,

rustfs/src/admin/handlers/site_replication.rs

@@ -31,6 +31,10 @@ use crate::admin::storage_api::bucket::metadata::{
 use crate::admin::storage_api::bucket::metadata_sys;
 use crate::admin::storage_api::bucket::quota::BucketQuota;
 use crate::admin::storage_api::bucket::replication;
+use crate::admin::storage_api::bucket::replication::{
+    assign_site_replication_rule_priorities, is_site_replication_rule, merge_incoming_replication_config,
+    replication_target_arn_deployment_id,
+};
 use crate::admin::storage_api::bucket::target::{ARN, BucketTarget, BucketTargetType, BucketTargets, Credentials};
 use crate::admin::storage_api::bucket::target_sys::BucketTargetSys;
 use crate::admin::storage_api::bucket::utils::{deserialize, serialize};
@@ -1118,6 +1122,36 @@ async fn load_site_replication_state() -> S3Result<SiteReplicationState> {
     }
 }
 
+/// Whether this deployment participates in site replication (two or more
+/// peers in the persisted state). Read by the S3 interface layer to gate
+/// replication-config edits (MinIO `ErrReplicationDenyEditError` semantics,
+/// issue #1948); a state-read failure propagates so the gate fails closed.
+pub(crate) async fn site_replication_enabled() -> S3Result<bool> {
+    Ok(load_site_replication_state().await?.enabled())
+}
+
+/// Deployment ids of the remote peers the reconciler derives a
+/// `site-repl-<id>` rule for on every bucket (the same peer filter as
+/// `build_site_replication_config`); empty when site replication is not
+/// enabled. Read by the bucket usecase so an S3 replication-config edit keeps
+/// exactly the reconciler-owned rules (issue #1948); a state-read failure
+/// propagates so the edit fails closed.
+pub(crate) async fn site_replication_remote_peer_deployment_ids() -> S3Result<HashSet<String>> {
+    let state = load_site_replication_state().await?;
+    if !state.enabled() {
+        return Ok(HashSet::new());
+    }
+    let local_peer = current_local_runtime_peer(&state);
+    Ok(state
+        .peers
+        .values()
+        .filter(|peer| {
+            peer.deployment_id != local_peer.deployment_id && !same_identity_endpoint(&peer.endpoint, &local_peer.endpoint)
+        })
+        .map(|peer| peer.deployment_id.clone())
+        .collect())
+}
+
 async fn load_site_replication_state_no_lock(store: Arc<ECStore>) -> S3Result<SiteReplicationState> {
     match read_config_no_lock(store, SITE_REPLICATION_STATE_PATH).await {
         Ok(data) => parse_site_replication_state(&data),
@@ -7762,20 +7796,6 @@ fn bucket_target_deployment_id(target: &BucketTarget) -> Option<String> {
     replication_target_arn_deployment_id(&target.arn)
 }
 
-fn replication_target_arn_deployment_id(arn: &str) -> Option<String> {
-    let parts: Vec<_> = arn.split(':').collect();
-    if parts.len() == 6
-        && parts[0] == "arn"
-        && matches!(parts[1], "rustfs" | "minio")
-        && parts[2] == "replication"
-        && !parts[4].is_empty()
-    {
-        return Some(parts[4].to_string());
-    }
-
-    None
-}
-
 fn prune_removed_site_replication_bucket_targets(
     existing: BucketTargets,
     removed_deployment_ids: &HashSet<String>,
@@ -7800,10 +7820,6 @@ fn prune_removed_site_replication_bucket_targets(
     (BucketTargets { targets }, removed)
 }
 
-fn is_site_replication_rule(rule: &ReplicationRule) -> bool {
-    rule.id.as_deref().is_some_and(|id| id.starts_with("site-repl-"))
-}
-
 /// Whether every `site-repl-*` rule on this bucket resolves to a live remote target.
 ///
 /// The rule set alone cannot answer this: a rule can be perfectly formed while the endpoint
@@ -7829,52 +7845,6 @@ async fn site_replication_targets_online(bucket: &str, replication_config_xml: &
     true
 }
 
-/// Merge a peer's replication config into the local one.
-///
-/// `site-repl-*` rules encode the *sender's* outbound direction — their destination ARN
-/// names the receiver — so applying a peer's rule set verbatim replaces the receiver's
-/// reverse rule with one pointing at itself. No bucket target can satisfy that ARN
-/// (`reconcile_site_replication_bucket_targets` skips the local peer), so the receiver
-/// silently stops replicating back: the one-directional symptom. Only operator-authored
-/// rules travel between sites; each site owns its own `site-repl-*` rules.
-fn merge_incoming_replication_config(
-    incoming: Option<ReplicationConfiguration>,
-    local: Option<ReplicationConfiguration>,
-) -> Option<ReplicationConfiguration> {
-    let incoming_role = incoming.as_ref().map(|config| config.role.clone()).unwrap_or_default();
-    // Operator rules first, then the local site rules — the same order
-    // `ensure_site_replication_bucket_replication_config_with_runtime` produces, so its
-    // no-op check matches and the bucket metadata is written once per broadcast, not twice.
-    let mut rules: Vec<ReplicationRule> = incoming
-        .into_iter()
-        .flat_map(|config| config.rules)
-        .filter(|rule| !is_site_replication_rule(rule))
-        .collect();
-    rules.extend(
-        local
-            .into_iter()
-            .flat_map(|config| config.rules)
-            .filter(is_site_replication_rule),
-    );
-
-    if rules.is_empty() {
-        return None;
-    }
-
-    for (index, rule) in rules.iter_mut().enumerate() {
-        rule.priority = Some(i32::try_from(index + 1).unwrap_or(i32::MAX));
-    }
-
-    // A site-replication ARN in `role` is the sender's, and `site_replication_target_arns_by_peer`
-    // reads it — carrying it over would pin the receiver's targets to the sender's identity.
-    let role = match replication_target_arn_deployment_id(&incoming_role) {
-        Some(_) => String::new(),
-        None => incoming_role,
-    };
-
-    Some(ReplicationConfiguration { role, rules })
-}
-
 /// Merge a peer's ILM expiry document into the local lifecycle config.
 ///
 /// Mirrors MinIO's `mergeWithCurrentLCConfig` with one hardening: incoming
@@ -8213,9 +8183,7 @@ fn prune_removed_site_replication_rules(
         return (None, removed);
     }
 
-    for (index, rule) in config.rules.iter_mut().enumerate() {
-        rule.priority = Some(i32::try_from(index + 1).unwrap_or(i32::MAX));
-    }
+    assign_site_replication_rule_priorities(&mut config.rules, is_site_replication_rule);
 
     (Some(config), removed)
 }
@@ -8389,9 +8357,10 @@ async fn ensure_site_replication_bucket_replication_config_with_runtime(
         .cloned()
         .collect();
     rules.extend(desired.rules);
-    for (index, rule) in rules.iter_mut().enumerate() {
-        rule.priority = Some(i32::try_from(index + 1).unwrap_or(i32::MAX));
-    }
+    // Operator priorities are the operator's policy; only the derived rules
+    // take free slots, by the same function as the config merges so a merged
+    // write and this pass agree byte for byte.
+    assign_site_replication_rule_priorities(&mut rules, is_site_replication_rule);
 
     // Only a site-replication ARN in `role` is ours to drop — an operator-authored role is
     // part of the bucket's S3-visible configuration, and repairing a reverse rule must not
@@ -17089,7 +17058,7 @@ mod tests {
     }
 
     #[test]
-    fn test_prune_removed_site_replication_rules_removes_site_rule_and_reorders_priorities() {
+    fn test_prune_removed_site_replication_rules_removes_site_rule_and_keeps_operator_priority() {
         let removed_deployment_ids = HashSet::from(["removed-dep".to_string()]);
         let kept_rule = build_site_replication_rule("arn:rustfs:replication::kept-dep:photos", 3, "site-repl-kept-dep");
         let removed_rule = build_site_replication_rule("arn:rustfs:replication::removed-dep:photos", 1, "site-repl-removed-dep");
@@ -17106,9 +17075,9 @@ mod tests {
         assert!(updated.role.is_empty());
         assert_eq!(updated.rules.len(), 2);
         assert_eq!(updated.rules[0].id.as_deref(), Some("user-managed-rule"));
-        assert_eq!(updated.rules[0].priority, Some(1));
+        assert_eq!(updated.rules[0].priority, Some(9), "the operator's priority is policy and stays");
         assert_eq!(updated.rules[1].id.as_deref(), Some("site-repl-kept-dep"));
-        assert_eq!(updated.rules[1].priority, Some(2));
+        assert_eq!(updated.rules[1].priority, Some(1), "the derived rule moves to the lowest free slot");
     }
 
     #[test]

rustfs/src/admin/storage_api.rs

@@ -443,6 +443,8 @@ pub(crate) mod replication {
     pub(crate) use super::ecstore_bucket::replication::{
         REMOTE_TARGET_CAPABILITY_CONTRACT_VERSION, REMOTE_TARGET_UNSUPPORTED_FIELDS, REMOTE_TARGET_WRITABLE_FIELDS,
         REPLICATION_CAPABILITY_CONTRACT_VERSION, REPLICATION_READ_ONLY_HISTORICAL_FIELDS, REPLICATION_WRITABLE_FIELDS,
+        assign_site_replication_rule_priorities, is_site_replication_rule, merge_incoming_replication_config,
+        replication_target_arn_deployment_id,
     };
     pub(crate) type BucketReplicationResyncStatus = super::ecstore_bucket::replication::BucketReplicationResyncStatus;
     pub(crate) type BucketStats = super::ecstore_bucket::replication::BucketStats;

rustfs/src/app/bucket_usecase.rs

@@ -38,9 +38,9 @@ use super::storage_api::bucket_usecase::bucket::{
     metadata_sys,
     policy_sys::PolicySys,
     replication::{
-        ReplicationTargetValidationError, invalid_replication_config_status_field, replication_target_arns,
-        should_remove_replication_target, unsupported_replication_config_field, validate_replication_config_structure,
-        validate_replication_config_target_arns,
+        ReplicationTargetValidationError, invalid_replication_config_status_field, merge_user_replication_config,
+        replication_target_arns, should_remove_replication_target, unsupported_replication_config_field,
+        validate_replication_config_structure, validate_replication_config_target_arns,
     },
     target::{BucketTargetType, BucketTargets},
     utils::serialize,
@@ -623,11 +623,52 @@ async fn validate_bucket_replication_update(bucket: &str, config: &ReplicationCo
     validate_replication_config_targets(&targets, config)
 }
 
-async fn replication_targets_without_config_targets(
+/// Defense in depth for site-replication-managed buckets (issue #1948): an S3
+/// PutBucketReplication replaces the operator-authored rules but must not wipe
+/// the rules the reconciler derived for the current remote peers
+/// (`site_peer_deployment_ids`) — until its next pass (600s period) every
+/// peer link on this bucket would be silently dead. The same merge also drops
+/// incoming impostors of those rules. An empty peer set (site replication
+/// disabled) keeps the verbatim overwrite semantics: rule ids are not
+/// reserved, so an operator's own `site-repl-*` rule is ordinary state there.
+fn merge_user_replication_config_update(
+    incoming: ReplicationConfiguration,
+    existing: Option<ReplicationConfiguration>,
+    site_peer_deployment_ids: &HashSet<String>,
+) -> ReplicationConfiguration {
+    if site_peer_deployment_ids.is_empty() {
+        return incoming;
+    }
+    // `incoming` passed structure validation, so it holds at least one rule;
+    // `None` is only reachable when every incoming rule impersonates a
+    // reconciler rule, and then the stored reconciler rules are what remains.
+    merge_user_replication_config(Some(incoming.clone()), existing, site_peer_deployment_ids).unwrap_or(incoming)
+}
+
+/// Split of an S3 DeleteBucketReplication on the stored config (issue #1948):
+/// the operator-authored rules are removed, the rules the reconciler derived
+/// for the current remote peers survive (`None` means nothing survives and
+/// the config is deleted), and the returned ARNs are the ones whose bucket
+/// targets may be garbage-collected — never an ARN a surviving reconciler
+/// rule still points at.
+fn split_replication_config_for_user_delete(
+    config: ReplicationConfiguration,
+    site_peer_deployment_ids: &HashSet<String>,
+) -> (Option<ReplicationConfiguration>, HashSet<String>) {
+    let mut removable_arns = replication_target_arns(&config);
+    let remaining = merge_user_replication_config(None, Some(config), site_peer_deployment_ids);
+    if let Some(remaining) = remaining.as_ref() {
+        for rule in &remaining.rules {
+            removable_arns.remove(rule.destination.bucket.trim());
+        }
+    }
+    (remaining, removable_arns)
+}
+
+async fn replication_targets_without_arns(
     bucket: &str,
-    config: &ReplicationConfiguration,
+    target_arns: &HashSet<String>,
 ) -> S3Result<Option<(BucketTargets, usize)>> {
-    let target_arns = replication_target_arns(config);
     if target_arns.is_empty() {
         return Ok(None);
     }
@@ -638,7 +679,7 @@ async fn replication_targets_without_config_targets(
         Err(err) => return Err(ApiError::from(err).into()),
     };
 
-    let removed = remove_replication_targets_from_config_targets(&mut targets, &target_arns);
+    let removed = remove_replication_targets_from_config_targets(&mut targets, target_arns);
     if removed == 0 {
         return Ok(None);
     }
@@ -1582,9 +1623,15 @@ impl DefaultBucketUsecase {
         Ok(S3Response::new(DeleteBucketPolicyOutput {}))
     }
 
+    /// `site_peers` is the set of remote site-replication peer deployment ids
+    /// (empty when site replication is disabled). The interface layer reads it
+    /// from the persisted state and fails closed on a read error, so this
+    /// usecase stays a pure function of its inputs (layer rule: app never
+    /// imports interface).
     pub async fn execute_delete_bucket_replication(
         &self,
         req: S3Request<DeleteBucketReplicationInput>,
+        site_peers: HashSet<String>,
     ) -> S3Result<S3Response<DeleteBucketReplicationOutput>> {
         let expected_incarnation_id = bucket_config_mutation_incarnation(&req, &req.input.bucket)?;
         let request_context = req.extensions.get::<request_context::RequestContext>().cloned();
@@ -1604,15 +1651,29 @@ impl DefaultBucketUsecase {
             Err(StorageError::ConfigNotFound) => None,
             Err(err) => return Err(ApiError::from(err).into()),
         };
-        let updated_targets = if let Some(config) = replication_config.as_ref() {
-            replication_targets_without_config_targets(&bucket, config).await?
+        let (remaining_config, updated_targets) = if let Some(config) = replication_config.as_ref() {
+            let (remaining, removable_arns) = split_replication_config_for_user_delete(config.clone(), &site_peers);
+            let targets = replication_targets_without_arns(&bucket, &removable_arns).await?;
+            (remaining, targets)
         } else {
-            None
+            (None, None)
         };
 
-        delete_bucket_config_for_incarnation(&bucket, BUCKET_REPLICATION_CONFIG, expected_incarnation_id)
-            .await
-            .map_err(ApiError::from)?;
+        match remaining_config {
+            // Site-replication rules and the targets backing them survive the
+            // S3 delete (issue #1948); only the operator-authored rules go.
+            Some(remaining) => {
+                let data = serialize_config(&remaining)?;
+                update_bucket_config_for_incarnation(&bucket, BUCKET_REPLICATION_CONFIG, data, expected_incarnation_id)
+                    .await
+                    .map_err(ApiError::from)?;
+            }
+            None => {
+                delete_bucket_config_for_incarnation(&bucket, BUCKET_REPLICATION_CONFIG, expected_incarnation_id)
+                    .await
+                    .map_err(ApiError::from)?;
+            }
+        }
         if let Some((targets, removed)) = updated_targets
             && let Err(err) =
                 write_replication_targets_after_config_delete(&bucket, &targets, removed, expected_incarnation_id).await
@@ -2459,9 +2520,11 @@ impl DefaultBucketUsecase {
         Ok(S3Response::new(PutBucketCorsOutput::default()))
     }
 
+    /// See [`Self::execute_delete_bucket_replication`] for `site_peers`.
     pub async fn execute_put_bucket_replication(
         &self,
         req: S3Request<PutBucketReplicationInput>,
+        site_peers: HashSet<String>,
     ) -> S3Result<S3Response<PutBucketReplicationOutput>> {
         let expected_incarnation_id = bucket_config_mutation_incarnation(&req, &req.input.bucket)?;
         let request_context = req.extensions.get::<request_context::RequestContext>().cloned();
@@ -2485,6 +2548,13 @@ impl DefaultBucketUsecase {
 
         let targets_guard = lock_bucket_targets_metadata(&bucket).await;
         validate_bucket_replication_update(&bucket, &replication_configuration).await?;
+        let existing_config = match metadata_sys::get_replication_config(&bucket).await {
+            Ok((config, _)) => Some(config),
+            Err(StorageError::ConfigNotFound) => None,
+            Err(err) => return Err(ApiError::from(err).into()),
+        };
+        let replication_configuration =
+            merge_user_replication_config_update(replication_configuration, existing_config, &site_peers);
         let data = serialize_config(&replication_configuration)?;
         update_bucket_config_for_incarnation(&bucket, BUCKET_REPLICATION_CONFIG, data, expected_incarnation_id)
             .await
@@ -3114,6 +3184,185 @@ mod tests {
         assert!(arns.contains(destination));
     }
 
+    fn replication_rule_with_id(arn: &str, id: &str, priority: i32) -> ReplicationRule {
+        let mut rule = replication_rule_for_target(arn);
+        rule.id = Some(id.to_string());
+        rule.priority = Some(priority);
+        rule
+    }
+
+    fn site_peers(deployment_ids: &[&str]) -> HashSet<String> {
+        deployment_ids.iter().map(|id| id.to_string()).collect()
+    }
+
+    #[test]
+    fn put_replication_merge_preserves_site_replication_rules() {
+        let existing = ReplicationConfiguration {
+            role: String::new(),
+            rules: vec![
+                replication_rule_with_id("arn:rustfs:replication::peer-dep:bucket", "site-repl-peer-dep", 1),
+                replication_rule_with_id("arn:rustfs:replication:us-east-1:old:bucket", "old-user-rule", 2),
+            ],
+        };
+        let incoming = ReplicationConfiguration {
+            role: String::new(),
+            rules: vec![
+                replication_rule_with_id("arn:rustfs:replication:us-east-1:new:bucket", "new-user-rule", 1),
+                replication_rule_with_id("arn:rustfs:replication::forged-dep:bucket", "site-repl-peer-dep", 2),
+                replication_rule_with_id("arn:rustfs:replication::other-dep:bucket", "site-repl-other", 3),
+            ],
+        };
+
+        let merged = merge_user_replication_config_update(incoming, Some(existing), &site_peers(&["peer-dep"]));
+
+        let rules: Vec<_> = merged
+            .rules
+            .iter()
+            .map(|rule| (rule.id.as_deref().unwrap_or_default(), rule.destination.bucket.as_str()))
+            .collect();
+        assert_eq!(
+            rules,
+            vec![
+                ("new-user-rule", "arn:rustfs:replication:us-east-1:new:bucket"),
+                ("site-repl-other", "arn:rustfs:replication::other-dep:bucket"),
+                ("site-repl-peer-dep", "arn:rustfs:replication::peer-dep:bucket"),
+            ],
+            "user rules replaced, the reconciler rule for the current peer kept over the incoming impostor, \
+             a site-repl-* id that names no current peer is ordinary operator state"
+        );
+    }
+
+    // Rule ids do not reserve `site-repl-*`: outside site replication an
+    // owner's `site-repl-user` rule is ordinary state, so PUT stores it
+    // verbatim and DELETE removes it and garbage-collects its target.
+    #[test]
+    fn put_then_delete_replication_without_site_replication_treats_site_repl_id_as_user_rule() {
+        let user_arn = "arn:minio:replication:us-east-1:2f1c-remote:bucket";
+        let incoming = ReplicationConfiguration {
+            role: String::new(),
+            rules: vec![replication_rule_with_id(user_arn, "site-repl-user", 1)],
+        };
+
+        let stored = merge_user_replication_config_update(incoming.clone(), None, &HashSet::new());
+        assert_eq!(stored, incoming, "PUT on a non-site-replication bucket is verbatim");
+
+        let (remaining, removable) = split_replication_config_for_user_delete(stored, &HashSet::new());
+        assert!(remaining.is_none(), "DELETE must remove the operator's site-repl-* rule");
+        assert_eq!(removable, HashSet::from([user_arn.to_string()]));
+    }
+
+    // Under site replication only a rule the reconciler would derive — id
+    // `site-repl-<peer>` for a current peer, destination ARN naming the same
+    // peer — is reconciler-owned. Everything else is operator state.
+    #[test]
+    fn delete_replication_split_keeps_only_reconciler_derived_rules() {
+        let peer_arn = "arn:rustfs:replication::peer-dep:bucket";
+        let user_arn = "arn:minio:replication:us-east-1:2f1c-remote:bucket";
+        let config = ReplicationConfiguration {
+            role: String::new(),
+            rules: vec![
+                replication_rule_with_id(user_arn, "site-repl-user", 1),
+                replication_rule_with_id(user_arn, "site-repl-peer-dep", 2),
+                replication_rule_with_id("arn:rustfs:replication::gone-dep:bucket", "site-repl-gone-dep", 3),
+                replication_rule_with_id(peer_arn, "site-repl-peer-dep", 4),
+            ],
+        };
+
+        let (remaining, removable) = split_replication_config_for_user_delete(config, &site_peers(&["peer-dep"]));
+
+        let remaining = remaining.expect("the reconciler-derived rule must survive");
+        assert_eq!(remaining.rules.len(), 1);
+        assert_eq!(remaining.rules[0].destination.bucket, peer_arn);
+        assert_eq!(
+            removable,
+            HashSet::from([user_arn.to_string(), "arn:rustfs:replication::gone-dep:bucket".to_string()]),
+            "targets of operator rules and of a removed peer are garbage-collected"
+        );
+    }
+
+    #[test]
+    fn put_replication_merge_returns_incoming_verbatim_without_site_rules() {
+        let existing = ReplicationConfiguration {
+            role: String::new(),
+            rules: vec![replication_rule_with_id(
+                "arn:rustfs:replication:us-east-1:old:bucket",
+                "old-user-rule",
+                7,
+            )],
+        };
+        let incoming = ReplicationConfiguration {
+            role: String::new(),
+            rules: vec![replication_rule_with_id(
+                "arn:rustfs:replication:us-east-1:new:bucket",
+                "new-user-rule",
+                5,
+            )],
+        };
+
+        let merged = merge_user_replication_config_update(incoming.clone(), Some(existing), &HashSet::new());
+
+        assert_eq!(merged.role, incoming.role);
+        assert_eq!(merged.rules, incoming.rules, "non-SR buckets keep the verbatim overwrite semantics");
+    }
+
+    #[test]
+    fn delete_replication_split_keeps_site_rules_and_their_targets() {
+        let sr_arn = "arn:rustfs:replication::peer-dep:bucket";
+        let user_arn = "arn:rustfs:replication:us-east-1:user:bucket";
+        let config = ReplicationConfiguration {
+            role: String::new(),
+            rules: vec![
+                replication_rule_with_id(user_arn, "user-rule", 1),
+                replication_rule_with_id(sr_arn, "site-repl-peer-dep", 2),
+            ],
+        };
+
+        let (remaining, removable) = split_replication_config_for_user_delete(config, &site_peers(&["peer-dep"]));
+
+        let remaining = remaining.expect("site-replication rules must survive a user delete");
+        let ids: Vec<_> = remaining
+            .rules
+            .iter()
+            .map(|rule| rule.id.as_deref().unwrap_or_default())
+            .collect();
+        assert_eq!(ids, vec!["site-repl-peer-dep"]);
+        assert_eq!(removable, HashSet::from([user_arn.to_string()]));
+    }
+
+    #[test]
+    fn delete_replication_split_protects_targets_shared_with_site_rules() {
+        let sr_arn = "arn:rustfs:replication::peer-dep:bucket";
+        let config = ReplicationConfiguration {
+            role: String::new(),
+            rules: vec![
+                replication_rule_with_id(sr_arn, "user-rule-on-sr-target", 1),
+                replication_rule_with_id(sr_arn, "site-repl-peer-dep", 2),
+            ],
+        };
+
+        let (remaining, removable) = split_replication_config_for_user_delete(config, &site_peers(&["peer-dep"]));
+
+        assert!(remaining.is_some());
+        assert!(
+            removable.is_empty(),
+            "a target still referenced by a surviving site-replication rule must not be removed"
+        );
+    }
+
+    #[test]
+    fn delete_replication_split_removes_everything_without_site_rules() {
+        let user_arn = "arn:rustfs:replication:us-east-1:user:bucket";
+        let config = ReplicationConfiguration {
+            role: String::new(),
+            rules: vec![replication_rule_with_id(user_arn, "user-rule", 1)],
+        };
+
+        let (remaining, removable) = split_replication_config_for_user_delete(config, &site_peers(&["peer-dep"]));
+
+        assert!(remaining.is_none(), "without site-replication rules the whole config is deleted");
+        assert_eq!(removable, HashSet::from([user_arn.to_string()]));
+    }
+
     fn replication_targets_with_arn(arns: &[&str]) -> BucketTargets {
         BucketTargets {
             targets: arns
@@ -3451,7 +3700,10 @@ mod tests {
         let req = build_request(input, Method::DELETE);
         let usecase = DefaultBucketUsecase::without_context();
 
-        let err = usecase.execute_delete_bucket_replication(req).await.unwrap_err();
+        let err = usecase
+            .execute_delete_bucket_replication(req, HashSet::new())
+            .await
+            .unwrap_err();
         assert_eq!(err.code(), &S3ErrorCode::InternalError);
     }
 
@@ -4537,7 +4789,7 @@ mod tests {
         let req = build_request(input, Method::PUT);
         let usecase = DefaultBucketUsecase::without_context();
 
-        let err = usecase.execute_put_bucket_replication(req).await.unwrap_err();
+        let err = usecase.execute_put_bucket_replication(req, HashSet::new()).await.unwrap_err();
         assert_eq!(err.code(), &S3ErrorCode::InternalError);
     }
 
@@ -4555,7 +4807,7 @@ mod tests {
             .unwrap();
 
         let err = DefaultBucketUsecase::without_context()
-            .execute_put_bucket_replication(build_request(input, Method::PUT))
+            .execute_put_bucket_replication(build_request(input, Method::PUT), HashSet::new())
             .await
             .expect_err("unsupported fields must be rejected before store access");
 

rustfs/src/app/storage_api.rs

@@ -619,6 +619,8 @@ pub(crate) mod bucket {
 
         use crate::storage::storage_api::ecstore_bucket::replication as replication_contracts;
 
+        pub(crate) use replication_contracts::merge_user_replication_config;
+
         type ReplicationObjectBridge = crate::storage::storage_api::ecstore_bucket::replication::ReplicationObjectBridge;
         pub(crate) type DeleteReplicationConfigSnapshot =
             crate::storage::storage_api::ecstore_bucket::replication::DeleteReplicationConfigSnapshot;

rustfs/src/storage/ecfs.rs

@@ -63,6 +63,69 @@ use crate::app::storage_api::object_usecase::bucket::replication::{
 };
 use crate::storage::storage_api::ecfs_consumer::StorageObjectOptions as ObjectOptions;
 
+#[cfg(test)]
+static SITE_REPLICATION_GATE_TEST_OVERRIDE: std::sync::atomic::AtomicU8 = std::sync::atomic::AtomicU8::new(0);
+#[cfg(test)]
+const SITE_REPLICATION_GATE_FORCE_DISABLED: u8 = 1;
+#[cfg(test)]
+const SITE_REPLICATION_GATE_FORCE_ENABLED: u8 = 2;
+
+async fn site_replication_gate_enabled() -> S3Result<bool> {
+    #[cfg(test)]
+    match SITE_REPLICATION_GATE_TEST_OVERRIDE.load(std::sync::atomic::Ordering::SeqCst) {
+        SITE_REPLICATION_GATE_FORCE_DISABLED => return Ok(false),
+        SITE_REPLICATION_GATE_FORCE_ENABLED => return Ok(true),
+        _ => {}
+    }
+    crate::admin::handlers::site_replication::site_replication_enabled().await
+}
+
+/// Remote site-replication peer deployment ids handed to the bucket usecase
+/// so an S3 replication-config edit keeps exactly the reconciler-owned rules
+/// (issue #1948). Read here, in the interface layer, because the usecase must
+/// not import the admin handlers (layer guard); a state-read failure
+/// propagates so the edit fails closed.
+async fn site_replication_peer_deployment_ids_for_edit() -> S3Result<std::collections::HashSet<String>> {
+    // While the gate override is in effect the test exercises the deny/allow
+    // branch, not the peer set; there is no persisted state to read.
+    #[cfg(test)]
+    if SITE_REPLICATION_GATE_TEST_OVERRIDE.load(std::sync::atomic::Ordering::SeqCst) != 0 {
+        return Ok(std::collections::HashSet::new());
+    }
+    crate::admin::handlers::site_replication::site_replication_remote_peer_deployment_ids().await
+}
+
+/// MinIO `ErrReplicationDenyEditError`.
+fn replication_deny_edit_error() -> S3Error {
+    let mut err = S3Error::with_message(
+        S3ErrorCode::Custom("XMinioReplicationDenyEdit".into()),
+        "Sub-User is not allowed to edit Replication configuration",
+    );
+    err.set_status_code(StatusCode::BAD_REQUEST);
+    err
+}
+
+/// Site-replication gate for S3 replication-config edits (issue #1948).
+///
+/// On a site-replication deployment the bucket's replication config carries
+/// the operator-managed `site-repl-*` rules that keep every peer in sync, and
+/// a successful edit is broadcast to all peers — so a user holding only
+/// bucket-scoped `s3:PutReplicationConfiguration` could rewrite or erase
+/// replication net-wide. MinIO parity (`ErrReplicationDenyEditError`): only
+/// owner credentials (root or root-parented) may edit. Runs after the policy
+/// authorization in the access layer and only on the external S3 path — the
+/// reconciler and peer bucket-meta ingestion never route through these
+/// handlers.
+async fn deny_replication_config_edit_for_non_owner<T>(req: &S3Request<T>) -> S3Result<()> {
+    if crate::storage::access::req_info_ref(req)?.is_owner {
+        return Ok(());
+    }
+    if site_replication_gate_enabled().await? {
+        return Err(replication_deny_edit_error());
+    }
+    Ok(())
+}
+
 #[derive(Debug, Clone)]
 pub struct FS {
     /// This server's late-bound application-context slot (backlog#1052 S2).
@@ -500,8 +563,10 @@ impl S3 for FS {
         &self,
         req: S3Request<DeleteBucketReplicationInput>,
     ) -> S3Result<S3Response<DeleteBucketReplicationOutput>> {
+        deny_replication_config_edit_for_non_owner(&req).await?;
+        let site_peers = site_replication_peer_deployment_ids_for_edit().await?;
         let usecase = s3_api::bucket_usecase_for(self);
-        usecase.execute_delete_bucket_replication(req).await
+        usecase.execute_delete_bucket_replication(req, site_peers).await
     }
 
     #[instrument(level = "debug", skip(self))]
@@ -1353,8 +1418,10 @@ impl S3 for FS {
         &self,
         req: S3Request<PutBucketReplicationInput>,
     ) -> S3Result<S3Response<PutBucketReplicationOutput>> {
+        deny_replication_config_edit_for_non_owner(&req).await?;
+        let site_peers = site_replication_peer_deployment_ids_for_edit().await?;
         let usecase = s3_api::bucket_usecase_for(self);
-        usecase.execute_put_bucket_replication(req).await
+        usecase.execute_put_bucket_replication(req, site_peers).await
     }
 
     async fn put_bucket_request_payment(
@@ -1919,3 +1986,103 @@ impl S3 for FS {
         Box::pin(usecase.execute_upload_part_copy(req)).await
     }
 }
+
+#[cfg(test)]
+mod tests {
+    use super::{
+        FS, SITE_REPLICATION_GATE_FORCE_DISABLED, SITE_REPLICATION_GATE_FORCE_ENABLED, SITE_REPLICATION_GATE_TEST_OVERRIDE,
+    };
+    use crate::storage::access::ReqInfo;
+    use http::Method;
+    use http::StatusCode;
+    use s3s::dto::{DeleteBucketReplicationInput, PutBucketReplicationInput, ReplicationConfiguration};
+    use s3s::{S3, S3Error, S3ErrorCode, S3Request};
+    use std::sync::atomic::Ordering;
+
+    fn replication_config_edit_request<T>(input: T, is_owner: bool) -> S3Request<T> {
+        let mut req = S3Request {
+            input,
+            method: Method::PUT,
+            uri: http::Uri::from_static("/"),
+            headers: http::HeaderMap::new(),
+            extensions: http::Extensions::new(),
+            credentials: None,
+            region: None,
+            service: None,
+            trailing_headers: None,
+        };
+        req.extensions.insert(ReqInfo {
+            is_owner,
+            ..Default::default()
+        });
+        req
+    }
+
+    fn put_bucket_replication_input() -> PutBucketReplicationInput {
+        PutBucketReplicationInput {
+            bucket: "test-bucket".to_string(),
+            checksum_algorithm: None,
+            content_md5: None,
+            expected_bucket_owner: None,
+            replication_configuration: ReplicationConfiguration {
+                role: String::new(),
+                rules: Vec::new(),
+            },
+            token: None,
+        }
+    }
+
+    fn delete_bucket_replication_input() -> DeleteBucketReplicationInput {
+        DeleteBucketReplicationInput {
+            bucket: "test-bucket".to_string(),
+            expected_bucket_owner: None,
+        }
+    }
+
+    fn assert_replication_deny_edit(err: &S3Error) {
+        match err.code() {
+            S3ErrorCode::Custom(code) => assert_eq!(code, "XMinioReplicationDenyEdit"),
+            other => panic!("expected XMinioReplicationDenyEdit, got {other:?}"),
+        }
+        assert_eq!(err.status_code(), Some(StatusCode::BAD_REQUEST));
+    }
+
+    /// Single test on purpose: the branches share the process-wide gate
+    /// override, and parallel tests would race it.
+    #[tokio::test]
+    async fn replication_config_edit_gate_denies_only_non_owner_under_site_replication() {
+        let fs = FS::new();
+        SITE_REPLICATION_GATE_TEST_OVERRIDE.store(SITE_REPLICATION_GATE_FORCE_ENABLED, Ordering::SeqCst);
+
+        // Non-owner PUT/DELETE through the real S3 handlers: denied by the
+        // gate before the usecase (and thus the store) is ever touched.
+        let err = fs
+            .put_bucket_replication(replication_config_edit_request(put_bucket_replication_input(), false))
+            .await
+            .expect_err("non-owner PutBucketReplication must be denied while site replication is enabled");
+        assert_replication_deny_edit(&err);
+        let err = fs
+            .delete_bucket_replication(replication_config_edit_request(delete_bucket_replication_input(), false))
+            .await
+            .expect_err("non-owner DeleteBucketReplication must be denied while site replication is enabled");
+        assert_replication_deny_edit(&err);
+
+        // Owner passes the gate (the usecase's empty-rules structure error
+        // proves the request reached the usecase instead of the deny path).
+        let err = fs
+            .put_bucket_replication(replication_config_edit_request(put_bucket_replication_input(), true))
+            .await
+            .expect_err("owner request should pass the gate and fail later on config validation");
+        assert_eq!(err.code(), &S3ErrorCode::InvalidRequest);
+
+        // Without site replication the policy check alone still governs the edit.
+        SITE_REPLICATION_GATE_TEST_OVERRIDE.store(SITE_REPLICATION_GATE_FORCE_DISABLED, Ordering::SeqCst);
+        let err = fs
+            .put_bucket_replication(replication_config_edit_request(put_bucket_replication_input(), false))
+            .await
+            .expect_err("non-owner request should pass the gate and fail later on config validation");
+        assert_eq!(err.code(), &S3ErrorCode::InvalidRequest);
+
+        SITE_REPLICATION_GATE_TEST_OVERRIDE.store(0, Ordering::SeqCst);
+    }
+}

rustfs/src/storage/rpc/node_service.rs

@@ -1912,7 +1912,7 @@ impl Node for NodeService {
 
     async fn background_heal_status(
         &self,
-        request: Request<BackgroundHealStatusRequest>,
+        _request: Request<BackgroundHealStatusRequest>,
     ) -> Result<Response<BackgroundHealStatusResponse>, Status> {
         if self.resolve_object_store().is_none() {
             return Ok(Response::new(BackgroundHealStatusResponse {
@@ -1922,7 +1922,7 @@ impl Node for NodeService {
             }));
         }
         let snapshot = heal::capture_node_heal_status(rustfs_scanner::scanner::BackgroundHealInfo::default()).await;
-        match heal::encode_node_heal_status(&snapshot, request.into_inner().protocol_version) {
+        match heal::encode_node_heal_status(&snapshot) {
             Ok(bg_heal_state) => Ok(Response::new(BackgroundHealStatusResponse {
                 success: true,
                 bg_heal_state: bg_heal_state.into(),

rustfs/src/storage/rpc/node_service/heal.rs

@@ -25,8 +25,7 @@ use std::io::Cursor;
 
 use super::super::encode_msgpack_map;
 
-const NODE_HEAL_STATUS_PREVIOUS_VERSION: u8 = 1;
-const NODE_HEAL_STATUS_VERSION: u8 = 2;
+const NODE_HEAL_STATUS_VERSION: u8 = 1;
 const NODE_HEAL_STATUS_MAX_SIZE: usize = 64 * 1024;
 const NODE_REPLACEMENT_RECOVERY_STATUS_VERSION: u8 = 1;
 const NODE_REPLACEMENT_RECOVERY_STATUS_MAX_SIZE: usize = 64 * 1024;
@@ -173,38 +172,13 @@ pub(crate) fn heal_topology_fingerprint(endpoint_pools: &EndpointServerPools) ->
     Ok(hex_simd::encode_to_string(hasher.finalize(), hex_simd::AsciiCase::Lower))
 }
 
-pub(crate) type NodeHealProgress = rustfs_heal::HealProgress;
-
-#[derive(Debug, Clone, Serialize, Deserialize)]
+#[derive(Debug, Clone, Default, PartialEq, Eq, Serialize, Deserialize)]
 #[serde(rename_all = "camelCase", deny_unknown_fields)]
-struct NodeHealProgressV1 {
-    objects_scanned: u64,
-    objects_healed: u64,
-    objects_failed: u64,
-    bytes_processed: u64,
-}
-
-impl From<&NodeHealProgress> for NodeHealProgressV1 {
-    fn from(progress: &NodeHealProgress) -> Self {
-        Self {
-            objects_scanned: progress.objects_scanned,
-            objects_healed: progress.objects_healed,
-            objects_failed: progress.objects_failed,
-            bytes_processed: progress.bytes_processed,
-        }
-    }
-}
-
-impl From<NodeHealProgressV1> for NodeHealProgress {
-    fn from(progress: NodeHealProgressV1) -> Self {
-        Self {
-            objects_scanned: progress.objects_scanned,
-            objects_healed: progress.objects_healed,
-            objects_failed: progress.objects_failed,
-            bytes_processed: progress.bytes_processed,
-            ..Default::default()
-        }
-    }
+pub(crate) struct NodeHealProgress {
+    pub objects_scanned: u64,
+    pub objects_healed: u64,
+    pub objects_failed: u64,
+    pub bytes_processed: u64,
 }
 
 #[derive(Debug, Clone, Serialize, Deserialize)]
@@ -246,48 +220,6 @@ pub(crate) struct NodeHealStatusSnapshot {
     pub progress: Option<NodeHealProgress>,
 }
 
-#[derive(Debug, Clone, Serialize, Deserialize)]
-#[serde(rename_all = "camelCase", deny_unknown_fields)]
-struct NodeHealStatusSnapshotV1 {
-    version: u8,
-    services_enabled: bool,
-    initialized: bool,
-    info: NodeHealInfo,
-    operations: HealOperationsSnapshot,
-    progress: Option<NodeHealProgressV1>,
-}
-
-impl From<&NodeHealStatusSnapshot> for NodeHealStatusSnapshotV1 {
-    fn from(snapshot: &NodeHealStatusSnapshot) -> Self {
-        Self {
-            version: NODE_HEAL_STATUS_PREVIOUS_VERSION,
-            services_enabled: snapshot.services_enabled,
-            initialized: snapshot.initialized,
-            info: snapshot.info.clone(),
-            operations: snapshot.operations,
-            progress: snapshot.progress.as_ref().map(NodeHealProgressV1::from),
-        }
-    }
-}
-
-impl From<NodeHealStatusSnapshotV1> for NodeHealStatusSnapshot {
-    fn from(snapshot: NodeHealStatusSnapshotV1) -> Self {
-        Self {
-            version: snapshot.version,
-            services_enabled: snapshot.services_enabled,
-            initialized: snapshot.initialized,
-            info: snapshot.info,
-            operations: snapshot.operations,
-            progress: snapshot.progress.map(NodeHealProgress::from),
-        }
-    }
-}
-
-#[derive(Deserialize)]
-struct NodeHealStatusVersion {
-    version: u8,
-}
-
 impl NodeHealStatusSnapshot {
     #[cfg(test)]
     pub(crate) fn for_test(
@@ -317,7 +249,14 @@ impl NodeHealStatusSnapshot {
 }
 
 pub(crate) async fn capture_node_heal_status(info: BackgroundHealInfo) -> NodeHealStatusSnapshot {
-    let progress = rustfs_heal::current_heal_progress_snapshot().await;
+    let progress = rustfs_heal::current_heal_progress_snapshot()
+        .await
+        .map(|progress| NodeHealProgress {
+            objects_scanned: progress.objects_scanned,
+            objects_healed: progress.objects_healed,
+            objects_failed: progress.objects_failed,
+            bytes_processed: progress.bytes_processed,
+        });
 
     NodeHealStatusSnapshot {
         version: NODE_HEAL_STATUS_VERSION,
@@ -329,44 +268,23 @@ pub(crate) async fn capture_node_heal_status(info: BackgroundHealInfo) -> NodeHe
     }
 }
 
-pub(crate) fn encode_node_heal_status(snapshot: &NodeHealStatusSnapshot, protocol_version: u32) -> Result<Vec<u8>, String> {
-    let encoded = if protocol_version < rustfs_protos::BACKGROUND_HEAL_STATUS_PROTOCOL_VERSION {
-        encode_msgpack_map(&NodeHealStatusSnapshotV1::from(snapshot))
-    } else {
-        let mut snapshot = snapshot.clone();
-        snapshot.version = NODE_HEAL_STATUS_VERSION;
-        encode_msgpack_map(&snapshot)
-    };
-    encoded.map_err(|err| format!("failed to encode node heal status: {err}"))
+pub(crate) fn encode_node_heal_status(snapshot: &NodeHealStatusSnapshot) -> Result<Vec<u8>, String> {
+    encode_msgpack_map(snapshot).map_err(|err| format!("failed to encode node heal status: {err}"))
 }
 
 pub(crate) fn decode_node_heal_status(data: &[u8]) -> Result<NodeHealStatusSnapshot, String> {
     if data.len() > NODE_HEAL_STATUS_MAX_SIZE {
         return Err("node heal status exceeds size limit".to_string());
     }
-    let decode_version = || {
-        let mut deserializer = Deserializer::new(Cursor::new(data));
-        NodeHealStatusVersion::deserialize(&mut deserializer)
-            .map(|version| (version, deserializer))
-            .map_err(|err| format!("failed to decode node heal status: {err}"))
-    };
-    let (version, deserializer) = decode_version()?;
-    if usize::try_from(deserializer.get_ref().position()).ok() != Some(data.len()) {
-        return Err("node heal status contains trailing data".to_string());
-    }
-
     let mut deserializer = Deserializer::new(Cursor::new(data));
-    let snapshot = match version.version {
-        NODE_HEAL_STATUS_PREVIOUS_VERSION => {
-            NodeHealStatusSnapshotV1::deserialize(&mut deserializer).map(NodeHealStatusSnapshot::from)
-        }
-        NODE_HEAL_STATUS_VERSION => NodeHealStatusSnapshot::deserialize(&mut deserializer),
-        version => return Err(format!("unsupported node heal status version: {version}")),
-    }
-    .map_err(|err| format!("failed to decode node heal status: {err}"))?;
+    let snapshot = NodeHealStatusSnapshot::deserialize(&mut deserializer)
+        .map_err(|err| format!("failed to decode node heal status: {err}"))?;
     if usize::try_from(deserializer.get_ref().position()).ok() != Some(data.len()) {
         return Err("node heal status contains trailing data".to_string());
     }
+    if snapshot.version != NODE_HEAL_STATUS_VERSION {
+        return Err(format!("unsupported node heal status version: {}", snapshot.version));
+    }
     Ok(snapshot)
 }
 
@@ -469,10 +387,9 @@ pub(crate) fn decode_node_replacement_recovery_status(data: &[u8]) -> Result<Nod
 #[cfg(test)]
 mod tests {
     use super::{
-        NODE_HEAL_STATUS_MAX_SIZE, NODE_HEAL_STATUS_PREVIOUS_VERSION, NODE_HEAL_STATUS_VERSION, NodeHealProgress,
-        NodeHealStatusSnapshot, NodeReplacementRecoveryStatusSnapshot, decode_node_heal_status,
-        decode_node_replacement_recovery_status, encode_node_heal_status, encode_node_replacement_recovery_status,
-        heal_control_coordinator, heal_topology_fingerprint,
+        NODE_HEAL_STATUS_MAX_SIZE, NODE_HEAL_STATUS_VERSION, NodeHealProgress, NodeHealStatusSnapshot,
+        NodeReplacementRecoveryStatusSnapshot, decode_node_heal_status, decode_node_replacement_recovery_status,
+        encode_node_heal_status, encode_node_replacement_recovery_status, heal_control_coordinator, heal_topology_fingerprint,
     };
     use crate::storage::storage_api::{
         Endpoint,
@@ -616,72 +533,20 @@ mod tests {
                 ..Default::default()
             },
             Some(NodeHealProgress {
-                kind: rustfs_heal::heal::progress::HealProgressKind::ObjectSweep,
                 objects_scanned: 7,
                 objects_healed: 5,
                 objects_failed: 1,
-                skipped_objects: 1,
-                objects_total_count: 10,
-                objects_total_size: 2048,
                 bytes_processed: 1024,
-                progress_percentage: 50.0,
-                progress_state: rustfs_heal::heal::progress::HealProgressState::Running,
-                baseline_generation: Some(42),
-                baseline_known: true,
-                ..Default::default()
             }),
         );
 
-        let encoded = encode_node_heal_status(&snapshot, rustfs_protos::BACKGROUND_HEAL_STATUS_PROTOCOL_VERSION)
-            .expect("snapshot should encode");
+        let encoded = encode_node_heal_status(&snapshot).expect("snapshot should encode");
         let decoded = decode_node_heal_status(&encoded).expect("snapshot should decode");
         assert_eq!(decoded.version, NODE_HEAL_STATUS_VERSION);
         assert_eq!(decoded.operations.queue_length, 2);
         assert_eq!(decoded.progress, snapshot.progress);
     }
 
-    #[test]
-    fn node_heal_status_v1_encoding_preserves_rolling_compatibility() {
-        let snapshot = NodeHealStatusSnapshot::for_test(
-            true,
-            true,
-            BackgroundHealInfo::default(),
-            HealOperationsSnapshot::default(),
-            Some(NodeHealProgress {
-                objects_scanned: 7,
-                objects_healed: 5,
-                objects_failed: 1,
-                skipped_objects: 3,
-                bytes_processed: 1024,
-                progress_state: rustfs_heal::heal::progress::HealProgressState::Running,
-                baseline_known: true,
-                ..Default::default()
-            }),
-        );
-
-        let encoded =
-            encode_node_heal_status(&snapshot, u32::from(NODE_HEAL_STATUS_PREVIOUS_VERSION)).expect("v1 snapshot should encode");
-        let wire: serde_json::Value = rmp_serde::from_slice(&encoded).expect("v1 snapshot should decode as JSON");
-        let progress = wire["progress"].as_object().expect("v1 progress should be a map");
-        assert_eq!(wire["version"], NODE_HEAL_STATUS_PREVIOUS_VERSION);
-        assert_eq!(progress.len(), 4);
-        assert_eq!(progress["objectsScanned"], 7);
-        assert!(!progress.contains_key("skippedObjects"));
-
-        let decoded = decode_node_heal_status(&encoded).expect("v1 snapshot should decode");
-        let progress = decoded.progress.as_ref().expect("v1 progress should be present");
-        assert_eq!(decoded.version, NODE_HEAL_STATUS_PREVIOUS_VERSION);
-        assert_eq!(progress.objects_scanned, 7);
-        assert_eq!(progress.progress_state, rustfs_heal::heal::progress::HealProgressState::Unknown);
-        assert!(!progress.baseline_known);
-
-        let upgraded = encode_node_heal_status(&decoded, rustfs_protos::BACKGROUND_HEAL_STATUS_PROTOCOL_VERSION)
-            .expect("decoded v1 snapshot should upgrade to v2");
-        let upgraded = decode_node_heal_status(&upgraded).expect("upgraded snapshot should decode");
-        assert_eq!(upgraded.version, NODE_HEAL_STATUS_VERSION);
-        assert_eq!(upgraded.progress.as_ref(), Some(progress));
-    }
-
     #[test]
     fn node_heal_status_rejects_unknown_version() {
         let mut snapshot = NodeHealStatusSnapshot::for_test(
@@ -693,7 +558,7 @@ mod tests {
         );
         snapshot.version += 1;
 
-        let encoded = rmp_serde::to_vec_named(&snapshot).expect("snapshot should encode");
+        let encoded = encode_node_heal_status(&snapshot).expect("snapshot should encode");
         let err = decode_node_heal_status(&encoded).expect_err("unknown version should fail closed");
         assert!(err.contains("unsupported node heal status version"));
     }
@@ -714,22 +579,13 @@ mod tests {
                 "activeBySource": {"scanner": 0, "admin": 1, "autoHeal": 0, "internal": 0, "readRepair": 0},
                 "retryingBySource": {"scanner": 0, "admin": 0, "autoHeal": 0, "internal": 0, "readRepair": 0}
             },
-            "progress": {
-                "objectsScanned": 7,
-                "objectsHealed": 5,
-                "objectsFailed": 1,
-                "bytesProcessed": 1024
-            }
+            "progress": null
         });
         let encoded = rmp_serde::to_vec_named(&fixture).expect("fixture should encode");
         let decoded = decode_node_heal_status(&encoded).expect("fixed v1 fixture should decode");
         assert_eq!(decoded.info().bitrot_start_cycle, 9);
         assert_eq!(decoded.operations.queue_length, 2);
         assert_eq!(decoded.operations.queued_by_source.mrf, 0);
-        let progress = decoded.progress.expect("legacy progress should decode");
-        assert_eq!(progress.objects_scanned, 7);
-        assert!(!progress.baseline_known);
-        assert_eq!(progress.progress_state, rustfs_heal::heal::progress::HealProgressState::Unknown);
     }
 
     #[test]
@@ -775,8 +631,7 @@ mod tests {
             HealOperationsSnapshot::default(),
             None,
         );
-        let encoded = encode_node_heal_status(&snapshot, rustfs_protos::BACKGROUND_HEAL_STATUS_PROTOCOL_VERSION)
-            .expect("snapshot should encode");
+        let encoded = encode_node_heal_status(&snapshot).expect("snapshot should encode");
         let encoded_json: serde_json::Value = rmp_serde::from_slice(&encoded).expect("encoded snapshot should decode as JSON");
         assert_eq!(encoded_json["info"]["bitrotStartTime"], serde_json::json!("2023-11-14T22:13:20.123456Z"));
     }