From b0f491549a7a6dc51c57c2074c7a1bf9be2ea5e9 Mon Sep 17 00:00:00 2001 From: Zhengchao An Date: Wed, 1 Jul 2026 15:39:02 +0800 Subject: [PATCH] refactor(ecstore): hide replication runtime types (#4131) --- .../ecstore/src/bucket/replication/README.md | 8 +- .../replication_bandwidth_boundary.rs | 4 +- .../replication_storage_boundary.rs | 1 + .../bucket/replication/runtime_boundary.rs | 8 +- .../architecture/ecstore-module-split-plan.md | 7 +- scripts/check_architecture_migration_rules.sh | 83 +++++++++++++++++++ 6 files changed, 101 insertions(+), 10 deletions(-) diff --git a/crates/ecstore/src/bucket/replication/README.md b/crates/ecstore/src/bucket/replication/README.md index 5b11950f3..312a02925 100644 --- a/crates/ecstore/src/bucket/replication/README.md +++ b/crates/ecstore/src/bucket/replication/README.md @@ -13,7 +13,7 @@ and lifecycle/heal scheduling paths. | `datatypes.rs` | Replication status and operation DTOs. | Publicly re-exported through the ECStore replication facade. | | `replication_pool.rs` | Replication queue, worker pool, MRF persistence, bucket stats, and delete/object scheduling. | Depends on bucket target sys, bucket metadata sys, metadata paths, and file metadata replication contracts through local boundaries, config storage, storage contracts through the replication storage boundary, runtime sources, and notification state. | | `replication_resyncer.rs` | Object replication, delete replication, resync, MRF encode/decode, target calls, and multipart target upload paths. | Depends on target calls and target config types through the replication target boundary, metadata paths and metadata systems through the replication metadata boundary, file metadata replication contracts through the filemeta boundary, error contracts through the error boundary, versioning systems, storage contracts through the replication storage boundary, config-derived storage class labels through the config store, runtime sources, notification events and local event host selection through the event sink, bandwidth reader wrapping, and SetDisks lock timing. | -| `replication_state.rs` | Replication queue/stat state and worker accounting. | Reads runtime sources, file metadata replication contracts, and error contracts through local boundaries, and owns shared replication pool/stat state. | +| `replication_state.rs` | Replication queue/stat state and worker accounting. | Reads runtime sources, file metadata replication contracts, error contracts, and bucket monitor handles through local boundaries, and owns shared replication pool/stat state. | | `rule.rs` | Rule evaluation helpers for object replication options. | Depends on ECStore replication object option types. | | `mod.rs` | Compatibility re-export facade for the current ECStore owner. | Must stay stable until downstream scanner, lifecycle, heal, and metrics paths compile through replacement contracts. | @@ -28,7 +28,7 @@ and lifecycle/heal scheduling paths. | `ReplicationFileMeta` | Replication status, decisions, MRF entries, resync decisions, and target reset helpers. | `rustfs_filemeta` replication contracts are concentrated in `replication_filemeta_boundary.rs`; `FileInfo` remains in the storage boundary for storage trait bindings and walk options. | | `ReplicationErrorBoundary` | ECStore error/result contracts and replication-specific error classifiers. | `crate::error` imports are concentrated in `replication_error_boundary.rs`. | | `ReplicationTargetStore` | Bucket target listing, target client lookup, target offline checks, target config types, and target operation option types. | Bucket target sys access, `BucketTargets`, and target operation types are concentrated in `replication_target_boundary.rs`. | -| `ReplicationRuntime` | Worker pool, queue sizing, stats, bucket monitor, local node identity, cancellation, and admission state. | Direct runtime source/global access and shared replication pool/stat state. | +| `ReplicationRuntime` | Worker pool, queue sizing, stats, bucket monitor, local node identity, cancellation, and admission state. | Direct runtime source/global access and shared replication pool/stat state; ECStore object store and bucket monitor implementation types stay behind local storage/bandwidth boundaries. | | `ReplicationBandwidthLimiter` | Target reader wrapping for replication bandwidth accounting and throttling. | Direct bucket bandwidth reader imports from resyncer paths. | | `ReplicationEventSink` | Notification and audit events for skipped, failed, pending, and completed replication operations. | Event notification service calls and local event host selection are concentrated in `replication_event_sink.rs`. | | `ReplicationTagFilter` | Decode object tag strings for rule and metadata replication decisions. | Direct bucket tagging helper imports from replication workers. | @@ -48,7 +48,9 @@ and lifecycle/heal scheduling paths. lifecycle, scanner, OBS, heal, and tests compile through replacement paths. 5. Keep imports between modules in this directory relative to the local replication module, not `crate::bucket::replication::*` self paths. -6. Move at most one contract boundary per code-bearing PR and verify it with +6. Keep runtime source access from importing ECStore object store or bucket + monitor implementation types directly; use local boundary-owned aliases. +7. Move at most one contract boundary per code-bearing PR and verify it with focused replication tests before broad gates. ## First Code-Bearing Step diff --git a/crates/ecstore/src/bucket/replication/replication_bandwidth_boundary.rs b/crates/ecstore/src/bucket/replication/replication_bandwidth_boundary.rs index c307c2f4a..13741ce37 100644 --- a/crates/ecstore/src/bucket/replication/replication_bandwidth_boundary.rs +++ b/crates/ecstore/src/bucket/replication/replication_bandwidth_boundary.rs @@ -19,9 +19,11 @@ use tokio::io::AsyncRead; use crate::bucket::bandwidth::monitor::Monitor; use crate::bucket::bandwidth::reader::{BucketOptions, MonitorReaderOptions, MonitoredReader}; +pub(crate) type ReplicationBucketMonitor = Monitor; + pub(crate) fn wrap_reader( stream: Box, - monitor: Arc, + monitor: Arc, bucket: &str, arn: &str, header_size: usize, diff --git a/crates/ecstore/src/bucket/replication/replication_storage_boundary.rs b/crates/ecstore/src/bucket/replication/replication_storage_boundary.rs index bd50e125c..d1f03548c 100644 --- a/crates/ecstore/src/bucket/replication/replication_storage_boundary.rs +++ b/crates/ecstore/src/bucket/replication/replication_storage_boundary.rs @@ -16,6 +16,7 @@ use rustfs_filemeta::FileInfo; use tokio_util::sync::CancellationToken; use super::replication_error_boundary::Error; +pub(crate) type ReplicationObjectStore = crate::store::ECStore; pub(crate) use crate::client::api_get_options::{AdvancedGetOptions, StatObjectOptions}; pub(crate) use crate::object_api::{GetObjectReader, ObjectInfo, ObjectOptions, PutObjReader}; pub(crate) use crate::storage_api_contracts::list::{ diff --git a/crates/ecstore/src/bucket/replication/runtime_boundary.rs b/crates/ecstore/src/bucket/replication/runtime_boundary.rs index ae4de9ebe..d930c6119 100644 --- a/crates/ecstore/src/bucket/replication/runtime_boundary.rs +++ b/crates/ecstore/src/bucket/replication/runtime_boundary.rs @@ -14,13 +14,13 @@ use std::sync::Arc; +use super::replication_bandwidth_boundary::ReplicationBucketMonitor; use super::replication_pool::DynReplicationPool; use super::replication_state::ReplicationStats; -use crate::bucket::bandwidth::monitor::Monitor; +use super::replication_storage_boundary::ReplicationObjectStore; use crate::runtime::sources; -use crate::store::ECStore; -pub(crate) fn object_store_handle() -> Option> { +pub(crate) fn object_store_handle() -> Option> { sources::object_store_handle() } @@ -40,6 +40,6 @@ pub(crate) fn replication_runtime_initialized() -> bool { sources::replication_runtime_initialized() } -pub(crate) fn bucket_monitor() -> Option> { +pub(crate) fn bucket_monitor() -> Option> { sources::bucket_monitor() } diff --git a/docs/architecture/ecstore-module-split-plan.md b/docs/architecture/ecstore-module-split-plan.md index 014a94d78..95e6dfe7a 100644 --- a/docs/architecture/ecstore-module-split-plan.md +++ b/docs/architecture/ecstore-module-split-plan.md @@ -117,7 +117,9 @@ Current coupling: - global replication pool/stat initialization still lives with ECStore runtime compatibility state; - modules inside `bucket/replication` use local relative paths rather than the - ECStore owner path for replication self-imports. + ECStore owner path for replication self-imports; +- replication runtime source access uses storage/bandwidth boundary aliases for + ECStore object store and bucket monitor implementation types. Required contracts before crate movement: @@ -154,7 +156,8 @@ Required contracts before crate movement: are concentrated in `crates/ecstore/src/bucket/replication/replication_target_boundary.rs`. - `ReplicationRuntime`: pool, stats, worker admission, bucket monitor, local - node identity, cancellation, and queue sizing. + node identity, cancellation, and queue sizing. Concrete ECStore object store + and bucket monitor types stay behind local storage/bandwidth boundaries. - `ReplicationBandwidthLimiter`: target reader wrapping for replication bandwidth accounting and throttling. - `ReplicationEventSink`: notification/audit events for skipped, failed, and diff --git a/scripts/check_architecture_migration_rules.sh b/scripts/check_architecture_migration_rules.sh index 78fa1d823..751d3157c 100755 --- a/scripts/check_architecture_migration_rules.sh +++ b/scripts/check_architecture_migration_rules.sh @@ -208,6 +208,7 @@ REPLICATION_LOCAL_PATH_BYPASS_HITS_FILE="${TMP_DIR}/replication_local_path_bypas REPLICATION_LOCK_BOUNDARY_BYPASS_HITS_FILE="${TMP_DIR}/replication_lock_boundary_bypass_hits.txt" REPLICATION_METADATA_BOUNDARY_BYPASS_HITS_FILE="${TMP_DIR}/replication_metadata_boundary_bypass_hits.txt" REPLICATION_MSGP_BOUNDARY_BYPASS_HITS_FILE="${TMP_DIR}/replication_msgp_boundary_bypass_hits.txt" +REPLICATION_RUNTIME_TYPE_BYPASS_HITS_FILE="${TMP_DIR}/replication_runtime_type_bypass_hits.txt" REPLICATION_STORAGE_BOUNDARY_BYPASS_HITS_FILE="${TMP_DIR}/replication_storage_boundary_bypass_hits.txt" REPLICATION_TARGET_BOUNDARY_BYPASS_HITS_FILE="${TMP_DIR}/replication_target_boundary_bypass_hits.txt" REPLICATION_TAGGING_BOUNDARY_BYPASS_HITS_FILE="${TMP_DIR}/replication_tagging_boundary_bypass_hits.txt" @@ -2570,6 +2571,88 @@ if [[ -s "$REPLICATION_LOCAL_PATH_BYPASS_HITS_FILE" ]]; then report_failure "replication modules must use local relative paths instead of crate-qualified replication self paths: $(paste -sd '; ' "$REPLICATION_LOCAL_PATH_BYPASS_HITS_FILE")" fi +( + cd "$ROOT_DIR" + find crates/ecstore/src/bucket/replication -type f -name '*.rs' -print0 | + xargs -0 perl -0ne ' + next if $ARGV =~ m{replication_(storage|bandwidth)_boundary\.rs$}; + + sub blank { + my ($text) = @_; + $text =~ s/[^\n]/ /g; + return $text; + } + + sub blank_block_comments { + my ($text) = @_; + my $out = $text; + my $len = length($text); + my $i = 0; + + while ($i < $len) { + if ($i + 1 < $len && substr($text, $i, 2) eq "/*") { + my $start = $i; + my $depth = 0; + + while ($i < $len) { + if ($i + 1 < $len && substr($text, $i, 2) eq "/*") { + $depth++; + $i += 2; + next; + } + if ($i + 1 < $len && substr($text, $i, 2) eq "*/") { + $depth--; + $i += 2; + last if $depth == 0; + next; + } + $i++; + } + + substr($out, $start, $i - $start) = blank(substr($text, $start, $i - $start)); + next; + } + + $i++; + } + + return $out; + } + + my $source = $_; + my $hash = chr(35); + $source =~ s{b?r(${hash}*)".*?"\1}{blank($&)}egs; + $source =~ s{b?"(?:\\.|[^"\\])*"}{blank($&)}egs; + $source =~ s{'\''(?:\\.|[^'\''\\])+'\''}{blank($&)}egs; + $source = blank_block_comments($source); + $source =~ s{//[^\n]*}{blank($&)}eg; + + for my $statement (split /;/, $source) { + my $normalized = $statement; + $normalized =~ s/\s+//g; + + my $hits_ecstore = + $normalized =~ /\bECStore\b/ + && $normalized =~ /crate::(?:store(?:::|::\{)|\{.*store(?:::|::\{))/s; + my $hits_monitor = + $normalized =~ /\bMonitor\b/ + && $normalized =~ /crate::(?:bucket(?:::|::\{)|\{.*bucket(?:::|::\{))/s + && $normalized =~ /bucket(?:::|::\{).*bandwidth(?:::|::\{).*monitor(?:::|::\{)/s; + + if ($hits_ecstore || $hits_monitor) { + my $prefix = substr($source, 0, index($source, $statement)); + my $line = ($prefix =~ tr/\n//) + 1; + $normalized =~ s/\s+/ /g; + print "$ARGV:$line:$normalized\n"; + } + } + ' || true +) >"$REPLICATION_RUNTIME_TYPE_BYPASS_HITS_FILE" + +if [[ -s "$REPLICATION_RUNTIME_TYPE_BYPASS_HITS_FILE" ]]; then + report_failure "replication runtime concrete types must stay behind local storage or bandwidth boundaries: $(paste -sd '; ' "$REPLICATION_RUNTIME_TYPE_BYPASS_HITS_FILE")" +fi + ( cd "$ROOT_DIR" rg -n --with-filename 'crate::runtime::sources(\s+as\s+runtime_sources|::)' \