SERVER-39708 Optimize Btree multikey updates

5 files changed, 56 insertions, 2 deletions

diff --git a/src/mongo/db/catalog/multi_index_block.cpp b/src/mongo/db/catalog/multi_index_block.cpp
index 1eb834bd5bb..ee75f9ed64e 100644
--- a/src/mongo/db/catalog/multi_index_block.cpp
+++ b/src/mongo/db/catalog/multi_index_block.cpp
@@ -838,6 +838,12 @@ Status MultiIndexBlock::commit(OperationContext* opCtx,
                 boost::optional<MultikeyPaths>(MultikeyPathTracker::get(opCtx).getMultikeyPathInfo(
                     collection->ns(), _indexes[i].block->getIndexName()));
             if (multikeyPaths) {
+                // Upon reaching this point, multikeyPaths must either have at least one (possibly
+                // empty) element unless it is of an index with a type that doesn't support tracking
+                // multikeyPaths via the multikeyPaths array or has "special" multikey semantics.
+                invariant(!multikeyPaths.get().empty() ||
+                          !IndexBuildInterceptor::typeCanFastpathMultikeyUpdates(
+                              _indexes[i].block->getEntry()->descriptor()->getIndexType()));
                 _indexes[i].block->getEntry()->setMultikey(opCtx, *multikeyPaths);
             }
         }
diff --git a/src/mongo/db/index/index_build_interceptor.cpp b/src/mongo/db/index/index_build_interceptor.cpp
index c55a569442f..048163d6c6e 100644
--- a/src/mongo/db/index/index_build_interceptor.cpp
+++ b/src/mongo/db/index/index_build_interceptor.cpp
@@ -55,6 +55,18 @@ namespace mongo {
 
 MONGO_FAIL_POINT_DEFINE(hangDuringIndexBuildDrainYield);
 
+bool IndexBuildInterceptor::typeCanFastpathMultikeyUpdates(IndexType indexType) {
+    // Ensure no new indexes are added without considering whether they use the multikeyPaths
+    // vector.
+    invariant(indexType == INDEX_BTREE || indexType == INDEX_2D || indexType == INDEX_HAYSTACK ||
+              indexType == INDEX_2DSPHERE || indexType == INDEX_TEXT || indexType == INDEX_HASHED ||
+              indexType == INDEX_WILDCARD);
+    // Only BTREE indexes are guaranteed to use the multikeyPaths vector. Other index types either
+    // do not track path-level multikey information or have "special" handling of multikey
+    // information.
+    return (indexType == INDEX_BTREE);
+}
+
 IndexBuildInterceptor::IndexBuildInterceptor(OperationContext* opCtx, IndexCatalogEntry* entry)
     : _indexCatalogEntry(entry),
       _sideWritesTable(
@@ -64,6 +76,15 @@ IndexBuildInterceptor::IndexBuildInterceptor(OperationContext* opCtx, IndexCatal
     if (entry->descriptor()->unique()) {
         _duplicateKeyTracker = std::make_unique<DuplicateKeyTracker>(opCtx, entry);
     }
+    // `mergeMultikeyPaths` is sensitive to the two inputs having the same multikey
+    // "shape". Initialize `_multikeyPaths` with the right shape from the IndexCatalogEntry.
+    auto indexType = entry->descriptor()->getIndexType();
+    if (typeCanFastpathMultikeyUpdates(indexType)) {
+        auto numFields = entry->descriptor()->getNumFields();
+        _multikeyPaths = MultikeyPaths{};
+        auto it = _multikeyPaths->begin();
+        _multikeyPaths->insert(it, numFields, {});
+    }
 }
 
 void IndexBuildInterceptor::deleteTemporaryTables(OperationContext* opCtx) {
@@ -379,12 +400,17 @@ Status IndexBuildInterceptor::sideWrite(OperationContext* opCtx,
                                         Op op,
                                         int64_t* const numKeysOut) {
     invariant(opCtx->lockState()->inAWriteUnitOfWork());
-
     // Maintain parity with IndexAccessMethods handling of key counting. Only include
     // `multikeyMetadataKeys` when inserting.
     *numKeysOut = keys.size() + (op == Op::kInsert ? multikeyMetadataKeys.size() : 0);
 
-    if (op == Op::kInsert) {
+    auto indexType = _indexCatalogEntry->descriptor()->getIndexType();
+
+    // No need to take the multikeyPaths mutex if this is a trivial multikey update.
+    bool canBypassMultikeyMutex = typeCanFastpathMultikeyUpdates(indexType) &&
+        MultikeyPathTracker::isMultikeyPathsTrivial(multikeyPaths);
+
+    if (op == Op::kInsert && !canBypassMultikeyMutex) {
         // SERVER-39705: It's worth noting that a document may not generate any keys, but be
         // described as being multikey. This step must be done to maintain parity with `validate`s
         // expectations.
@@ -392,6 +418,10 @@ Status IndexBuildInterceptor::sideWrite(OperationContext* opCtx,
         if (_multikeyPaths) {
             MultikeyPathTracker::mergeMultikeyPaths(&_multikeyPaths.get(), multikeyPaths);
         } else {
+            // All indexes that support pre-initialization of _multikeyPaths during
+            // IndexBuildInterceptor construction time should have been initialized already.
+            invariant(!typeCanFastpathMultikeyUpdates(indexType));
+
             // `mergeMultikeyPaths` is sensitive to the two inputs having the same multikey
             // "shape". Initialize `_multikeyPaths` with the right shape from the first result.
             _multikeyPaths = multikeyPaths;
diff --git a/src/mongo/db/index/index_build_interceptor.h b/src/mongo/db/index/index_build_interceptor.h
index 4bdaaebf6f0..97dca244576 100644
--- a/src/mongo/db/index/index_build_interceptor.h
+++ b/src/mongo/db/index/index_build_interceptor.h
@@ -48,6 +48,8 @@ class IndexBuildInterceptor {
 public:
     enum class Op { kInsert, kDelete };
 
+    static bool typeCanFastpathMultikeyUpdates(IndexType type);
+
     /**
      * Creates a temporary table for writes during an index build. Additionally creates a temporary
      * table to store any duplicate key constraint violations found during the build, if the index
@@ -91,6 +93,7 @@ public:
      */
     Status checkDuplicateKeyConstraints(OperationContext* opCtx) const;
 
+
     /**
      * Performs a resumable scan on the side writes table, and either inserts or removes each key
      * from the underlying IndexAccessMethod. This will only insert as many records as are visible
diff --git a/src/mongo/db/multi_key_path_tracker.cpp b/src/mongo/db/multi_key_path_tracker.cpp
index d78271932e5..003a600e6bf 100644
--- a/src/mongo/db/multi_key_path_tracker.cpp
+++ b/src/mongo/db/multi_key_path_tracker.cpp
@@ -68,6 +68,15 @@ void MultikeyPathTracker::mergeMultikeyPaths(MultikeyPaths* toMergeInto,
     }
 }
 
+bool MultikeyPathTracker::isMultikeyPathsTrivial(const MultikeyPaths& paths) {
+    for (auto&& path : paths) {
+        if (!path.empty()) {
+            return false;
+        }
+    }
+    return true;
+}
+
 void MultikeyPathTracker::addMultikeyPathInfo(MultikeyPathInfo info) {
     invariant(_trackMultikeyPathInfo);
     // Merge the `MultikeyPathInfo` input into the accumulated value being tracked for the
diff --git a/src/mongo/db/multi_key_path_tracker.h b/src/mongo/db/multi_key_path_tracker.h
index 1fb1ee20fac..b155bc1fe58 100644
--- a/src/mongo/db/multi_key_path_tracker.h
+++ b/src/mongo/db/multi_key_path_tracker.h
@@ -63,6 +63,12 @@ public:
 
     static void mergeMultikeyPaths(MultikeyPaths* toMergeInto, const MultikeyPaths& newPaths);
 
+    /**
+     * Returns whether paths contains only empty sets, i.e., {{}, {}, {}}. This includes the case
+     * where the MultikeyPaths vector itself has no elements, e.g., {}.
+     */
+    static bool isMultikeyPathsTrivial(const MultikeyPaths& paths);
+
     // Decoration requires a default constructor.
     MultikeyPathTracker() = default;