/** * Copyright (C) 2022-present MongoDB, Inc. * * This program is free software: you can redistribute it and/or modify * it under the terms of the Server Side Public License, version 1, * as published by MongoDB, Inc. * * This program is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * Server Side Public License for more details. * * You should have received a copy of the Server Side Public License * along with this program. If not, see * . * * As a special exception, the copyright holders give permission to link the * code of portions of this program with the OpenSSL library under certain * conditions as described in each individual source file and distribute * linked combinations including the program with the OpenSSL library. You * must comply with the Server Side Public License in all respects for * all of the code used other than as permitted herein. If you modify file(s) * with this exception, you may extend this exception to your version of the * file(s), but you are not obligated to do so. If you do not wish to do so, * delete this exception statement from your version. If you delete this * exception statement from all source files in the program, then also delete * it in the license file. */ #pragma once #include "mongo/db/concurrency/d_concurrency.h" #include "mongo/db/operation_context.h" #include "mongo/db/repl/replica_set_aware_service.h" #include "mongo/util/concurrency/thread_pool.h" #include "mongo/util/uuid.h" namespace mongo { /** * Acquires the config db lock in IX mode and the collection lock for config.rangeDeletions in X * mode. */ class ScopedRangeDeleterLock { public: ScopedRangeDeleterLock(OperationContext* opCtx); ScopedRangeDeleterLock(OperationContext* opCtx, const UUID& collectionUuid); private: Lock::DBLock _configLock; Lock::CollectionLock _rangeDeletionLock; boost::optional _collectionUuidLock; }; /** * The BalancerStatsRegistry is used to cache metadata on shards, such as the orphan documents * count. The blancer (on the config sever) periodically fetches this metadata through the * _shardsvrGetStatsForBalancing command and uses it to evaluate balancing status of collections. * * The BalancerStatsRegistry is active only on replicaset primary nodes, it is initialized on stepUp * and terminated on stepDown. */ class BalancerStatsRegistry : public ReplicaSetAwareServiceShardSvr { BalancerStatsRegistry(const BalancerStatsRegistry&) = delete; BalancerStatsRegistry& operator=(const BalancerStatsRegistry&) = delete; public: BalancerStatsRegistry() = default; /** * Obtains the service-wide instance. */ static BalancerStatsRegistry* get(ServiceContext* serviceContext); static BalancerStatsRegistry* get(OperationContext* opCtx); /** * Non blocking initialization. Performs an asyncronous initialization of this registry. */ void initializeAsync(OperationContext* opCtx); void terminate(); /** * Update orphan document count for a specific collection. * `delta` is the increment/decrement that will be applied to the current cached count. * * If the registy is not initialized this function will be a noop. */ void updateOrphansCount(const UUID& collectionUUID, long long delta); void onRangeDeletionTaskInsertion(const UUID& collectionUUID, long long numOrphanDocs); void onRangeDeletionTaskDeletion(const UUID& collectionUUID, long long numOrphanDocs); long long getCollNumOrphanDocs(const UUID& collectionUUID) const; /** * Retrieves the numOrphanDocs from the balancer stats registry if initialized or runs an * aggregation on disk if not. */ long long getCollNumOrphanDocsFromDiskIfNeeded(OperationContext* opCtx, const UUID& collectionUUID) const; private: void onInitialDataAvailable(OperationContext* opCtx, bool isMajorityDataAvailable) override final {} void onStepUpBegin(OperationContext* opCtx, long long term) override final {} void onBecomeArbiter() override final {} void onShutdown() override final {} void onStartup(OperationContext* opCtx) override final; void onStepUpComplete(OperationContext* opCtx, long long term) override final; void onStepDown() override final; void _loadOrphansCount(OperationContext* opCtx); bool _isInitialized() const { return _state.load() == State::kInitialized; } struct CollectionStats { // Number of orphan documents for this collection long long numOrphanDocs; // Number of range deletion tasks long long numRangeDeletionTasks; }; enum class State { kPrimaryIdle, // The node is primary but the registry is not initialzed kInitializing, kInitialized, kTerminating, kSecondary, }; mutable Mutex _stateMutex = MONGO_MAKE_LATCH("BalancerStatsRegistry::_stateMutex"); AtomicWord _state{State::kSecondary}; ServiceContext::UniqueOperationContext _initOpCtxHolder; mutable Mutex _mutex = MONGO_MAKE_LATCH("BalancerStatsRegistry::_mutex"); // Map containing all the currently cached collection stats stdx::unordered_map _collStatsMap; // Thread pool used for asyncronous initialization std::shared_ptr _threadPool; }; } // namespace mongo