path: root/src/mongo/db/repl/oplog_applier.h

/**
 *    Copyright (C) 2018-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
 *    <http://www.mongodb.com/licensing/server-side-public-license>.
 *
 *    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 <boost/optional.hpp>
#include <vector>

#include "mongo/base/status.h"
#include "mongo/base/status_with.h"
#include "mongo/db/repl/oplog_buffer.h"
#include "mongo/db/repl/oplog_entry.h"
#include "mongo/db/repl/storage_interface.h"
#include "mongo/executor/task_executor.h"
#include "mongo/stdx/mutex.h"
#include "mongo/util/concurrency/thread_pool.h"
#include "mongo/util/functional.h"
#include "mongo/util/future.h"
#include "mongo/util/net/hostandport.h"

namespace mongo {
namespace repl {

/**
 * Applies oplog entries.
 * Reads from an OplogBuffer batches of operations that may be applied in parallel.
 */
class OplogApplier {
    OplogApplier(const OplogApplier&) = delete;
    OplogApplier& operator=(const OplogApplier&) = delete;

public:
    /**
     * Used to configure behavior of this OplogApplier.
     **/
    class Options {
    public:
        bool allowNamespaceNotFoundErrorsOnCrudOps = false;
        bool relaxUniqueIndexConstraints = false;
        bool skipWritesToOplog = false;

        // For initial sync only. If an update fails, the missing document is fetched from
        // this sync source to insert into the local collection.
        boost::optional<HostAndPort> missingDocumentSourceForInitialSync;

        // Used to determine which operations should be applied. Only initial sync will set this to
        // be something other than the null optime.
        OpTime beginApplyingOpTime = OpTime();

        // For replication recovery only. During replication rollback, this is used to keep track
        // of the stable timestamp from which we replay the oplog.
        boost::optional<Timestamp> stableTimestampForRecovery;
    };

    /**
     * Controls what can popped from the oplog buffer into a single batch of operations that can be
     * applied using multiApply().
     */
    class BatchLimits {
    public:
        size_t bytes = 0;
        size_t ops = 0;

        // If provided, the batch will not include any operations with timestamps after this point.
        // This is intended for implementing slaveDelay, so it should be some number of seconds
        // before now.
        boost::optional<Date_t> slaveDelayLatestTimestamp = {};

        // If non-null, the batch will include operations with timestamps either
        // before-and-including this point or after it, not both.
        Timestamp forceBatchBoundaryAfter;
    };

    // Used to report oplog application progress.
    class Observer;

    using Operations = std::vector<OplogEntry>;

    // Used by SyncTail to access batching logic.
    using GetNextApplierBatchFn = std::function<StatusWith<OplogApplier::Operations>(
        OperationContext* opCtx, const BatchLimits& batchLimits)>;

    /**
     * Creates thread pool for writer tasks.
     */
    static std::unique_ptr<ThreadPool> makeWriterPool();
    static std::unique_ptr<ThreadPool> makeWriterPool(int threadCount);

    /**
     * Returns maximum number of operations in each batch that can be applied using multiApply().
     */
    static std::size_t getBatchLimitOperations();

    /**
     * Calculates batch limit size (in bytes) using the maximum capped collection size of the oplog
     * size.
     * Batches are limited to 10% of the oplog.
     */
    static std::size_t calculateBatchLimitBytes(OperationContext* opCtx,
                                                StorageInterface* storageInterface);

    /**
     * Constructs this OplogApplier with specific options.
     * Obtains batches of operations from the OplogBuffer to apply.
     * Reports oplog application progress using the Observer.
     */
    OplogApplier(executor::TaskExecutor* executor, OplogBuffer* oplogBuffer, Observer* observer);

    virtual ~OplogApplier() = default;

    /**
     * Returns this applier's buffer.
     */
    OplogBuffer* getBuffer() const;

    /**
     * Starts this OplogApplier.
     * Use the Future object to be notified when this OplogApplier has finished shutting down.
     */
    Future<void> startup();

    /**
     * Starts the shutdown process for this OplogApplier.
     * It is safe to call shutdown() multiplie times.
     */
    void shutdown();

    /**
     * Returns true if we are shutting down.
     */
    bool inShutdown() const;

    /**
     * Pushes operations read into oplog buffer.
     * Accepts both Operations (OplogEntry) and OplogBuffer::Batch (BSONObj) iterators.
     * This supports current implementations of OplogFetcher and OplogBuffer which work in terms of
     * BSONObj.
     */
    void enqueue(OperationContext* opCtx,
                 Operations::const_iterator begin,
                 Operations::const_iterator end);
    void enqueue(OperationContext* opCtx,
                 OplogBuffer::Batch::const_iterator begin,
                 OplogBuffer::Batch::const_iterator end);

    /**
     * Returns a new batch of ops to apply.
     * A batch may consist of:
     *     at most "BatchLimits::ops" OplogEntries
     *     at most "BatchLimits::bytes" worth of OplogEntries
     *     only OplogEntries from before the "BatchLimits::slaveDelayLatestTimestamp" point
     *     a single command OplogEntry (excluding applyOps, which are grouped with CRUD ops)
     */
    StatusWith<Operations> getNextApplierBatch(OperationContext* opCtx,
                                               const BatchLimits& batchLimits);

    /**
     * Applies a batch of oplog entries by writing the oplog entries to the local oplog and then
     * using a set of threads to apply the operations.
     *
     * If the batch application is successful, returns the optime of the last op applied, which
     * should be the last op in the batch.
     * Returns ErrorCodes::CannotApplyOplogWhilePrimary if the node has become primary.
     *
     * To provide crash resilience, this function will advance the persistent value of 'minValid'
     * to at least the last optime of the batch. If 'minValid' is already greater than or equal
     * to the last optime of this batch, it will not be updated.
     *
     * TODO: remove when enqueue() is implemented.
     */
    StatusWith<OpTime> multiApply(OperationContext* opCtx, Operations ops);

private:
    /**
     * Pops the operation at the front of the OplogBuffer.
     */
    void _consume(OperationContext* opCtx, OplogBuffer* oplogBuffer);

    /**
     * Called from startup() to run oplog application loop.
     * Currently applicable to steady state replication only.
     * Implemented in subclasses but not visible otherwise.
     */
    virtual void _run(OplogBuffer* oplogBuffer) = 0;

    /**
     * Called from shutdown to signals oplog application loop to stop running.
     * Currently applicable to steady state replication only.
     * Implemented in subclasses but not visible otherwise.
     */
    virtual void _shutdown() = 0;

    /**
     * Called from multiApply() to apply a batch of operations in parallel.
     * Implemented in subclasses but not visible otherwise.
     */
    virtual StatusWith<OpTime> _multiApply(OperationContext* opCtx, Operations ops) = 0;

    // Used to schedule task for oplog application loop.
    // Not owned by us.
    executor::TaskExecutor* const _executor;

    // Not owned by us.
    OplogBuffer* const _oplogBuffer;

    // Not owned by us.
    Observer* const _observer;

    // Protects member data of OplogApplier.
    mutable stdx::mutex _mutex;

    // Set to true if shutdown() has been called.
    bool _inShutdown = false;
};

/**
 * The OplogApplier reports its progress using the Observer interface.
 */
class OplogApplier::Observer {
public:
    virtual ~Observer() = default;

    /**
     * Called when the OplogApplier is ready to start applying a batch of operations read from the
     * OplogBuffer.
     **/
    virtual void onBatchBegin(const OplogApplier::Operations& operations) = 0;

    /**
     * When the OplogApplier has completed applying a batch of operations, it will call this
     * function to report the last optime applied on success. Any errors during oplog application
     * will also be here.
     */
    virtual void onBatchEnd(const StatusWith<OpTime>& lastOpTimeApplied,
                            const OplogApplier::Operations& operations) = 0;

    /**
     * Called when documents are fetched and inserted into the collection in order to
     * apply an update operation.
     * Applies to initial sync only.
     *
     * TODO: Delegate fetching behavior to OplogApplier owner.
     */
    using FetchInfo = std::pair<OplogEntry, BSONObj>;
    virtual void onMissingDocumentsFetchedAndInserted(
        const std::vector<FetchInfo>& documentsFetchedAndInserted) = 0;
};

}  // namespace repl
}  // namespace mongo