YugabyteDB (2.13.1.0-b60, 21121d69985fbf76aa6958d8f04a9bfa936293b5)

// Licensed to the Apache Software Foundation (ASF) under one

// or more contributor license agreements.  See the NOTICE file

// distributed with this work for additional information

// regarding copyright ownership.  The ASF licenses this file

// to you under the Apache License, Version 2.0 (the

// "License"); you may not use this file except in compliance

// with the License.  You may obtain a copy of the License at

//

//   http://www.apache.org/licenses/LICENSE-2.0

//

// Unless required by applicable law or agreed to in writing,

// software distributed under the License is distributed on an

// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY

// KIND, either express or implied.  See the License for the

// specific language governing permissions and limitations

// under the License.

//

// The following only applies to changes made to this file as part of YugaByte development.

//

// Portions Copyright (c) YugaByte, Inc.

//

// Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except

// in compliance with the License.  You may obtain a copy of the License at

//

// http://www.apache.org/licenses/LICENSE-2.0

//

// Unless required by applicable law or agreed to in writing, software distributed under the License

// is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express

// or implied.  See the License for the specific language governing permissions and limitations

// under the License.

//

#ifndef YB_CONSENSUS_LOG_H_

#define YB_CONSENSUS_LOG_H_

#include <pthread.h>

#include <sys/types.h>

#include <atomic>

#include <condition_variable>

#include <map>

#include <memory>

#include <mutex>

#include <string>

#include <vector>

#include <boost/atomic.hpp>

#include <glog/logging.h>

#include "yb/common/common_fwd.h"

#include "yb/consensus/consensus_fwd.h"

#include "yb/consensus/log_util.h"

#include "yb/fs/fs_manager.h"

#include "yb/gutil/macros.h"

#include "yb/gutil/ref_counted.h"

#include "yb/gutil/spinlock.h"

#include "yb/util/status_fwd.h"

#include "yb/util/locks.h"

#include "yb/util/monotime.h"

#include "yb/util/mutex.h"

#include "yb/util/opid.h"

#include "yb/util/promise.h"

#include "yb/util/shared_lock.h"

#include "yb/util/status_callback.h"

#include "yb/util/threadpool.h"

namespace yb {

class MetricEntity;

class ThreadPool;

namespace cdc {

class CDCServiceTestMaxRentionTime_TestLogRetentionByOpId_MaxRentionTime_Test;

class CDCServiceTestMinSpace_TestLogRetentionByOpId_MinSpace_Test;

}

namespace log {

YB_STRONGLY_TYPED_BOOL(CreateNewSegment);

YB_DEFINE_ENUM(

    SegmentAllocationState,

    (kAllocationNotStarted)  // No segment allocation requested

    (kAllocationInProgress)  // Next segment allocation started

    (kAllocationFinished)    // Next segment ready

);

// Log interface, inspired by Raft's (logcabin) Log. Provides durability to YugaByte as a normal

// Write Ahead Log and also plays the role of persistent storage for the consensus state machine.

//

// Note: This class is not thread safe, the caller is expected to synchronize Log::Reserve() and

// Log::Append() calls.

//

// Log uses group commit to improve write throughput and latency without compromising ordering and

// durability guarantees.

//

// To add operations to the log, the caller must obtain the lock and call Reserve() with the

// collection of operations to be added. Then, the caller may release the lock and call

// AsyncAppend(). Reserve() reserves a slot on a queue for the log entry; AsyncAppend() indicates

// that the entry in the slot is safe to write to disk and adds a callback that will be invoked once

// the entry is written and synchronized to disk.

//

// For sample usage see mt-log-test.cc

//

// Methods on this class are _not_ thread-safe and must be externally synchronized unless otherwise

// noted.

//

// Note: The Log needs to be Close()d before any log-writing class is destroyed, otherwise the Log

// might hold references to these classes to execute the callbacks after each write.

class Log : public RefCountedThreadSafe<Log> {

 public:

  static const Status kLogShutdownStatus;

  // Opens or continues a log and sets 'log' to the newly built Log.

  // After a successful Open() the Log is ready to receive entries, if create_new_segment is true.

  static CHECKED_STATUS Open(const LogOptions &options,

                             const std::string& tablet_id,

                             const std::string& wal_dir,

                             const std::string& peer_uuid,

                             const Schema& schema,

                             uint32_t schema_version,

                             const scoped_refptr<MetricEntity>& table_metric_entity,

                             const scoped_refptr<MetricEntity>& tablet_metric_entity,

                             ThreadPool *append_thread_pool,

                             ThreadPool* allocation_thread_pool,

                             int64_t cdc_min_replicated_index,

                             scoped_refptr<Log> *log,

                             CreateNewSegment create_new_segment = CreateNewSegment::kTrue);

  ~Log();

  // Reserves a spot in the log's queue for 'entry_batch'.

  //

  // 'reserved_entry' is initialized by this method and any resources associated with it will be

  // released in AsyncAppend().  In order to ensure correct ordering of operations across multiple

  // threads, calls to this method must be externally synchronized.

  //

  // WARNING: the caller _must_ call AsyncAppend() or else the log will "stall" and will never be

  // able to make forward progress.

  void Reserve(LogEntryTypePB type, LogEntryBatchPB* entry_batch, LogEntryBatch** reserved_entry);

  // Asynchronously appends 'entry' to the log. Once the append completes and is synced, 'callback'

  // will be invoked.

  CHECKED_STATUS AsyncAppend(LogEntryBatch* entry,

                             const StatusCallback& callback);

  CHECKED_STATUS TEST_AsyncAppendWithReplicates(

      LogEntryBatch* entry, const ReplicateMsgs& replicates, const StatusCallback& callback);

  // Synchronously append a new entry to the log.  Log does not take ownership of the passed

  // 'entry'. If skip_wal_write is true, only update consensus metadata and LogIndex, skip write

  // to wal.

  // TODO get rid of this method, transition to the asynchronous API.

  CHECKED_STATUS Append(LogEntryPB* entry,

                        LogEntryMetadata entry_metadata,

                        bool skip_wal_write = false);

  // Append the given set of replicate messages, asynchronously.  This requires that the replicates

  // have already been assigned OpIds.

  CHECKED_STATUS AsyncAppendReplicates(const ReplicateMsgs& replicates, const OpId& committed_op_id,

                                       RestartSafeCoarseTimePoint batch_mono_time,

                                       const StatusCallback& callback);

  // Blocks the current thread until all the entries in the log queue are flushed and fsynced (if

  // fsync of log entries is enabled).

  CHECKED_STATUS WaitUntilAllFlushed();

  // The closure submitted to allocation_pool_ to allocate a new segment.

  void SegmentAllocationTask();

  // Syncs all state and closes the log.

  CHECKED_STATUS Close();

  // Return true if there is any on-disk data for the given tablet.

  static bool HasOnDiskData(FsManager* fs_manager, const std::string& tablet_id);

  // Delete all WAL data from the log associated with this tablet.

  // REQUIRES: The Log must be closed.

  static CHECKED_STATUS DeleteOnDiskData(Env* env,

                                         const std::string& tablet_id,

                                         const std::string& wal_dir,

                                         const std::string& peer_uuid);

  // Returns a reader that is able to read through the previous segments. The reader pointer is

  // guaranteed to be live as long as the log itself is initialized and live.

  LogReader* GetLogReader() const;

  CHECKED_STATUS GetSegmentsSnapshot(SegmentSequence* segments) const;

  void SetMaxSegmentSizeForTests(uint64_t max_segment_size) {

    max_segment_size_ = max_segment_size;

  }

  void DisableSync() {

    sync_disabled_ = true;

  }

  // If we previous called DisableSync(), we should restore the default behavior and then call

  // Sync() which will perform the actual syncing if required.

  CHECKED_STATUS ReEnableSyncIfRequired() {

    sync_disabled_ = false;

    return Sync();

  }

  // Get ID of tablet.

  const std::string& tablet_id() const {

    return tablet_id_;

  }

  // Gets the last-used OpId written to the log.  If no entry has ever been written to the log,

  // returns (0, 0)

  yb::OpId GetLatestEntryOpId() const;

  int64_t GetMinReplicateIndex() const;

  // Runs the garbage collector on the set of previous segments. Segments that only refer to in-mem

  // state that has been flushed are candidates for garbage collection.

  //

  // 'min_op_idx' is the minimum operation index required to be retained.  If successful, num_gced

  // is set to the number of deleted log segments.

  //

  // This method is thread-safe.

  CHECKED_STATUS GC(int64_t min_op_idx, int* num_gced);

  // Computes the amount of bytes that would have been GC'd if Log::GC had been called.

  CHECKED_STATUS GetGCableDataSize(int64_t min_op_idx, int64_t* total_size) const;

  // Returns the file system location of the currently active WAL segment.

  const WritableLogSegment* ActiveSegmentForTests() const {

    return active_segment_.get();

  }

  // If active segment is not empty, forces the Log to allocate a new segment and roll over.

  // This can be used to make sure all entries appended up to this point are available in closed,

  // readable segments. Note that this assumes there is already a valid active_segment_.

  CHECKED_STATUS AllocateSegmentAndRollOver();

  // For a log created with CreateNewSegment::kFalse, this is used to finish log initialization by

  // allocating a new segment.

  CHECKED_STATUS EnsureInitialNewSegmentAllocated();

  // Returns the total size of the current segments, in bytes.

  // Returns 0 if the log is shut down.

  uint64_t OnDiskSize();

  // Set the schema for the _next_ log segment.

  //

  // This method is thread-safe.

  void SetSchemaForNextLogSegment(const Schema& schema, uint32_t version);

  void set_wal_retention_secs(uint32_t wal_retention_secs);

  uint32_t wal_retention_secs() const;

  // Waits until specified op id is added to log.

  // Returns current op id after waiting, which could be greater than or equal to specified op id.

  //

  // On timeout returns default constructed OpId.

  yb::OpId WaitForSafeOpIdToApply(const yb::OpId& op_id, MonoDelta duration = MonoDelta());

  // Return a readable segment with the given sequence number, or nullptr if it

  // cannot be found (e.g. if it has already been GCed).

  scoped_refptr<ReadableLogSegment> GetSegmentBySequenceNumber(int64_t seq) const;

  void TEST_SetSleepDuration(const std::chrono::nanoseconds& duration) {

    sleep_duration_.store(duration, std::memory_order_release);

  }

  void TEST_SetAllOpIdsSafe(bool value) {

    all_op_ids_safe_ = value;

  }

  uint64_t active_segment_sequence_number() const;

  CHECKED_STATUS TEST_SubmitFuncToAppendToken(const std::function<void()>& func);

  // Returns the number of segments.

  size_t num_segments() const;

  const std::string& LogPrefix() const {

    return log_prefix_;

  }

  std::string wal_dir() const {

    return wal_dir_;

  }

  void set_cdc_min_replicated_index(int64_t cdc_min_replicated_index) {

    // TODO: check that the passed index is greater than the current index.

    cdc_min_replicated_index_.store(cdc_min_replicated_index, std::memory_order_release);

  }

  int64_t cdc_min_replicated_index() {

    return cdc_min_replicated_index_.load(std::memory_order_acquire);

  }

  CHECKED_STATUS FlushIndex();

  // Copies log to a new dir.

  // Flushes necessary files and uses hard links where it is safe.

  CHECKED_STATUS CopyTo(const std::string& dest_wal_dir);

  // Waits until all entries flushed, then reset last received op id to specified one.

  CHECKED_STATUS ResetLastSyncedEntryOpId(const OpId& op_id);

 private:

  friend class LogTest;

  friend class LogTestBase;

  FRIEND_TEST(LogTest, TestMultipleEntriesInABatch);

  FRIEND_TEST(LogTest, TestReadLogWithReplacedReplicates);

  FRIEND_TEST(LogTest, TestWriteAndReadToAndFromInProgressSegment);

  FRIEND_TEST(LogTest, TestLogMetrics);

  FRIEND_TEST(cdc::CDCServiceTestMaxRentionTime, TestLogRetentionByOpId_MaxRentionTime);

  FRIEND_TEST(cdc::CDCServiceTestMinSpace, TestLogRetentionByOpId_MinSpace);

  class Appender;

  // Log state.

  enum LogState {

    kLogInitialized,

    kLogWriting,

    kLogClosed

  };

  Log(LogOptions options,

      std::string wal_dir,

      std::string tablet_id,

      std::string peer_uuid,

      const Schema& schema,

      uint32_t schema_version,

      const scoped_refptr<MetricEntity>& table_metric_entity,

      const scoped_refptr<MetricEntity>& tablet_metric_entity,

      ThreadPool* append_thread_pool,

      ThreadPool* allocation_thread_pool,

      CreateNewSegment create_new_segment = CreateNewSegment::kTrue);

  Env* get_env() {

    return options_.env;

  }

  // Initializes a new one or continues an existing log.

  CHECKED_STATUS Init();

  // Make segments roll over. Note this assumes there was an existing valid active_segment_ we are

  // rolling over from.

  CHECKED_STATUS RollOver();

  // Writes the footer and closes the current segment.

  CHECKED_STATUS CloseCurrentSegment();

  // Sets 'out' to a newly created temporary file (see Env::NewTempWritableFile()) for a placeholder

  // segment. Sets 'result_path' to the fully qualified path to the unique filename created for the

  // segment.

  CHECKED_STATUS CreatePlaceholderSegment(const WritableFileOptions& opts,

                                          std::string* result_path,

                                          std::shared_ptr<WritableFile>* out);

  // Creates a new WAL segment on disk, writes the next_segment_header_ to disk as the header, and

  // sets active_segment_ to point to this new segment.

  CHECKED_STATUS SwitchToAllocatedSegment();

  // Preallocates the space for a new segment.

  CHECKED_STATUS PreAllocateNewSegment();

  // Returns the desired size for the next log segment to be created.

  uint64_t NextSegmentDesiredSize();

  // Writes serialized contents of 'entry' to the log. Called inside AppenderThread. If

  // 'caller_owns_operation' is true, then the 'operation' field of the entry will be released after

  // the entry is appended. If skip_wal_write is true, only update consensus metadata and LogIndex,

  // skip WAL write.

  //

  // TODO once Append() is removed, 'caller_owns_operation' and associated logic will no longer be

  // needed.

  CHECKED_STATUS DoAppend(

      LogEntryBatch* entry, bool caller_owns_operation = true, bool skip_wal_write = false);

  // Update footer_builder_ to reflect the log indexes seen in 'batch'.

  void UpdateFooterForBatch(LogEntryBatch* batch);

  // Update the LogIndex to include entries for the replicate messages found in 'batch'. The index

  // entry points to the offset 'start_offset' in the current log segment.

  CHECKED_STATUS UpdateIndexForBatch(const LogEntryBatch& batch);

  // Replaces the last "empty" segment in 'log_reader_', i.e. the one currently being written to, by

  // the same segment once properly closed.

  CHECKED_STATUS ReplaceSegmentInReaderUnlocked();

  CHECKED_STATUS Sync();

  // Helper method to get the segment sequence to GC based on the provided min_op_idx.

  CHECKED_STATUS GetSegmentsToGCUnlocked(int64_t min_op_idx, SegmentSequence* segments_to_gc) const;

  // Kick off an asynchronous task that pre-allocates a new log-segment, setting

  // 'allocation_status_'. To wait for the result of the task, use allocation_status_.Get().

  CHECKED_STATUS AsyncAllocateSegment();

  SegmentAllocationState allocation_state() {

    return allocation_state_.load(std::memory_order_acquire);

  }

  LogEntryBatch* ReserveMarker(LogEntryTypePB type);

  LogOptions options_;

  // The dir path where the write-ahead log for this tablet is stored.

  std::string wal_dir_;

  // The ID of the tablet this log is dedicated to.

  std::string tablet_id_;

  // Peer this log is dedicated to.

  std::string peer_uuid_;

  // Lock to protect modifications to schema_ and schema_version_.

  mutable rw_spinlock schema_lock_;

  // The current schema of the tablet this log is dedicated to.

  std::unique_ptr<Schema> schema_;

  // The schema version

  uint32_t schema_version_;

  // The currently active segment being written.

  std::unique_ptr<WritableLogSegment> active_segment_;

  // The current (active) segment sequence number. Initialized in the Log constructor based on

  // LogOptions.

  std::atomic<uint64_t> active_segment_sequence_number_;

  // The writable file for the next allocated segment

  std::shared_ptr<WritableFile> next_segment_file_;

  // The path for the next allocated segment.

  std::string next_segment_path_;

  // Lock to protect mutations to log_state_ and other shared state variables.

  mutable percpu_rwlock state_lock_;

  LogState log_state_;

  // A reader for the previous segments that were not yet GC'd.

  std::unique_ptr<LogReader> reader_;

  // Index which translates between operation indexes and the position of the operation in the log.

  scoped_refptr<LogIndex> log_index_;

  // Lock for notification of last_synced_entry_op_id_ changes.

  mutable std::mutex last_synced_entry_op_id_mutex_;

  mutable std::condition_variable last_synced_entry_op_id_cond_;

  // The last known OpId for a REPLICATE message appended and synced to this log (any segment).

  // NOTE: this op is not necessarily durable unless gflag durable_wal_write is true.

  boost::atomic<yb::OpId> last_synced_entry_op_id_{yb::OpId()};

  // The last know OpId for a REPLICATE message appended to this log (any segment).

  // This variable is not accessed concurrently.

  yb::OpId last_appended_entry_op_id_;

  yb::OpId last_submitted_op_id_;

  // A footer being prepared for the current segment.  When the segment is closed, it will be

  // written.

  LogSegmentFooterPB footer_builder_;

  // The maximum segment size, in bytes.

  uint64_t max_segment_size_;

  // The maximum segment size we want for the current WAL segment, in bytes.  This value keeps

  // doubling (for each subsequent WAL segment) till it gets to max_segment_size_.

  uint64_t cur_max_segment_size_;

  // Appender manages a TaskStream writing to the log. We will use one taskstream per tablet.

  std::unique_ptr<Appender> appender_;

  // A thread pool for asynchronously pre-allocating new log segments.

  std::unique_ptr<ThreadPoolToken> allocation_token_;

  // If true, sync on all appends.

  bool durable_wal_write_;

  // If non-zero, sync every interval of time.

  MonoDelta interval_durable_wal_write_;

  // If non-zero, sync if more than given amount of data to sync.

  int32_t bytes_durable_wal_write_mb_;

  // Keeps track of oldest entry which needs to be synced.

  MonoTime periodic_sync_earliest_unsync_entry_time_ = MonoTime::kMin;

  // For periodic sync, indicates if there are entries to be sync'ed.

  std::atomic<bool> periodic_sync_needed_ = {false};

  // For periodic sync, indicates number of bytes which need to be sync'ed.

  size_t periodic_sync_unsynced_bytes_ = 0;

  // If true, ignore the 'durable_wal_write_' flags above.  This is used to disable fsync during

  // bootstrap.

  bool sync_disabled_;

  // The status of the most recent log-allocation action.

  Promise<Status> allocation_status_;

  std::atomic<SegmentAllocationState> allocation_state_;

  scoped_refptr<MetricEntity> table_metric_entity_;

  scoped_refptr<MetricEntity> tablet_metric_entity_;

  std::unique_ptr<LogMetrics> metrics_;

  // The cached on-disk size of the log, used to track its size even if it has been closed.

  std::atomic<uint64_t> on_disk_size_;

  // Listener that will be invoked after new entry was appended to the log.

  std::function<void()> post_append_listener_;

  // Used in tests delay writing log entries.

  std::atomic<std::chrono::nanoseconds> sleep_duration_{std::chrono::nanoseconds(0)};

  // Used in tests to declare all operations as safe.

  bool all_op_ids_safe_ = false;

  const std::string log_prefix_;

  std::atomic<uint32_t> wal_retention_secs_{0};

  // Minimum replicate index for the current log being written. Used for CDC read initialization.

  std::atomic<int64_t> min_replicate_index_{-1};

  // The current replicated index that CDC has read.  Used for CDC read cache optimization.

  std::atomic<int64_t> cdc_min_replicated_index_{std::numeric_limits<int64_t>::max()};

  std::mutex log_copy_mutex_;

  // Used by GetSegmentsToGCUnlocked() as an anchor.

  int64_t log_copy_min_index_ GUARDED_BY(state_lock_) = std::numeric_limits<int64_t>::max();

  CreateNewSegment create_new_segment_at_start_;

  DISALLOW_COPY_AND_ASSIGN(Log);

};

}  // namespace log

}  // namespace yb

#endif /* YB_CONSENSUS_LOG_H_ */