@nxtedition/rocksdb 15.4.0 → 15.5.0

package/deps/rocksdb/rocksdb/monitoring/perf_context.cc

@@ -73,6 +73,7 @@ struct PerfContextByLevelInt {
   defCmd(compressed_sec_cache_compressed_bytes)    \
   defCmd(block_checksum_time)                      \
   defCmd(block_decompress_time)                    \
+  defCmd(block_decompress_count)                   \
   defCmd(get_read_bytes)                           \
   defCmd(multiget_read_bytes)                      \
   defCmd(iter_read_bytes)                          \
@@ -155,7 +156,12 @@ struct PerfContextByLevelInt {
   defCmd(decrypt_data_nanos)                       \
   defCmd(number_async_seek)                        \
   defCmd(file_ingestion_nanos)                     \
-  defCmd(file_ingestion_blocking_live_writes_nanos)
+  defCmd(file_ingestion_blocking_live_writes_nanos)\
+  defCmd(data_block_read_byte)                     \
+  defCmd(index_block_read_byte)                    \
+  defCmd(filter_block_read_byte)                   \
+  defCmd(compression_dict_block_read_byte)         \
+  defCmd(metadata_block_read_byte)
 // clang-format on
 
 struct PerfContextInt {
@@ -259,10 +265,10 @@ void PerfContext::Reset() {
 #endif
 }
 
-void PerfContextByLevel::Reset(){
+void PerfContextByLevel::Reset() {
 #ifndef NPERF_CONTEXT
 #define EMIT_FIELDS(x) x = 0;
-    DEF_PERF_CONTEXT_LEVEL_METRICS(EMIT_FIELDS)
+  DEF_PERF_CONTEXT_LEVEL_METRICS(EMIT_FIELDS)
 #undef EMIT_FIELDS
 #endif
 }

package/deps/rocksdb/rocksdb/monitoring/statistics.cc

@@ -93,6 +93,7 @@ const std::vector<std::pair<Tickers, std::string>> TickersNameMap = {
     {COMPACTION_OPTIMIZED_DEL_DROP_OBSOLETE,
      "rocksdb.compaction.optimized.del.drop.obsolete"},
     {COMPACTION_CANCELLED, "rocksdb.compaction.cancelled"},
+    {COMPACTION_ABORTED, "rocksdb.compaction.aborted"},
     {NUMBER_KEYS_WRITTEN, "rocksdb.number.keys.written"},
     {NUMBER_KEYS_READ, "rocksdb.number.keys.read"},
     {NUMBER_KEYS_UPDATED, "rocksdb.number.keys.updated"},
@@ -169,8 +170,8 @@ const std::vector<std::pair<Tickers, std::string>> TickersNameMap = {
     {BLOB_DB_NUM_KEYS_READ, "rocksdb.blobdb.num.keys.read"},
     {BLOB_DB_BYTES_WRITTEN, "rocksdb.blobdb.bytes.written"},
     {BLOB_DB_BYTES_READ, "rocksdb.blobdb.bytes.read"},
-    {BLOB_DB_WRITE_INLINED, "rocksdb.blobdb.write.inlined"},
-    {BLOB_DB_WRITE_INLINED_TTL, "rocksdb.blobdb.write.inlined.ttl"},
+    {BLOB_DB_WRITE_INLINED_DEPRECATED, "rocksdb.blobdb.write.inlined"},
+    {BLOB_DB_WRITE_INLINED_TTL_DEPRECATED, "rocksdb.blobdb.write.inlined.ttl"},
     {BLOB_DB_WRITE_BLOB, "rocksdb.blobdb.write.blob"},
     {BLOB_DB_WRITE_BLOB_TTL, "rocksdb.blobdb.write.blob.ttl"},
     {BLOB_DB_BLOB_FILE_BYTES_WRITTEN, "rocksdb.blobdb.blob.file.bytes.written"},
@@ -280,6 +281,21 @@ const std::vector<std::pair<Tickers, std::string>> TickersNameMap = {
     {NUMBER_WBWI_INGEST, "rocksdb.number.wbwi.ingest"},
     {SST_USER_DEFINED_INDEX_LOAD_FAIL_COUNT,
      "rocksdb.sst.user.defined.index.load.fail.count"},
+    {MULTISCAN_PREPARE_CALLS, "rocksdb.multiscan.prepare.calls"},
+    {MULTISCAN_PREPARE_ERRORS, "rocksdb.multiscan.prepare.errors"},
+    {MULTISCAN_BLOCKS_PREFETCHED, "rocksdb.multiscan.blocks.prefetched"},
+    {MULTISCAN_BLOCKS_FROM_CACHE, "rocksdb.multiscan.blocks.from.cache"},
+    {MULTISCAN_PREFETCH_BYTES, "rocksdb.multiscan.prefetch.bytes"},
+    {MULTISCAN_PREFETCH_BLOCKS_WASTED,
+     "rocksdb.multiscan.prefetch.blocks.wasted"},
+    {MULTISCAN_IO_REQUESTS, "rocksdb.multiscan.io.requests"},
+    {MULTISCAN_IO_COALESCED_NONADJACENT,
+     "rocksdb.multiscan.io.coalesced.nonadjacent"},
+    {MULTISCAN_SEEK_ERRORS, "rocksdb.multiscan.seek.errors"},
+    {PREFETCH_MEMORY_BYTES_GRANTED, "rocksdb.prefetch.memory.bytes.granted"},
+    {PREFETCH_MEMORY_BYTES_RELEASED, "rocksdb.prefetch.memory.bytes.released"},
+    {PREFETCH_MEMORY_REQUESTS_BLOCKED,
+     "rocksdb.prefetch.memory.requests.blocked"},
 };
 
 const std::vector<std::pair<Histograms, std::string>> HistogramsNameMap = {
@@ -354,6 +370,9 @@ const std::vector<std::pair<Histograms, std::string>> HistogramsNameMap = {
     {NUM_OP_PER_TRANSACTION, "rocksdb.num.op.per.transaction"},
     {MULTISCAN_PREPARE_ITERATORS,
      "rocksdb.multiscan.op.prepare.iterators.micros"},
+    {MULTISCAN_PREPARE_MICROS, "rocksdb.multiscan.prepare.micros"},
+    {MULTISCAN_BLOCKS_PER_PREPARE, "rocksdb.multiscan.blocks.per.prepare"},
+    {BLOCK_KEY_DISTRIBUTION_CV, "rocksdb.block.key.distribution.cv"},
 };
 
 std::shared_ptr<Statistics> CreateDBStatistics() {

package/deps/rocksdb/rocksdb/monitoring/stats_history_test.cc

@@ -185,7 +185,7 @@ TEST_F(StatsHistoryTest, GetStatsHistoryInMemory) {
 
 TEST_F(StatsHistoryTest, InMemoryStatsHistoryPurging) {
   constexpr int kPeriodSec = 1;
-  constexpr int kEstimatedOneSliceSize = 16000;
+  constexpr int kEstimatedOneSliceSize = 22100;
 
   Options options;
   options.create_if_missing = true;
@@ -277,7 +277,7 @@ TEST_F(StatsHistoryTest, InMemoryStatsHistoryPurging) {
   // If `slice_count == 0` when new statistics are added, consider increasing
   // `kEstimatedOneSliceSize`
   ASSERT_EQ(slice_count, 1);
-  ASSERT_TRUE(stats_history_size_reopen < 16000 &&
+  ASSERT_TRUE(stats_history_size_reopen < kEstimatedOneSliceSize &&
               stats_history_size_reopen > 0);
   ASSERT_TRUE(stats_count_reopen < stats_count && stats_count_reopen > 0);
   Close();

package/deps/rocksdb/rocksdb/options/cf_options.cc

@@ -30,6 +30,7 @@
 #include "rocksdb/utilities/object_registry.h"
 #include "rocksdb/utilities/options_type.h"
 #include "util/cast_util.h"
+#include "util/string_util.h"
 
 // NOTE: in this file, many option flags that were deprecated
 // and removed from the rest of the code have to be kept here
@@ -310,6 +311,14 @@ static std::unordered_map<std::string, OptionTypeInfo>
         {"trivial_copy_buffer_size",
          {offsetof(struct CompactionOptionsFIFO, trivial_copy_buffer_size),
           OptionType::kUInt64T, OptionVerificationType::kNormal,
+          OptionTypeFlags::kMutable}},
+        {"max_data_files_size",
+         {offsetof(struct CompactionOptionsFIFO, max_data_files_size),
+          OptionType::kUInt64T, OptionVerificationType::kNormal,
+          OptionTypeFlags::kMutable}},
+        {"use_kv_ratio_compaction",
+         {offsetof(struct CompactionOptionsFIFO, use_kv_ratio_compaction),
+          OptionType::kBoolean, OptionVerificationType::kNormal,
           OptionTypeFlags::kMutable}}};
 
 static std::unordered_map<std::string, OptionTypeInfo>
@@ -910,6 +919,11 @@ static std::unordered_map<std::string, OptionTypeInfo>
          {offsetof(struct ImmutableCFOptions, cf_allow_ingest_behind),
           OptionType::kBoolean, OptionVerificationType::kNormal,
           OptionTypeFlags::kNone}},
+        {"memtable_batch_lookup_optimization",
+         {offsetof(struct ImmutableCFOptions,
+                   memtable_batch_lookup_optimization),
+          OptionType::kBoolean, OptionVerificationType::kNormal,
+          OptionTypeFlags::kNone}},
 };
 
 const std::string OptionsHelper::kCFOptionsName = "ColumnFamilyOptions";
@@ -1050,7 +1064,9 @@ ImmutableCFOptions::ImmutableCFOptions(const ColumnFamilyOptions& cf_options)
       blob_cache(cf_options.blob_cache),
       persist_user_defined_timestamps(
           cf_options.persist_user_defined_timestamps),
-      cf_allow_ingest_behind(cf_options.cf_allow_ingest_behind) {}
+      cf_allow_ingest_behind(cf_options.cf_allow_ingest_behind),
+      memtable_batch_lookup_optimization(
+          cf_options.memtable_batch_lookup_optimization) {}
 
 ImmutableOptions::ImmutableOptions() : ImmutableOptions(Options()) {}
 
@@ -1259,6 +1275,10 @@ void MutableCFOptions::Dump(Logger* log) const {
                  compaction_options_fifo.max_table_files_size);
   ROCKS_LOG_INFO(log, "compaction_options_fifo.allow_compaction : %d",
                  compaction_options_fifo.allow_compaction);
+  ROCKS_LOG_INFO(log, "compaction_options_fifo.max_data_files_size : %" PRIu64,
+                 compaction_options_fifo.max_data_files_size);
+  ROCKS_LOG_INFO(log, "compaction_options_fifo.use_kv_ratio_compaction : %d",
+                 compaction_options_fifo.use_kv_ratio_compaction);
 
   // Blob file related options
   ROCKS_LOG_INFO(log, "                        enable_blob_files: %s",

package/deps/rocksdb/rocksdb/options/cf_options.h

@@ -84,6 +84,8 @@ struct ImmutableCFOptions {
   bool persist_user_defined_timestamps;
 
   bool cf_allow_ingest_behind;
+
+  bool memtable_batch_lookup_optimization;
 };
 
 struct ImmutableOptions : public ImmutableDBOptions, public ImmutableCFOptions {

package/deps/rocksdb/rocksdb/options/customizable_test.cc

@@ -1281,8 +1281,6 @@ class MockSliceTransform : public SliceTransform {
   Slice Transform(const Slice& /*key*/) const override { return Slice(); }
 
   bool InDomain(const Slice& /*key*/) const override { return false; }
-
-  bool InRange(const Slice& /*key*/) const override { return false; }
 };
 
 class MockMemoryAllocator : public BaseMemoryAllocator {

package/deps/rocksdb/rocksdb/options/db_options.cc

@@ -136,6 +136,10 @@ static std::unordered_map<std::string, OptionTypeInfo>
          {offsetof(struct MutableDBOptions, manifest_preallocation_size),
           OptionType::kSizeT, OptionVerificationType::kNormal,
           OptionTypeFlags::kMutable}},
+        {"verify_manifest_content_on_close",
+         {offsetof(struct MutableDBOptions, verify_manifest_content_on_close),
+          OptionType::kBoolean, OptionVerificationType::kNormal,
+          OptionTypeFlags::kMutable}},
         {"daily_offpeak_time_utc",
          {offsetof(struct MutableDBOptions, daily_offpeak_time_utc),
           OptionType::kString, OptionVerificationType::kNormal,
@@ -230,6 +234,10 @@ static std::unordered_map<std::string, OptionTypeInfo>
          {offsetof(struct ImmutableDBOptions, paranoid_checks),
           OptionType::kBoolean, OptionVerificationType::kNormal,
           OptionTypeFlags::kNone}},
+        {"open_files_async",
+         {offsetof(struct ImmutableDBOptions, open_files_async),
+          OptionType::kBoolean, OptionVerificationType::kNormal,
+          OptionTypeFlags::kNone}},
         {"flush_verify_memtable_count",
          {offsetof(struct ImmutableDBOptions, flush_verify_memtable_count),
           OptionType::kBoolean, OptionVerificationType::kNormal,
@@ -259,9 +267,7 @@ static std::unordered_map<std::string, OptionTypeInfo>
           OptionType::kBoolean, OptionVerificationType::kNormal,
           OptionTypeFlags::kNone}},
         {"skip_checking_sst_file_sizes_on_db_open",
-         {offsetof(struct ImmutableDBOptions,
-                   skip_checking_sst_file_sizes_on_db_open),
-          OptionType::kBoolean, OptionVerificationType::kNormal,
+         {0, OptionType::kBoolean, OptionVerificationType::kDeprecated,
           OptionTypeFlags::kNone}},
         {"new_table_reader_for_compaction_inputs",
          {0, OptionType::kBoolean, OptionVerificationType::kDeprecated,
@@ -373,6 +379,11 @@ static std::unordered_map<std::string, OptionTypeInfo>
          {offsetof(struct ImmutableDBOptions, avoid_flush_during_recovery),
           OptionType::kBoolean, OptionVerificationType::kNormal,
           OptionTypeFlags::kNone}},
+        {"enforce_write_buffer_manager_during_recovery",
+         {offsetof(struct ImmutableDBOptions,
+                   enforce_write_buffer_manager_during_recovery),
+          OptionType::kBoolean, OptionVerificationType::kNormal,
+          OptionTypeFlags::kNone}},
         {"allow_ingest_behind",
          {offsetof(struct ImmutableDBOptions, allow_ingest_behind),
           OptionType::kBoolean, OptionVerificationType::kNormal,
@@ -719,6 +730,7 @@ ImmutableDBOptions::ImmutableDBOptions(const DBOptions& options)
       create_missing_column_families(options.create_missing_column_families),
       error_if_exists(options.error_if_exists),
       paranoid_checks(options.paranoid_checks),
+      open_files_async(options.open_files_async),
       flush_verify_memtable_count(options.flush_verify_memtable_count),
       compaction_verify_record_count(options.compaction_verify_record_count),
       track_and_verify_wals_in_manifest(
@@ -767,14 +779,14 @@ ImmutableDBOptions::ImmutableDBOptions(const DBOptions& options)
       write_thread_max_yield_usec(options.write_thread_max_yield_usec),
       write_thread_slow_yield_usec(options.write_thread_slow_yield_usec),
       skip_stats_update_on_db_open(options.skip_stats_update_on_db_open),
-      skip_checking_sst_file_sizes_on_db_open(
-          options.skip_checking_sst_file_sizes_on_db_open),
       wal_recovery_mode(options.wal_recovery_mode),
       allow_2pc(options.allow_2pc),
       row_cache(options.row_cache),
       wal_filter(options.wal_filter),
       dump_malloc_stats(options.dump_malloc_stats),
       avoid_flush_during_recovery(options.avoid_flush_during_recovery),
+      enforce_write_buffer_manager_during_recovery(
+          options.enforce_write_buffer_manager_during_recovery),
       allow_ingest_behind(options.allow_ingest_behind),
       two_write_queues(options.two_write_queues),
       manual_wal_flush(options.manual_wal_flush),
@@ -818,6 +830,8 @@ void ImmutableDBOptions::Dump(Logger* log) const {
                    create_if_missing);
   ROCKS_LOG_HEADER(log, "                        Options.paranoid_checks: %d",
                    paranoid_checks);
+  ROCKS_LOG_HEADER(log, "                       Options.open_files_async: %d",
+                   open_files_async);
   ROCKS_LOG_HEADER(log, "            Options.flush_verify_memtable_count: %d",
                    flush_verify_memtable_count);
   ROCKS_LOG_HEADER(log, "         Options.compaction_verify_record_count: %d",
@@ -939,6 +953,9 @@ void ImmutableDBOptions::Dump(Logger* log) const {
 
   ROCKS_LOG_HEADER(log, "            Options.avoid_flush_during_recovery: %d",
                    avoid_flush_during_recovery);
+  ROCKS_LOG_HEADER(
+      log, "        Options.enforce_write_buffer_manager_during_recovery: %d",
+      enforce_write_buffer_manager_during_recovery);
   ROCKS_LOG_HEADER(log, "            Options.allow_ingest_behind: %d",
                    allow_ingest_behind);
   ROCKS_LOG_HEADER(log, "            Options.two_write_queues: %d",
@@ -1045,6 +1062,8 @@ MutableDBOptions::MutableDBOptions(const DBOptions& options)
       max_manifest_file_size(options.max_manifest_file_size),
       max_manifest_space_amp_pct(options.max_manifest_space_amp_pct),
       manifest_preallocation_size(options.manifest_preallocation_size),
+      verify_manifest_content_on_close(
+          options.verify_manifest_content_on_close),
       daily_offpeak_time_utc(options.daily_offpeak_time_utc) {}
 
 void MutableDBOptions::Dump(Logger* log) const {
@@ -1098,6 +1117,8 @@ void MutableDBOptions::Dump(Logger* log) const {
   ROCKS_LOG_HEADER(
       log, "            Options.manifest_preallocation_size: %" ROCKSDB_PRIszt,
       manifest_preallocation_size);
+  ROCKS_LOG_HEADER(log, "      Options.verify_manifest_content_on_close: %d",
+                   verify_manifest_content_on_close);
   ROCKS_LOG_HEADER(log, "Options.daily_offpeak_time_utc: %s",
                    daily_offpeak_time_utc.c_str());
 }

package/deps/rocksdb/rocksdb/options/db_options.h

@@ -24,6 +24,7 @@ struct ImmutableDBOptions {
   bool create_missing_column_families;
   bool error_if_exists;
   bool paranoid_checks;
+  bool open_files_async;
   bool flush_verify_memtable_count;
   bool compaction_verify_record_count;
   bool track_and_verify_wals_in_manifest;
@@ -70,13 +71,13 @@ struct ImmutableDBOptions {
   uint64_t write_thread_max_yield_usec;
   uint64_t write_thread_slow_yield_usec;
   bool skip_stats_update_on_db_open;
-  bool skip_checking_sst_file_sizes_on_db_open;
   WALRecoveryMode wal_recovery_mode;
   bool allow_2pc;
   std::shared_ptr<Cache> row_cache;
   WalFilter* wal_filter;
   bool dump_malloc_stats;
   bool avoid_flush_during_recovery;
+  bool enforce_write_buffer_manager_during_recovery;
   bool allow_ingest_behind;
   bool two_write_queues;
   bool manual_wal_flush;
@@ -147,6 +148,7 @@ struct MutableDBOptions {
   uint64_t max_manifest_file_size;
   int max_manifest_space_amp_pct;
   size_t manifest_preallocation_size;
+  bool verify_manifest_content_on_close;
   std::string daily_offpeak_time_utc;
 };
 

package/deps/rocksdb/rocksdb/options/options.cc

@@ -115,7 +115,9 @@ AdvancedColumnFamilyOptions::AdvancedColumnFamilyOptions(const Options& options)
       persist_user_defined_timestamps(options.persist_user_defined_timestamps),
       memtable_op_scan_flush_trigger(options.memtable_op_scan_flush_trigger),
       memtable_avg_op_scan_flush_trigger(
-          options.memtable_avg_op_scan_flush_trigger) {
+          options.memtable_avg_op_scan_flush_trigger),
+      memtable_batch_lookup_optimization(
+          options.memtable_batch_lookup_optimization) {
   assert(memtable_factory.get() != nullptr);
   if (max_bytes_for_level_multiplier_additional.size() <
       static_cast<unsigned int>(num_levels)) {
@@ -468,6 +470,8 @@ void ColumnFamilyOptions::Dump(Logger* log) const {
                    memtable_max_range_deletions);
   ROCKS_LOG_HEADER(log, "                 Options.cf_allow_ingest_behind: %s",
                    cf_allow_ingest_behind ? "true" : "false");
+  ROCKS_LOG_HEADER(log, "  Options.memtable_batch_lookup_optimization: %s",
+                   memtable_batch_lookup_optimization ? "true" : "false");
 }  // ColumnFamilyOptions::Dump
 
 void Options::Dump(Logger* log) const {

package/deps/rocksdb/rocksdb/options/options_helper.cc

@@ -65,6 +65,7 @@ void BuildDBOptions(const ImmutableDBOptions& immutable_db_options,
       immutable_db_options.create_missing_column_families;
   options.error_if_exists = immutable_db_options.error_if_exists;
   options.paranoid_checks = immutable_db_options.paranoid_checks;
+  options.open_files_async = immutable_db_options.open_files_async;
   options.flush_verify_memtable_count =
       immutable_db_options.flush_verify_memtable_count;
   options.compaction_verify_record_count =
@@ -149,8 +150,6 @@ void BuildDBOptions(const ImmutableDBOptions& immutable_db_options,
       immutable_db_options.write_thread_slow_yield_usec;
   options.skip_stats_update_on_db_open =
       immutable_db_options.skip_stats_update_on_db_open;
-  options.skip_checking_sst_file_sizes_on_db_open =
-      immutable_db_options.skip_checking_sst_file_sizes_on_db_open;
   options.wal_recovery_mode = immutable_db_options.wal_recovery_mode;
   options.allow_2pc = immutable_db_options.allow_2pc;
   options.row_cache = immutable_db_options.row_cache;
@@ -158,6 +157,8 @@ void BuildDBOptions(const ImmutableDBOptions& immutable_db_options,
   options.dump_malloc_stats = immutable_db_options.dump_malloc_stats;
   options.avoid_flush_during_recovery =
       immutable_db_options.avoid_flush_during_recovery;
+  options.enforce_write_buffer_manager_during_recovery =
+      immutable_db_options.enforce_write_buffer_manager_during_recovery;
   options.avoid_flush_during_shutdown =
       mutable_db_options.avoid_flush_during_shutdown;
   options.allow_ingest_behind = immutable_db_options.allow_ingest_behind;
@@ -188,6 +189,8 @@ void BuildDBOptions(const ImmutableDBOptions& immutable_db_options,
   options.lowest_used_cache_tier = immutable_db_options.lowest_used_cache_tier;
   options.enforce_single_del_contracts =
       immutable_db_options.enforce_single_del_contracts;
+  options.verify_manifest_content_on_close =
+      mutable_db_options.verify_manifest_content_on_close;
   options.daily_offpeak_time_utc = mutable_db_options.daily_offpeak_time_utc;
   options.follower_refresh_catchup_period_ms =
       immutable_db_options.follower_refresh_catchup_period_ms;
@@ -348,6 +351,8 @@ void UpdateColumnFamilyOptions(const ImmutableCFOptions& ioptions,
       ioptions.persist_user_defined_timestamps;
   cf_opts->default_temperature = ioptions.default_temperature;
   cf_opts->cf_allow_ingest_behind = ioptions.cf_allow_ingest_behind;
+  cf_opts->memtable_batch_lookup_optimization =
+      ioptions.memtable_batch_lookup_optimization;
 
   // TODO(yhchiang): find some way to handle the following derived options
   // * max_file_size