@nxtedition/rocksdb 10.1.5 → 10.2.0

package/deps/rocksdb/rocksdb/db/db_impl/db_impl.h

@@ -9,6 +9,7 @@
 #pragma once
 
 #include <atomic>
+#include <cstdint>
 #include <deque>
 #include <functional>
 #include <limits>
@@ -57,6 +58,7 @@
 #include "rocksdb/status.h"
 #include "rocksdb/trace_reader_writer.h"
 #include "rocksdb/transaction_log.h"
+#include "rocksdb/user_write_callback.h"
 #include "rocksdb/utilities/replayer.h"
 #include "rocksdb/write_buffer_manager.h"
 #include "table/merging_iterator.h"
@@ -231,6 +233,10 @@ class DBImpl : public DB {
   using DB::Write;
   Status Write(const WriteOptions& options, WriteBatch* updates) override;
 
+  using DB::WriteWithCallback;
+  Status WriteWithCallback(const WriteOptions& options, WriteBatch* updates,
+                           UserWriteCallback* user_write_cb) override;
+
   using DB::Get;
   Status Get(const ReadOptions& _read_options,
              ColumnFamilyHandle* column_family, const Slice& key,
@@ -501,8 +507,16 @@ class DBImpl : public DB {
   // All the returned filenames start with "/"
   Status GetLiveFiles(std::vector<std::string>&, uint64_t* manifest_file_size,
                       bool flush_memtable = true) override;
-  Status GetSortedWalFiles(VectorLogPtr& files) override;
-  Status GetCurrentWalFile(std::unique_ptr<LogFile>* current_log_file) override;
+  Status GetSortedWalFiles(VectorWalPtr& files) override;
+  Status GetSortedWalFilesImpl(VectorWalPtr& files, bool need_seqnos);
+
+  // Get the known flushed sizes of WALs that might still be written to
+  // or have pending sync.
+  // NOTE: unlike alive_log_files_, this function includes WALs that might
+  // be obsolete (but not obsolete to a pending Checkpoint) and not yet fully
+  // synced.
+  Status GetOpenWalSizes(std::map<uint64_t, uint64_t>& number_to_size);
+  Status GetCurrentWalFile(std::unique_ptr<WalFile>* current_log_file) override;
   Status GetCreationTimeOfOldestFile(uint64_t* creation_time) override;
 
   Status GetUpdatesSince(
@@ -688,7 +702,8 @@ class DBImpl : public DB {
   // thread to determine whether it is safe to perform the write.
   virtual Status WriteWithCallback(const WriteOptions& write_options,
                                    WriteBatch* my_batch,
-                                   WriteCallback* callback);
+                                   WriteCallback* callback,
+                                   UserWriteCallback* user_write_cb = nullptr);
 
   // Returns the sequence number that is guaranteed to be smaller than or equal
   // to the sequence number of any key that could be inserted into the current
@@ -1447,6 +1462,9 @@ class DBImpl : public DB {
   Status RenameTempFileToOptionsFile(const std::string& file_name);
   Status DeleteObsoleteOptionsFiles();
 
+  void NotifyOnManualFlushScheduled(autovector<ColumnFamilyData*> cfds,
+                                    FlushReason flush_reason);
+
   void NotifyOnFlushBegin(ColumnFamilyData* cfd, FileMetaData* file_meta,
                           const MutableCFOptions& mutable_cf_options,
                           int job_id, FlushReason flush_reason);
@@ -1497,6 +1515,7 @@ class DBImpl : public DB {
   // batch that does not have duplicate keys.
   Status WriteImpl(const WriteOptions& options, WriteBatch* updates,
                    WriteCallback* callback = nullptr,
+                   UserWriteCallback* user_write_cb = nullptr,
                    uint64_t* log_used = nullptr, uint64_t log_ref = 0,
                    bool disable_memtable = false, uint64_t* seq_used = nullptr,
                    size_t batch_cnt = 0,
@@ -1505,6 +1524,7 @@ class DBImpl : public DB {
 
   Status PipelinedWriteImpl(const WriteOptions& options, WriteBatch* updates,
                             WriteCallback* callback = nullptr,
+                            UserWriteCallback* user_write_cb = nullptr,
                             uint64_t* log_used = nullptr, uint64_t log_ref = 0,
                             bool disable_memtable = false,
                             uint64_t* seq_used = nullptr);
@@ -1531,7 +1551,8 @@ class DBImpl : public DB {
   // marks start of a new sub-batch.
   Status WriteImplWALOnly(
       WriteThread* write_thread, const WriteOptions& options,
-      WriteBatch* updates, WriteCallback* callback, uint64_t* log_used,
+      WriteBatch* updates, WriteCallback* callback,
+      UserWriteCallback* user_write_cb, uint64_t* log_used,
       const uint64_t log_ref, uint64_t* seq_used, const size_t sub_batch_cnt,
       PreReleaseCallback* pre_release_callback, const AssignOrder assign_order,
       const PublishLastSeq publish_last_seq, const bool disable_memtable);
@@ -1703,8 +1724,11 @@ class DBImpl : public DB {
       return w;
     }
     Status ClearWriter() {
-      // TODO: plumb Env::IOActivity, Env::IOPriority
-      Status s = writer->WriteBuffer(WriteOptions());
+      Status s;
+      if (writer->file()) {
+        // TODO: plumb Env::IOActivity, Env::IOPriority
+        s = writer->WriteBuffer(WriteOptions());
+      }
       delete writer;
       writer = nullptr;
       return s;
@@ -1719,10 +1743,16 @@ class DBImpl : public DB {
 
     void PrepareForSync() {
       assert(!getting_synced);
-      // Size is expected to be monotonically increasing.
-      assert(writer->file()->GetFlushedSize() >= pre_sync_size);
+      // Ensure the head of logs_ is marked as getting_synced if any is.
       getting_synced = true;
-      pre_sync_size = writer->file()->GetFlushedSize();
+      // If last sync failed on a later WAL, this could be a fully synced
+      // and closed WAL that just needs to be recorded as synced in the
+      // manifest.
+      if (writer->file()) {
+        // Size is expected to be monotonically increasing.
+        assert(writer->file()->GetFlushedSize() >= pre_sync_size);
+        pre_sync_size = writer->file()->GetFlushedSize();
+      }
     }
 
     void FinishSync() {
@@ -1920,7 +1950,7 @@ class DBImpl : public DB {
   void ReleaseFileNumberFromPendingOutputs(
       std::unique_ptr<std::list<uint64_t>::iterator>& v);
 
-  IOStatus SyncClosedLogs(const WriteOptions& write_options,
+  IOStatus SyncClosedWals(const WriteOptions& write_options,
                           JobContext* job_context, VersionEdit* synced_wals,
                           bool error_recovery_in_prog);
 
@@ -2179,7 +2209,8 @@ class DBImpl : public DB {
   void GenerateFlushRequest(const autovector<ColumnFamilyData*>& cfds,
                             FlushReason flush_reason, FlushRequest* req);
 
-  void SchedulePendingFlush(const FlushRequest& req);
+  // Returns true if `req` is successfully enqueued.
+  bool SchedulePendingFlush(const FlushRequest& req);
 
   void SchedulePendingCompaction(ColumnFamilyData* cfd);
   void SchedulePendingPurge(std::string fname, std::string dir_to_sync,
@@ -2255,6 +2286,11 @@ class DBImpl : public DB {
   ColumnFamilyData* PickCompactionFromQueue(
       std::unique_ptr<TaskLimiterToken>* token, LogBuffer* log_buffer);
 
+  IOStatus SyncWalImpl(bool include_current_wal,
+                       const WriteOptions& write_options,
+                       JobContext* job_context, VersionEdit* synced_wals,
+                       bool error_recovery_in_prog);
+
   // helper function to call after some of the logs_ were synced
   void MarkLogsSynced(uint64_t up_to, bool synced_dir, VersionEdit* edit);
   Status ApplyWALToManifest(const ReadOptions& read_options,
@@ -2312,6 +2348,9 @@ class DBImpl : public DB {
   bool HaveManualCompaction(ColumnFamilyData* cfd);
   bool MCOverlap(ManualCompactionState* m, ManualCompactionState* m1);
   void UpdateDeletionCompactionStats(const std::unique_ptr<Compaction>& c);
+
+  // May open and read table files for table property.
+  // Should not be called while holding mutex_.
   void BuildCompactionJobInfo(const ColumnFamilyData* cfd, Compaction* c,
                               const Status& st,
                               const CompactionJobStats& compaction_job_stats,
@@ -2527,8 +2566,10 @@ class DBImpl : public DB {
 
   bool persistent_stats_cfd_exists_ = true;
 
-  // alive_log_files_ is protected by mutex_ and log_write_mutex_ with details
-  // as follows:
+  // The current WAL file and those that have not been found obsolete from
+  // memtable flushes. A WAL not on this list might still be pending writer
+  // flush and/or sync and close and might still be in logs_. alive_log_files_
+  // is protected by mutex_ and log_write_mutex_ with details as follows:
   // 1. read by FindObsoleteFiles() which can be called in either application
   //    thread or RocksDB bg threads, both mutex_ and log_write_mutex_ are
   //    held.
@@ -2579,7 +2620,7 @@ class DBImpl : public DB {
   // 8. read by MarkLogsNotSynced() and MarkLogsSynced() are protected by
   //    log_write_mutex_.
   // 9. erase() by MarkLogsSynced() protected by log_write_mutex_.
-  // 10. read by SyncClosedLogs() protected by only log_write_mutex_. This can
+  // 10. read by SyncClosedWals() protected by only log_write_mutex_. This can
   //     happen in bg flush threads after DB::Open() returns success to
   //     applications.
   // 11. reads, e.g. front(), iteration, and back() called by PreprocessWrite()
@@ -2592,7 +2633,7 @@ class DBImpl : public DB {
   // 13. emplace_back() by SwitchMemtable() hold both mutex_ and
   //     log_write_mutex_. This happens in the write group leader. Can conflict
   //     with bg threads calling FindObsoleteFiles(), MarkLogsSynced(),
-  //     SyncClosedLogs(), etc. as well as application threads calling
+  //     SyncClosedWals(), etc. as well as application threads calling
   //     FlushWAL(), SyncWAL(), LockWAL(). This is fine because all parties
   //     require at least log_write_mutex_.
   // 14. iteration called in WriteToWAL(write_group) protected by

package/deps/rocksdb/rocksdb/db/db_impl/db_impl_compaction_flush.cc

@@ -87,6 +87,9 @@ bool DBImpl::ShouldRescheduleFlushRequestToRetainUDT(
   mutex_.AssertHeld();
   assert(flush_req.cfd_to_max_mem_id_to_persist.size() == 1);
   ColumnFamilyData* cfd = flush_req.cfd_to_max_mem_id_to_persist.begin()->first;
+  if (cfd->GetAndClearFlushSkipReschedule()) {
+    return false;
+  }
   uint64_t max_memtable_id =
       flush_req.cfd_to_max_mem_id_to_persist.begin()->second;
   if (cfd->IsDropped() ||
@@ -98,15 +101,20 @@ bool DBImpl::ShouldRescheduleFlushRequestToRetainUDT(
   // alleviated if we continue with the flush instead of postponing it.
   const auto& mutable_cf_options = *cfd->GetLatestMutableCFOptions();
 
-  // Taking the status of the active Memtable into consideration so that we are
-  // not just checking if DB is currently already in write stall mode.
-  int mem_to_flush = cfd->mem()->ApproximateMemoryUsageFast() >=
-                             cfd->mem()->write_buffer_size() / 2
-                         ? 1
-                         : 0;
+  // Use the same criteria as WaitUntilFlushWouldNotStallWrites does w.r.t
+  // defining what a write stall is about to happen means. If this uses a
+  // stricter criteria, for example, a write stall is about to happen if the
+  // last memtable is 10% full, there is a possibility that manual flush could
+  // be waiting in `WaitUntilFlushWouldNotStallWrites` with the incorrect
+  // expectation that others will clear up the excessive memtables and
+  // eventually let it proceed. The others in this case won't start clearing
+  // until the last memtable is 10% full. To avoid that scenario, the criteria
+  // this uses should be the same or less strict than
+  // `WaitUntilFlushWouldNotStallWrites` does.
   WriteStallCondition write_stall =
       ColumnFamilyData::GetWriteStallConditionAndCause(
-          cfd->imm()->NumNotFlushed() + mem_to_flush, /*num_l0_files=*/0,
+          cfd->GetUnflushedMemTableCountForWriteStallCheck(),
+          /*num_l0_files=*/0,
           /*num_compaction_needed_bytes=*/0, mutable_cf_options,
           *cfd->ioptions())
           .first;
@@ -116,89 +124,19 @@ bool DBImpl::ShouldRescheduleFlushRequestToRetainUDT(
   return true;
 }
 
-IOStatus DBImpl::SyncClosedLogs(const WriteOptions& write_options,
+IOStatus DBImpl::SyncClosedWals(const WriteOptions& write_options,
                                 JobContext* job_context,
                                 VersionEdit* synced_wals,
                                 bool error_recovery_in_prog) {
-  TEST_SYNC_POINT("DBImpl::SyncClosedLogs:Start");
-  InstrumentedMutexLock l(&log_write_mutex_);
-  autovector<log::Writer*, 1> logs_to_sync;
-  uint64_t current_log_number = logfile_number_;
-  while (logs_.front().number < current_log_number &&
-         logs_.front().IsSyncing()) {
-    log_sync_cv_.Wait();
-  }
-  for (auto it = logs_.begin();
-       it != logs_.end() && it->number < current_log_number; ++it) {
-    auto& log = *it;
-    log.PrepareForSync();
-    logs_to_sync.push_back(log.writer);
-  }
-
-  IOStatus io_s;
-  if (!logs_to_sync.empty()) {
-    log_write_mutex_.Unlock();
-
-    assert(job_context);
-
-    for (log::Writer* log : logs_to_sync) {
-      ROCKS_LOG_INFO(immutable_db_options_.info_log,
-                     "[JOB %d] Syncing log #%" PRIu64, job_context->job_id,
-                     log->get_log_number());
-      if (error_recovery_in_prog) {
-        log->file()->reset_seen_error();
-      }
-
-      IOOptions io_options;
-      io_s = WritableFileWriter::PrepareIOOptions(write_options, io_options);
-      if (!io_s.ok()) {
-        break;
-      }
-      io_s = log->file()->Sync(io_options, immutable_db_options_.use_fsync);
-      if (!io_s.ok()) {
-        break;
-      }
-
-      if (immutable_db_options_.recycle_log_file_num > 0) {
-        if (error_recovery_in_prog) {
-          log->file()->reset_seen_error();
-        }
-        // Normally the log file is closed when purging obsolete file, but if
-        // log recycling is enabled, the log file is closed here so that it
-        // can be reused.
-        io_s = log->Close(write_options);
-        if (!io_s.ok()) {
-          break;
-        }
-      }
-    }
-    if (io_s.ok()) {
-      IOOptions io_options;
-      io_s = WritableFileWriter::PrepareIOOptions(write_options, io_options);
-      if (io_s.ok()) {
-        io_s = directories_.GetWalDir()->FsyncWithDirOptions(
-            io_options, nullptr,
-            DirFsyncOptions(DirFsyncOptions::FsyncReason::kNewFileSynced));
-      }
-    }
-
-    TEST_SYNC_POINT_CALLBACK("DBImpl::SyncClosedLogs:BeforeReLock",
-                             /*arg=*/nullptr);
-    log_write_mutex_.Lock();
+  TEST_SYNC_POINT("DBImpl::SyncClosedWals:Start");
 
-    // "number <= current_log_number - 1" is equivalent to
-    // "number < current_log_number".
-    if (io_s.ok()) {
-      MarkLogsSynced(current_log_number - 1, true, synced_wals);
-    } else {
-      MarkLogsNotSynced(current_log_number - 1);
-    }
-    if (!io_s.ok()) {
-      TEST_SYNC_POINT("DBImpl::SyncClosedLogs:Failed");
-      return io_s;
-    }
+  IOStatus io_s = SyncWalImpl(/*include_current_wal*/ false, write_options,
+                              job_context, synced_wals, error_recovery_in_prog);
+  if (!io_s.ok()) {
+    TEST_SYNC_POINT("DBImpl::SyncClosedWals:Failed");
+  } else {
+    TEST_SYNC_POINT("DBImpl::SyncClosedWals:end");
   }
-  TEST_SYNC_POINT("DBImpl::SyncClosedLogs:end");
   return io_s;
 }
 
@@ -237,12 +175,12 @@ Status DBImpl::FlushMemTableToOutputFile(
   // If needs_to_sync_closed_wals is true, we need to record the current
   // maximum memtable ID of this column family so that a later PickMemtables()
   // call will not pick memtables whose IDs are higher. This is due to the fact
-  // that SyncClosedLogs() may release the db mutex, and memtable switch can
+  // that SyncClosedWals() may release the db mutex, and memtable switch can
   // happen for this column family in the meantime. The newly created memtables
   // have their data backed by unsynced WALs, thus they cannot be included in
   // this flush job.
   // Another reason why we must record the current maximum memtable ID of this
-  // column family: SyncClosedLogs() may release db mutex, thus it's possible
+  // column family: SyncClosedWals() may release db mutex, thus it's possible
   // for application to continue to insert into memtables increasing db's
   // sequence number. The application may take a snapshot, but this snapshot is
   // not included in `snapshot_seqs` which will be passed to flush job because
@@ -256,7 +194,7 @@ Status DBImpl::FlushMemTableToOutputFile(
 
   // If needs_to_sync_closed_wals is false, then the flush job will pick ALL
   // existing memtables of the column family when PickMemTable() is called
-  // later. Although we won't call SyncClosedLogs() in this case, we may still
+  // later. Although we won't call SyncClosedWals() in this case, we may still
   // call the callbacks of the listeners, i.e. NotifyOnFlushBegin() which also
   // releases and re-acquires the db mutex. In the meantime, the application
   // can still insert into the memtables and increase the db's sequence number.
@@ -286,12 +224,12 @@ Status DBImpl::FlushMemTableToOutputFile(
   bool need_cancel = false;
   IOStatus log_io_s = IOStatus::OK();
   if (needs_to_sync_closed_wals) {
-    // SyncClosedLogs() may unlock and re-lock the log_write_mutex multiple
+    // SyncClosedWals() may unlock and re-lock the log_write_mutex multiple
     // times.
     VersionEdit synced_wals;
     bool error_recovery_in_prog = error_handler_.IsRecoveryInProgress();
     mutex_.Unlock();
-    log_io_s = SyncClosedLogs(write_options, job_context, &synced_wals,
+    log_io_s = SyncClosedWals(write_options, job_context, &synced_wals,
                               error_recovery_in_prog);
     mutex_.Lock();
     if (log_io_s.ok() && synced_wals.IsWalAddition()) {
@@ -306,7 +244,7 @@ Status DBImpl::FlushMemTableToOutputFile(
       error_handler_.SetBGError(log_io_s, BackgroundErrorReason::kFlush);
     }
   } else {
-    TEST_SYNC_POINT("DBImpl::SyncClosedLogs:Skip");
+    TEST_SYNC_POINT("DBImpl::SyncClosedWals:Skip");
   }
   s = log_io_s;
 
@@ -580,7 +518,7 @@ Status DBImpl::AtomicFlushMemTablesToOutputFiles(
     VersionEdit synced_wals;
     bool error_recovery_in_prog = error_handler_.IsRecoveryInProgress();
     mutex_.Unlock();
-    log_io_s = SyncClosedLogs(write_options, job_context, &synced_wals,
+    log_io_s = SyncClosedWals(write_options, job_context, &synced_wals,
                               error_recovery_in_prog);
     mutex_.Lock();
     if (log_io_s.ok() && synced_wals.IsWalAddition()) {
@@ -1657,10 +1595,12 @@ Status DBImpl::CompactFilesImpl(
 
   ReleaseFileNumberFromPendingOutputs(pending_outputs_inserted_elem);
 
+  mutex_.Unlock();
   if (compaction_job_info != nullptr) {
     BuildCompactionJobInfo(cfd, c.get(), s, compaction_job_stats,
                            job_context->job_id, compaction_job_info);
   }
+  mutex_.Lock();
 
   if (status.ok()) {
     // Done
@@ -2310,6 +2250,23 @@ void DBImpl::GenerateFlushRequest(const autovector<ColumnFamilyData*>& cfds,
   }
 }
 
+void DBImpl::NotifyOnManualFlushScheduled(autovector<ColumnFamilyData*> cfds,
+                                          FlushReason flush_reason) {
+  if (immutable_db_options_.listeners.size() == 0U) {
+    return;
+  }
+  if (shutting_down_.load(std::memory_order_acquire)) {
+    return;
+  }
+  std::vector<ManualFlushInfo> info;
+  for (ColumnFamilyData* cfd : cfds) {
+    info.push_back({cfd->GetID(), cfd->GetName(), flush_reason});
+  }
+  for (const auto& listener : immutable_db_options_.listeners) {
+    listener->OnManualFlushScheduled(this, info);
+  }
+}
+
 Status DBImpl::FlushMemTable(ColumnFamilyData* cfd,
                              const FlushOptions& flush_options,
                              FlushReason flush_reason,
@@ -2414,7 +2371,14 @@ Status DBImpl::FlushMemTable(ColumnFamilyData* cfd,
         }
       }
       for (const auto& req : flush_reqs) {
-        SchedulePendingFlush(req);
+        assert(req.cfd_to_max_mem_id_to_persist.size() == 1);
+        ColumnFamilyData* loop_cfd =
+            req.cfd_to_max_mem_id_to_persist.begin()->first;
+        bool already_queued_for_flush = loop_cfd->queued_for_flush();
+        bool flush_req_enqueued = SchedulePendingFlush(req);
+        if (already_queued_for_flush || flush_req_enqueued) {
+          loop_cfd->SetFlushSkipReschedule();
+        }
       }
       MaybeScheduleFlushOrCompaction();
     }
@@ -2426,6 +2390,8 @@ Status DBImpl::FlushMemTable(ColumnFamilyData* cfd,
       }
     }
   }
+
+  NotifyOnManualFlushScheduled({cfd}, flush_reason);
   TEST_SYNC_POINT("DBImpl::FlushMemTable:AfterScheduleFlush");
   TEST_SYNC_POINT("DBImpl::FlushMemTable:BeforeWaitForBgFlush");
   if (s.ok() && flush_options.wait) {
@@ -2570,6 +2536,7 @@ Status DBImpl::AtomicFlushMemTables(
       }
     }
   }
+  NotifyOnManualFlushScheduled(cfds, flush_reason);
   TEST_SYNC_POINT("DBImpl::AtomicFlushMemTables:AfterScheduleFlush");
   TEST_SYNC_POINT("DBImpl::AtomicFlushMemTables:BeforeWaitForBgFlush");
   if (s.ok() && flush_options.wait) {
@@ -2627,7 +2594,9 @@ Status DBImpl::RetryFlushesForErrorRecovery(FlushReason flush_reason,
           flush_reason,
           {{cfd,
             std::numeric_limits<uint64_t>::max() /* max_mem_id_to_persist */}}};
-      SchedulePendingFlush(flush_req);
+      if (SchedulePendingFlush(flush_req)) {
+        cfd->SetFlushSkipReschedule();
+      };
     }
   }
   MaybeScheduleFlushOrCompaction();
@@ -2715,13 +2684,13 @@ Status DBImpl::WaitUntilFlushWouldNotStallWrites(ColumnFamilyData* cfd,
       // mode due to pending compaction bytes, but that's less common
       // No extra immutable Memtable will be created if the current Memtable is
       // empty.
-      int mem_to_flush = cfd->mem()->IsEmpty() ? 0 : 1;
-      write_stall_condition = ColumnFamilyData::GetWriteStallConditionAndCause(
-                                  cfd->imm()->NumNotFlushed() + mem_to_flush,
-                                  vstorage->l0_delay_trigger_count() + 1,
-                                  vstorage->estimated_compaction_needed_bytes(),
-                                  mutable_cf_options, *cfd->ioptions())
-                                  .first;
+      write_stall_condition =
+          ColumnFamilyData::GetWriteStallConditionAndCause(
+              cfd->GetUnflushedMemTableCountForWriteStallCheck(),
+              vstorage->l0_delay_trigger_count() + 1,
+              vstorage->estimated_compaction_needed_bytes(), mutable_cf_options,
+              *cfd->ioptions())
+              .first;
     } while (write_stall_condition != WriteStallCondition::kNormal);
   }
   return Status::OK();
@@ -3033,13 +3002,14 @@ ColumnFamilyData* DBImpl::PickCompactionFromQueue(
   return cfd;
 }
 
-void DBImpl::SchedulePendingFlush(const FlushRequest& flush_req) {
+bool DBImpl::SchedulePendingFlush(const FlushRequest& flush_req) {
   mutex_.AssertHeld();
+  bool enqueued = false;
   if (reject_new_background_jobs_) {
-    return;
+    return enqueued;
   }
   if (flush_req.cfd_to_max_mem_id_to_persist.empty()) {
-    return;
+    return enqueued;
   }
   if (!immutable_db_options_.atomic_flush) {
     // For the non-atomic flush case, we never schedule multiple column
@@ -3054,6 +3024,7 @@ void DBImpl::SchedulePendingFlush(const FlushRequest& flush_req) {
       cfd->set_queued_for_flush(true);
       ++unscheduled_flushes_;
       flush_queue_.push_back(flush_req);
+      enqueued = true;
     }
   } else {
     for (auto& iter : flush_req.cfd_to_max_mem_id_to_persist) {
@@ -3062,7 +3033,9 @@ void DBImpl::SchedulePendingFlush(const FlushRequest& flush_req) {
     }
     ++unscheduled_flushes_;
     flush_queue_.push_back(flush_req);
+    enqueued = true;
   }
+  return enqueued;
 }
 
 void DBImpl::SchedulePendingCompaction(ColumnFamilyData* cfd) {
@@ -4190,7 +4163,7 @@ void DBImpl::BuildCompactionJobInfo(
   compaction_job_info->base_input_level = c->start_level();
   compaction_job_info->output_level = c->output_level();
   compaction_job_info->stats = compaction_job_stats;
-  const auto& input_table_properties = c->GetInputTableProperties();
+  const auto& input_table_properties = c->GetOrInitInputTableProperties();
   const auto& output_table_properties = c->GetOutputTableProperties();
   compaction_job_info->table_properties.insert(input_table_properties.begin(),
                                                input_table_properties.end());

package/deps/rocksdb/rocksdb/db/db_impl/db_impl_files.cc

@@ -312,6 +312,26 @@ void DBImpl::FindObsoleteFiles(JobContext* job_context, bool force,
         // logs_ could have changed while we were waiting.
         continue;
       }
+      // This WAL file is not live, so it's OK if we never sync the rest of it.
+      // If it's already closed, then it's been fully synced. If
+      // !background_close_inactive_wals then we need to Close it before
+      // removing from logs_ but not blocking while holding log_write_mutex_.
+      if (!immutable_db_options_.background_close_inactive_wals &&
+          log.writer->file()) {
+        // We are taking ownership of and pinning the front entry, so we can
+        // expect it to be the same after releasing and re-acquiring the lock
+        log.PrepareForSync();
+        log_write_mutex_.Unlock();
+        // TODO: maybe check the return value of Close.
+        // TODO: plumb Env::IOActivity, Env::IOPriority
+        auto s = log.writer->file()->Close({});
+        s.PermitUncheckedError();
+        log_write_mutex_.Lock();
+        log.writer->PublishIfClosed();
+        assert(&log == &logs_.front());
+        log.FinishSync();
+        log_sync_cv_.SignalAll();
+      }
       logs_to_free_.push_back(log.ReleaseWriter());
       logs_.pop_front();
     }
@@ -410,12 +430,24 @@ void DBImpl::PurgeObsoleteFiles(JobContext& state, bool schedule_only) {
       state.manifest_delete_files.size());
   // We may ignore the dbname when generating the file names.
   for (auto& file : state.sst_delete_files) {
-    if (!file.only_delete_metadata) {
-      candidate_files.emplace_back(
-          MakeTableFileName(file.metadata->fd.GetNumber()), file.path);
-    }
-    if (file.metadata->table_reader_handle) {
-      table_cache_->Release(file.metadata->table_reader_handle);
+    auto* handle = file.metadata->table_reader_handle;
+    if (file.only_delete_metadata) {
+      if (handle) {
+        // Simply release handle of file that is not being deleted
+        table_cache_->Release(handle);
+      }
+    } else {
+      // File is being deleted (actually obsolete)
+      auto number = file.metadata->fd.GetNumber();
+      candidate_files.emplace_back(MakeTableFileName(number), file.path);
+      if (handle == nullptr) {
+        // For files not "pinned" in table cache
+        handle = TableCache::Lookup(table_cache_.get(), number);
+      }
+      if (handle) {
+        TableCache::ReleaseObsolete(table_cache_.get(), handle,
+                                    file.uncache_aggressiveness);
+      }
     }
     file.DeleteMetadata();
   }
@@ -491,7 +523,7 @@ void DBImpl::PurgeObsoleteFiles(JobContext& state, bool schedule_only) {
   for (const auto w : state.logs_to_free) {
     // TODO: maybe check the return value of Close.
     // TODO: plumb Env::IOActivity, Env::IOPriority
-    auto s = w->Close(WriteOptions());
+    auto s = w->Close({});
     s.PermitUncheckedError();
   }
 
@@ -577,8 +609,6 @@ void DBImpl::PurgeObsoleteFiles(JobContext& state, bool schedule_only) {
     std::string fname;
     std::string dir_to_sync;
     if (type == kTableFile) {
-      // evict from cache
-      TableCache::Evict(table_cache_.get(), number);
       fname = MakeTableFileName(candidate_file.file_path, number);
       dir_to_sync = candidate_file.file_path;
     } else if (type == kBlobFile) {

package/deps/rocksdb/rocksdb/db/db_impl/db_impl_follower.cc

@@ -5,6 +5,7 @@
 
 #include "db/db_impl/db_impl_follower.h"
 
+#include <algorithm>
 #include <cinttypes>
 
 #include "db/arena_wrapped_db_iter.h"