Add multi-batches write implementation

db/column_family.cc

@@ -1515,6 +1515,13 @@ Status ColumnFamilyData::ValidateOptions(
       }
     }
   }
+
+  if (db_options.enable_multi_batch_write &&
+      cf_options.max_successive_merges > 0) {
+    return Status::NotSupported(
+        "Multi thread write is only supported with no successive merges");
+  }
+
   return s;
 }
 

db/db_impl/db_impl.h

@@ -232,6 +232,10 @@ class DBImpl : public DB {
   virtual Status Write(const WriteOptions& options,
                        WriteBatch* updates) override;
 
+  using DB::MultiBatchWrite;
+  virtual Status MultiBatchWrite(const WriteOptions& options,
+                                 std::vector<WriteBatch*>&& updates) override;
+
   using DB::Get;
   virtual Status Get(const ReadOptions& options,
                      ColumnFamilyHandle* column_family, const Slice& key,
@@ -1475,6 +1479,13 @@ class DBImpl : public DB {
                    PreReleaseCallback* pre_release_callback = nullptr,
                    PostMemTableCallback* post_memtable_callback = nullptr);
 
+  Status MultiBatchWriteImpl(const WriteOptions& write_options,
+                             std::vector<WriteBatch*>&& my_batch,
+                             WriteCallback* callback = nullptr,
+                             uint64_t* log_used = nullptr, uint64_t log_ref = 0,
+                             uint64_t* seq_used = nullptr);
+  void MultiBatchWriteCommit(CommitRequest* request);
+
   Status PipelinedWriteImpl(const WriteOptions& options, WriteBatch* updates,
                             WriteCallback* callback = nullptr,
                             uint64_t* log_used = nullptr, uint64_t log_ref = 0,
@@ -2005,7 +2016,8 @@ class DBImpl : public DB {
       mutex_.Lock();
     }
 
-    if (!immutable_db_options_.unordered_write) {
+    if (!immutable_db_options_.unordered_write &&
+        !immutable_db_options_.enable_multi_batch_write) {
       // Then the writes are finished before the next write group starts
       return;
     }

db/db_impl/db_impl_open.cc

@@ -151,6 +151,16 @@ DBOptions SanitizeOptions(const std::string& dbname, const DBOptions& src,
     result.avoid_flush_during_recovery = false;
   }
 
+  // multi thread write do not support two-write-que or write in 2PC
+  if (result.two_write_queues || result.allow_2pc) {
+    result.enable_multi_batch_write = false;
+  }
+
+  if (result.enable_multi_batch_write) {
+    result.enable_pipelined_write = false;
+    result.allow_concurrent_memtable_write = true;
+  }
+
   ImmutableDBOptions immutable_db_options(result);
   if (!immutable_db_options.IsWalDirSameAsDBPath()) {
     // Either the WAL dir and db_paths[0]/db_name are not the same, or we
@@ -1290,7 +1300,7 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& wal_numbers,
       bool has_valid_writes = false;
       status = WriteBatchInternal::InsertInto(
           batch_to_use, column_family_memtables_.get(), &flush_scheduler_,
-          &trim_history_scheduler_, true, wal_number, this,
+          &trim_history_scheduler_, true, wal_number, 0, this,
           false /* concurrent_memtable_writes */, next_sequence,
           &has_valid_writes, seq_per_batch_, batch_per_txn_);
       MaybeIgnoreError(&status);
@@ -2230,7 +2240,6 @@ Status DBImpl::Open(const DBOptions& db_options, const std::string& dbname,
                            impl->immutable_db_options_.db_paths[0].path);
   }
 
-
   if (s.ok()) {
     ROCKS_LOG_HEADER(impl->immutable_db_options_.info_log, "DB pointer %p",
                      impl);

db/db_impl/db_impl_secondary.cc

@@ -285,7 +285,7 @@ Status DBImplSecondary::RecoverLogFiles(
         status = WriteBatchInternal::InsertInto(
             &batch, column_family_memtables_.get(),
             nullptr /* flush_scheduler */, nullptr /* trim_history_scheduler*/,
-            true, log_number, this, false /* concurrent_memtable_writes */,
+            true, log_number, 0, this, false /* concurrent_memtable_writes */,
             next_sequence, &has_valid_writes, seq_per_batch_, batch_per_txn_);
       }
       // If column family was not found, it might mean that the WAL write