Merge pull request ClickHouse#18481 from ClickHouse/disable_write_with_aio

alexey-milovidov · web-flow · commit 491f4817139f · 2020-12-25T04:24:43.000+03:00
Disable write with AIO even for big merges
diff --git a/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp b/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp
@@ -685,7 +685,6 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataMergerMutator::mergePartsToTempor
     if (disk->exists(new_part_tmp_path))
         throw Exception("Directory " + fullPath(disk, new_part_tmp_path) + " already exists", ErrorCodes::DIRECTORY_ALREADY_EXISTS);
 
-    MergeTreeData::DataPart::ColumnToSize merged_column_to_size;
 
     Names all_column_names = metadata_snapshot->getColumns().getNamesOfPhysical();
     NamesAndTypesList storage_columns = metadata_snapshot->getColumns().getAllPhysical();
@@ -767,6 +766,7 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataMergerMutator::mergePartsToTempor
         rows_sources_uncompressed_write_buf = tmp_disk->writeFile(rows_sources_file_path);
         rows_sources_write_buf = std::make_unique<CompressedWriteBuffer>(*rows_sources_uncompressed_write_buf);
 
+        MergeTreeData::DataPart::ColumnToSize merged_column_to_size;
         for (const MergeTreeData::DataPartPtr & part : parts)
             part->accumulateColumnSizes(merged_column_to_size);
 
@@ -921,7 +921,6 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataMergerMutator::mergePartsToTempor
         merging_columns,
         index_factory.getMany(metadata_snapshot->getSecondaryIndices()),
         compression_codec,
-        merged_column_to_size,
         data_settings->min_merge_bytes_to_use_direct_io,
         blocks_are_granules_size};
 
diff --git a/src/Storages/MergeTree/MergeTreeIOSettings.h b/src/Storages/MergeTree/MergeTreeIOSettings.h
@@ -46,7 +46,8 @@ struct MergeTreeWriterSettings
     bool rewrite_primary_key;
     bool blocks_are_granules_size;
 
-    /// true if we write temporary files during alter.
+    /// Used for AIO threshold comparsion
+    /// FIXME currently doesn't work because WriteBufferAIO contain obscure bug(s)
     size_t estimated_size = 0;
 };
 
diff --git a/src/Storages/MergeTree/MergedBlockOutputStream.cpp b/src/Storages/MergeTree/MergedBlockOutputStream.cpp
@@ -27,7 +27,6 @@ MergedBlockOutputStream::MergedBlockOutputStream(
         columns_list_,
         skip_indices,
         default_codec_,
-        {},
         data_part->storage.global_context.getSettings().min_bytes_to_use_direct_io,
         blocks_are_granules_size)
 {
@@ -39,7 +38,6 @@ MergedBlockOutputStream::MergedBlockOutputStream(
     const NamesAndTypesList & columns_list_,
     const MergeTreeIndices & skip_indices,
     CompressionCodecPtr default_codec_,
-    const MergeTreeData::DataPart::ColumnToSize & merged_column_to_size,
     size_t aio_threshold,
     bool blocks_are_granules_size)
     : IMergedBlockOutputStream(data_part, metadata_snapshot_)
@@ -54,16 +52,6 @@ MergedBlockOutputStream::MergedBlockOutputStream(
         /* rewrite_primary_key = */ true,
         blocks_are_granules_size);
 
-    if (aio_threshold > 0 && !merged_column_to_size.empty())
-    {
-        for (const auto & column : columns_list)
-        {
-            auto size_it = merged_column_to_size.find(column.name);
-            if (size_it != merged_column_to_size.end())
-                writer_settings.estimated_size += size_it->second;
-        }
-    }
-
     if (!part_path.empty())
         volume->getDisk()->createDirectories(part_path);
 
diff --git a/src/Storages/MergeTree/MergedBlockOutputStream.h b/src/Storages/MergeTree/MergedBlockOutputStream.h
@@ -27,7 +27,6 @@ class MergedBlockOutputStream final : public IMergedBlockOutputStream
         const NamesAndTypesList & columns_list_,
         const MergeTreeIndices & skip_indices,
         CompressionCodecPtr default_codec_,
-        const MergeTreeData::DataPart::ColumnToSize & merged_column_to_size,
         size_t aio_threshold,
         bool blocks_are_granules_size = false);
 
