diff --git a/analytic_engine/src/compaction/picker.rs b/analytic_engine/src/compaction/picker.rs
index e309772967..7669d02204 100644
--- a/analytic_engine/src/compaction/picker.rs
+++ b/analytic_engine/src/compaction/picker.rs
@@ -617,15 +617,13 @@ mod tests {
         },
     };
 
-    fn build_sst_meta_data(time_range: TimeRange, size: u64) -> SstMetaData {
+    fn build_sst_meta_data(time_range: TimeRange, _size: u64) -> SstMetaData {
         SstMetaData {
             min_key: Bytes::from_static(b"100"),
             max_key: Bytes::from_static(b"200"),
             time_range,
             max_sequence: 200,
             schema: build_schema(),
-            size,
-            row_num: 2,
             storage_format_opts: Default::default(),
             bloom_filter: Default::default(),
         }
@@ -769,6 +767,8 @@ mod tests {
             .map(|size| {
                 let file_meta = FileMeta {
                     id: 1,
+                    size,
+                    row_num: 0,
                     meta: build_sst_meta_data(TimeRange::empty(), size),
                 };
                 let queue = FilePurgeQueue::new(1, 1.into(), tx.clone());
diff --git a/analytic_engine/src/instance/flush_compaction.rs b/analytic_engine/src/instance/flush_compaction.rs
index 5a9b8df21b..1a5a04f604 100644
--- a/analytic_engine/src/instance/flush_compaction.rs
+++ b/analytic_engine/src/instance/flush_compaction.rs
@@ -502,7 +502,7 @@ impl Instance {
                 flushed_sequence = seq;
                 sst_num += files_to_level0.len();
                 for add_file in &files_to_level0 {
-                    local_metrics.observe_sst_size(add_file.file.meta.size);
+                    local_metrics.observe_sst_size(add_file.file.size);
                 }
             }
         }
@@ -511,7 +511,7 @@ impl Instance {
                 .dump_normal_memtable(table_data, request_id, mem)
                 .await?;
             if let Some(file) = file {
-                let sst_size = file.meta.size;
+                let sst_size = file.size;
                 files_to_level0.push(AddFile { level: 0, file });
 
                 // Set flushed sequence to max of the last_sequence of memtables.
@@ -619,14 +619,12 @@ impl Instance {
             let sst_file_path = table_data.set_sst_file_path(file_id);
 
             // TODO: min_key max_key set in sst_builder build
-            let mut sst_meta = SstMetaData {
+            let sst_meta = SstMetaData {
                 min_key: min_key.clone(),
                 max_key: max_key.clone(),
                 time_range: *time_range,
                 max_sequence,
                 schema: table_data.schema(),
-                size: 0,
-                row_num: 0,
                 storage_format_opts: StorageFormatOptions::new(
                     table_data.table_options().storage_format,
                 ),
@@ -663,10 +661,7 @@ impl Instance {
                         path: sst_file_path.to_string(),
                     })?;
 
-                // update sst metadata by built info.
-                sst_meta.row_num = sst_info.row_num as u64;
-                sst_meta.size = sst_info.file_size as u64;
-                Ok(sst_meta)
+                Ok((sst_info, sst_meta))
             });
 
             batch_record_senders.push(batch_record_sender);
@@ -698,12 +693,15 @@ impl Instance {
         batch_record_senders.clear();
 
         let ret = try_join_all(sst_handlers).await;
-        for (idx, sst_meta) in ret.context(RuntimeJoin)?.into_iter().enumerate() {
+        for (idx, info_and_meta) in ret.context(RuntimeJoin)?.into_iter().enumerate() {
+            let (sst_info, sst_meta) = info_and_meta?;
             files_to_level0.push(AddFile {
                 level: 0,
                 file: FileMeta {
                     id: file_ids[idx],
-                    meta: sst_meta?,
+                    size: sst_info.file_size as u64,
+                    row_num: sst_info.row_num as u64,
+                    meta: sst_meta,
                 },
             })
         }
@@ -727,14 +725,12 @@ impl Instance {
             }
         };
         let max_sequence = memtable_state.last_sequence();
-        let mut sst_meta = SstMetaData {
+        let sst_meta = SstMetaData {
             min_key,
             max_key,
             time_range: memtable_state.time_range,
             max_sequence,
             schema: table_data.schema(),
-            size: 0,
-            row_num: 0,
             storage_format_opts: StorageFormatOptions::new(table_data.storage_format()),
             bloom_filter: Default::default(),
         };
@@ -774,11 +770,11 @@ impl Instance {
             })?;
 
         // update sst metadata by built info.
-        sst_meta.row_num = sst_info.row_num as u64;
-        sst_meta.size = sst_info.file_size as u64;
 
         Ok(Some(FileMeta {
             id: file_id,
+            row_num: sst_info.row_num as u64,
+            size: sst_info.file_size as u64,
             meta: sst_meta,
         }))
     }
@@ -953,7 +949,7 @@ impl SpaceStore {
             row_iter::record_batch_with_key_iter_to_stream(merge_iter, &runtime)
         };
 
-        let mut sst_meta = file::merge_sst_meta(&input.files, schema);
+        let sst_meta = file::merge_sst_meta(&input.files, schema);
 
         // Alloc file id for the merged sst.
         let file_id = table_data.alloc_file_id();
@@ -979,16 +975,12 @@ impl SpaceStore {
                 path: sst_file_path.to_string(),
             })?;
 
-        // update sst metadata by built info.
-        sst_meta.row_num = sst_info.row_num as u64;
-        sst_meta.size = sst_info.file_size as u64;
-
         table_data
             .metrics
-            .compaction_observe_output_sst_size(sst_meta.size);
+            .compaction_observe_output_sst_size(sst_info.file_size as u64);
         table_data
             .metrics
-            .compaction_observe_output_sst_row_num(sst_meta.row_num);
+            .compaction_observe_output_sst_row_num(sst_info.row_num as u64);
 
         info!(
             "Instance files compacted, table:{}, table_id:{}, request_id:{}, output_path:{}, input_files:{:?}, sst_meta:{:?}",
@@ -1014,6 +1006,8 @@ impl SpaceStore {
             level: input.output_level,
             file: FileMeta {
                 id: file_id,
+                size: sst_info.file_size as u64,
+                row_num: sst_info.row_num as u64,
                 meta: sst_meta,
             },
         });
diff --git a/analytic_engine/src/sst/file.rs b/analytic_engine/src/sst/file.rs
index 37b92e0289..09cac771c3 100644
--- a/analytic_engine/src/sst/file.rs
+++ b/analytic_engine/src/sst/file.rs
@@ -185,7 +185,7 @@ impl FileHandle {
 
     #[inline]
     pub fn row_num(&self) -> u64 {
-        self.inner.meta.meta.row_num
+        self.inner.meta.row_num
     }
 
     #[inline]
@@ -235,7 +235,7 @@ impl FileHandle {
 
     #[inline]
     pub fn size(&self) -> u64 {
-        self.inner.meta.meta.size
+        self.inner.meta.size
     }
 
     #[inline]
@@ -424,6 +424,10 @@ impl FileHandleSet {
 pub struct FileMeta {
     /// Id of the sst file
     pub id: FileId,
+    /// file size in bytes
+    pub size: u64,
+    // total row number
+    pub row_num: u64,
     pub meta: SstMetaData,
 }
 
@@ -508,10 +512,6 @@ pub struct SstMetaData {
     /// Max sequence number in the sst
     pub max_sequence: SequenceNumber,
     pub schema: Schema,
-    /// file size in bytes
-    pub size: u64,
-    // total row number
-    pub row_num: u64,
     pub storage_format_opts: StorageFormatOptions,
     pub bloom_filter: Option<BloomFilter>,
 }
@@ -532,8 +532,6 @@ impl From<SstMetaData> for sst_pb::SstMetaData {
             max_sequence: src.max_sequence,
             time_range: Some(src.time_range.into()),
             schema: Some(common_pb::TableSchema::from(&src.schema)),
-            size: src.size,
-            row_num: src.row_num,
             storage_format_opts: Some(src.storage_format_opts.into()),
             bloom_filter: src.bloom_filter.map(|v| v.into()),
         }
@@ -564,8 +562,6 @@ impl TryFrom<sst_pb::SstMetaData> for SstMetaData {
             time_range,
             max_sequence: src.max_sequence,
             schema,
-            size: src.size,
-            row_num: src.row_num,
             storage_format_opts,
             bloom_filter,
         })
@@ -726,9 +722,6 @@ pub fn merge_sst_meta(files: &[FileHandle], schema: Schema) -> SstMetaData {
     let mut time_range_start = files[0].time_range().inclusive_start();
     let mut time_range_end = files[0].time_range().exclusive_end();
     let mut max_sequence = files[0].max_sequence();
-    // TODO(jiacai2050): what if format of different file is different?
-    // pick first now
-    let storage_format = files[0].storage_format();
 
     if files.len() > 1 {
         for file in &files[1..] {
@@ -746,11 +739,8 @@ pub fn merge_sst_meta(files: &[FileHandle], schema: Schema) -> SstMetaData {
         time_range: TimeRange::new(time_range_start, time_range_end).unwrap(),
         max_sequence,
         schema,
-        // we don't know file size and total row number yet
-        size: 0,
-        row_num: 0,
-        storage_format_opts: StorageFormatOptions::new(storage_format),
-        // bloom filter is rebuilt when write sst, so use default here
+        // we don't know those info yet
+        storage_format_opts: Default::default(),
         bloom_filter: Default::default(),
     }
 }
@@ -805,8 +795,6 @@ pub mod tests {
                 time_range: self.time_range,
                 max_sequence: self.max_sequence,
                 schema: self.schema.clone(),
-                row_num: 0,
-                size: 0,
                 storage_format_opts: Default::default(),
                 bloom_filter: Default::default(),
             }
diff --git a/analytic_engine/src/sst/manager.rs b/analytic_engine/src/sst/manager.rs
index 2d64a8fafb..59f8945e94 100644
--- a/analytic_engine/src/sst/manager.rs
+++ b/analytic_engine/src/sst/manager.rs
@@ -149,6 +149,8 @@ pub mod tests {
                     0,
                     FileMeta {
                         id: id as FileId,
+                        size: 0,
+                        row_num: 0,
                         meta: sst_meta,
                     },
                 );
diff --git a/analytic_engine/src/sst/meta_cache.rs b/analytic_engine/src/sst/meta_cache.rs
index f81fdf339c..a347dd08b3 100644
--- a/analytic_engine/src/sst/meta_cache.rs
+++ b/analytic_engine/src/sst/meta_cache.rs
@@ -48,11 +48,7 @@ impl MetaData {
     ///
     /// After the building, a new parquet meta data will be generated which
     /// contains no extended custom information.
-    pub fn try_new(
-        parquet_meta_data: &ParquetMetaData,
-        sst_size: usize,
-        ignore_bloom_filter: bool,
-    ) -> Result<Self> {
+    pub fn try_new(parquet_meta_data: &ParquetMetaData, ignore_bloom_filter: bool) -> Result<Self> {
         let file_meta_data = parquet_meta_data.file_metadata();
         let kv_metas = file_meta_data
             .key_value_metadata()
@@ -66,11 +62,6 @@ impl MetaData {
                 sst_meta.bloom_filter = None;
             }
 
-            // FIXME: After the issue fixed, let's remove the `sst_size` parameter.
-            // The size in sst_meta is always 0, so overwrite it here.
-            // Refer to https://github.com/CeresDB/ceresdb/issues/321
-            sst_meta.size = sst_size as u64;
-
             Arc::new(sst_meta)
         };
 
diff --git a/analytic_engine/src/sst/parquet/async_reader.rs b/analytic_engine/src/sst/parquet/async_reader.rs
index c552c65060..6148ca4f63 100644
--- a/analytic_engine/src/sst/parquet/async_reader.rs
+++ b/analytic_engine/src/sst/parquet/async_reader.rs
@@ -335,8 +335,7 @@ impl<'a> Reader<'a> {
             let parquet_meta_data = self.load_meta_data_from_storage().await?;
 
             let ignore_bloom_filter = avoid_update_cache && empty_predicate;
-            let file_size = self.file_reader.file_size().await.context(DecodeSstMeta)?;
-            MetaData::try_new(&parquet_meta_data, file_size, ignore_bloom_filter)
+            MetaData::try_new(&parquet_meta_data, ignore_bloom_filter)
                 .map_err(|e| Box::new(e) as _)
                 .context(DecodeSstMeta)?
         };
diff --git a/analytic_engine/src/sst/parquet/builder.rs b/analytic_engine/src/sst/parquet/builder.rs
index eda9008d9d..73039da19f 100644
--- a/analytic_engine/src/sst/parquet/builder.rs
+++ b/analytic_engine/src/sst/parquet/builder.rs
@@ -311,8 +311,6 @@ mod tests {
                 time_range: TimeRange::new_unchecked(Timestamp::new(1), Timestamp::new(2)),
                 max_sequence: 200,
                 schema: schema.clone(),
-                size: 10,
-                row_num: 2,
                 storage_format_opts: Default::default(),
                 bloom_filter: Default::default(),
             };
@@ -368,13 +366,7 @@ mod tests {
                     Arc::new(file_reader),
                     &sst_reader_options,
                 );
-                let mut sst_meta_readback = {
-                    // FIXME: size of SstMetaData is not what this file's size, so overwrite it
-                    // https://github.com/CeresDB/ceresdb/issues/321
-                    let mut meta = reader.meta_data().await.unwrap().clone();
-                    meta.size = sst_meta.size;
-                    meta
-                };
+                let mut sst_meta_readback = { reader.meta_data().await.unwrap().clone() };
                 // bloom filter is built insider sst writer, so overwrite to default for
                 // comparsion
                 sst_meta_readback.bloom_filter = Default::default();
@@ -461,8 +453,6 @@ mod tests {
                 time_range: Default::default(),
                 max_sequence: 1,
                 schema,
-                size: 0,
-                row_num: 0,
                 storage_format_opts: Default::default(),
                 bloom_filter: Default::default(),
             },
diff --git a/analytic_engine/src/sst/parquet/encoding.rs b/analytic_engine/src/sst/parquet/encoding.rs
index 1db6f7c5b5..c836b1f9bf 100644
--- a/analytic_engine/src/sst/parquet/encoding.rs
+++ b/analytic_engine/src/sst/parquet/encoding.rs
@@ -871,8 +871,6 @@ mod tests {
             time_range: TimeRange::new_unchecked(Timestamp::new(100), Timestamp::new(101)),
             max_sequence: 200,
             schema: schema.clone(),
-            size: 10,
-            row_num: 4,
             storage_format_opts,
             bloom_filter: Default::default(),
         };
@@ -1006,8 +1004,6 @@ mod tests {
             time_range: TimeRange::new_unchecked(Timestamp::new(100), Timestamp::new(101)),
             max_sequence: 200,
             schema: schema.clone(),
-            size: 10,
-            row_num: 4,
             storage_format_opts,
             bloom_filter: Default::default(),
         };
diff --git a/analytic_engine/src/table/version_edit.rs b/analytic_engine/src/table/version_edit.rs
index c0a10071aa..e5ee836f25 100644
--- a/analytic_engine/src/table/version_edit.rs
+++ b/analytic_engine/src/table/version_edit.rs
@@ -62,8 +62,8 @@ impl From<AddFile> for meta_pb::AddFileMeta {
             time_range: Some(v.file.meta.time_range.into()),
             max_seq: v.file.meta.max_sequence,
             schema: Some(common_pb::TableSchema::from(&v.file.meta.schema)),
-            size: v.file.meta.size,
-            row_num: v.file.meta.row_num,
+            size: v.file.size,
+            row_num: v.file.row_num,
             storage_format: analytic_common_pb::StorageFormat::from(v.file.meta.storage_format())
                 as i32,
         }
@@ -91,14 +91,14 @@ impl TryFrom<meta_pb::AddFileMeta> for AddFile {
                 .context(InvalidLevel { level: src.level })?,
             file: FileMeta {
                 id: src.file_id,
+                size: src.size,
+                row_num: src.row_num,
                 meta: SstMetaData {
                     min_key: Bytes::from(src.min_key),
                     max_key: Bytes::from(src.max_key),
                     time_range,
                     max_sequence: src.max_seq,
                     schema,
-                    size: src.size,
-                    row_num: src.row_num,
                     storage_format_opts: StorageFormatOptions::new(storage_format.into()),
                     bloom_filter: Default::default(),
                 },
@@ -186,6 +186,8 @@ pub mod tests {
                 level: 0,
                 file: FileMeta {
                     id: self.file_id,
+                    size: 0,
+                    row_num: 0,
                     meta: self.sst_meta.clone(),
                 },
             }
diff --git a/benchmarks/src/util.rs b/benchmarks/src/util.rs
index d2071a66c6..9c9256cd81 100644
--- a/benchmarks/src/util.rs
+++ b/benchmarks/src/util.rs
@@ -150,6 +150,8 @@ pub async fn file_handles_from_ssts(
         let sst_meta = meta_from_sst(store, &path, meta_cache).await;
         let file_meta = FileMeta {
             id: *file_id,
+            size: 0,
+            row_num: 0,
             meta: sst_meta,
         };
 
diff --git a/proto/protos/sst.proto b/proto/protos/sst.proto
index 32b136e447..6fb5305334 100644
--- a/proto/protos/sst.proto
+++ b/proto/protos/sst.proto
@@ -25,8 +25,6 @@ message SstMetaData {
   // The time range of the sst
   common.TimeRange time_range = 4;
   common.TableSchema schema = 5;
-  uint64 size = 6;
-  uint64 row_num = 7;
-  analytic_common.StorageFormatOptions storage_format_opts = 8;
-  SstBloomFilter bloom_filter = 9;
+  analytic_common.StorageFormatOptions storage_format_opts = 6;
+  SstBloomFilter bloom_filter = 7;
 }